The extend information of a time zone file with last transition < now
could result in a wrong cached zone because it used the zone of the
last transition.
This could lead to wrong zones in systems with slim zoneinfo.
Fixes#42216
Change-Id: I7c57c35b5cfa58482ac7925b5d86618c52f5444d
Reviewed-on: https://go-review.googlesource.com/c/go/+/264939
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Some programs have a lot of timers that they adjust both forward and
backward in time. This can cause a large number of timerModifiedEarlier
timers. In practice these timers are used for I/O deadlines and are
rarely reached. The effect is that the runtime spends a lot of time
in adjusttimers making sure that there are no timerModifiedEarlier
timers, but the effort is wasted because none of the adjusted timers
are near the top of the timer heap anyhow.
Avoid much of this extra work by keeping track of the earliest known
timerModifiedEarlier timer. This lets us skip adjusttimers if we know
that none of the timers will be ready to run anyhow. We will still
eventually run it, when we reach the deadline of the earliest known
timerModifiedEarlier, although in practice that timer has likely
been removed. When we do run adjusttimers, we will reset all of the
timerModifiedEarlier timers, and clear our notion of when we need
to run adjusttimers again.
This effect should be to significantly reduce the number of times we
walk through the timer list in adjusttimers.
Fixes#41699
Change-Id: I38eb2be611fb34e3017bb33d0a9ed40d75fb414f
Reviewed-on: https://go-review.googlesource.com/c/go/+/258303
Trust: Ian Lance Taylor <iant@golang.org>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Pretty minor concern, but after auditing the compiler/runtime for
conversions from pointers to go:notinheap types to unsafe.Pointer,
this is the only remaining one I found.
Update #42076
Change-Id: I81d5b893c9ada2fc19a51c2559262f2e9ff71c35
Reviewed-on: https://go-review.googlesource.com/c/go/+/265757
Trust: Keith Randall <khr@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
pointers to go:notinheap types should be treated as scalars. That
means they shouldn't be stored directly in interfaces, or directly
in reflect.Value.ptr.
Also be sure to use uintpr to compare such pointers in reflect.DeepEqual.
Fixes#42076
Change-Id: I53735f6d434e9c3108d4940bd1bae14c61ef2a74
Reviewed-on: https://go-review.googlesource.com/c/go/+/264480
Trust: Keith Randall <khr@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
storeType splits compound stores up into a scalar parts and a pointer parts.
The scalar part happens unconditionally, and the pointer part happens
under the guard of a write barrier check.
Types which are declared as pointers, but are represented as scalars because
they might have "bad" values, were not handled correctly here. They ended
up not getting stored in either set.
Fixes#42032
Change-Id: I46f6600075c0c370e640b807066247237f93c7ac
Reviewed-on: https://go-review.googlesource.com/c/go/+/264300
Trust: Keith Randall <khr@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
rewrite.go has two identical functions isPowerOfTwo and
isPowerOfTwo64; the former has been there for a while, while the
latter was added together with isPowerOfTwo{8,16,32} for use in typed
rules.
This change deletes isPowerOfTwo and switch to using isPowerOfTwo64
everywhere.
Change-Id: If26c94565d2393fac6f0ba117ee7ee2fc915f7cd
Reviewed-on: https://go-review.googlesource.com/c/go/+/265417
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Changes:
- makeValAndOff is deleted in favour of MakeValAndOff{32,64}
- canAdd is renamed to canAdd64 to uniform with existing canAdd32
- addOffset{32,64} is simplified by directly using MakeValAndOff{32,64}
- ValAndOff.Int64 is removed
Change-Id: Ic01db7fa31ddfe0aaaf1d1d77af823d48a7bee84
Reviewed-on: https://go-review.googlesource.com/c/go/+/265357
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Instead of always polling 500ms, start with an interval of 1ms and
exponentially back off to at most 500ms. 10% jitter is added to each
interval.
This makes Shutdown more responsive when connections and listeners
close quickly.
Also removes the need for the polling interval to be changed in tests
since if tests' connections and listeners close quickly Shutdown will
also return quickly.
Fixes#42156
Change-Id: I5e59844a2980c09adebff57ae8b58817965e6db4
Reviewed-on: https://go-review.googlesource.com/c/go/+/264479
Run-TryBot: Emmanuel Odeke <emmanuel@orijtech.com>
Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Trust: Bryan C. Mills <bcmills@google.com>
Combine (AND m (SRWconst x)) or (SRWconst (AND m x)) when mask m is
and the shift value produce constant which can be encoded into an
RLWINM instruction.
Combine (CLRLSLDI (SRWconst x)) if the combining of the underling rotate
masks produces a constant which can be encoded into RLWINM.
Likewise for (SLDconst (SRWconst x)) and (CLRLSDI (RLWINM x)).
Combine rotate word + and operations which can be encoded as a single
RLWINM/RLWNM instruction.
The most notable performance improvements arise from the crypto
benchmarks below (GOARCH=power8 on a ppc64le/linux):
pkg:golang.org/x/crypto/blowfish goos:linux goarch:ppc64le
ExpandKeyWithSalt 52.2µs ± 0% 47.5µs ± 0% -8.88%
ExpandKey 44.4µs ± 0% 40.3µs ± 0% -9.15%
pkg:golang.org/x/crypto/ssh/internal/bcrypt_pbkdf goos:linux goarch:ppc64le
Key 57.6ms ± 0% 52.3ms ± 0% -9.13%
pkg:golang.org/x/crypto/bcrypt goos:linux goarch:ppc64le
Equal 90.9ms ± 0% 82.6ms ± 0% -9.13%
DefaultCost 91.0ms ± 0% 82.7ms ± 0% -9.12%
Change-Id: I59a0ca29face38f4ab46e37124c32906f216c4ce
Reviewed-on: https://go-review.googlesource.com/c/go/+/260798
Run-TryBot: Carlos Eduardo Seo <carlos.seo@linaro.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Carlos Eduardo Seo <carlos.seo@linaro.com>
Trust: Lynn Boger <laboger@linux.vnet.ibm.com>
Change the scheduler to treat expired timers with the same approach it
uses to steal runnable G's.
Previously the scheduler ignored timers on P's not marked for
preemption. That had the downside that any G's waiting on those expired
timers starved until the G running on their P completed or was
preempted. That could take as long as 20ms if sysmon was in a 10ms
wake up cycle.
In addition, a spinning P that ignored an expired timer and found no
other work would stop despite there being available work, missing the
opportunity for greater parallelism.
With this change the scheduler no longer ignores timers on
non-preemptable P's or relies on sysmon as a backstop to start threads
when timers expire. Instead it wakes an idle P, if needed, when
creating a new timer because it cannot predict if the current P will
have a scheduling opportunity before the new timer expires. The P it
wakes will determine how long to sleep and block on the netpoller for
the required time, potentially stealing the new timer when it wakes.
This change also eliminates a race between a spinning P transitioning
to idle concurrently with timer creation using the same pattern used
for submission of new goroutines in the same window.
Benchmark analysis:
CL 232199, which was included in Go 1.15 improved timer latency over Go
1.14 by allowing P's to steal timers from P's not marked for preemption.
The benchmarks added in this CL measure that improvement in the
ParallelTimerLatency benchmark seen below. However, Go 1.15 still relies
on sysmon to notice expired timers in some situations and sysmon can
sleep for up to 10ms before waking to check timers. This CL fixes that
shortcoming with modest regression on other benchmarks.
name \ avg-late-ns go14.time.bench go15.time.bench fix.time.bench
ParallelTimerLatency-8 17.3M ± 3% 7.9M ± 0% 0.2M ± 3%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=1-8 53.4k ±23% 50.7k ±31% 252.4k ± 9%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=2-8 204k ±14% 90k ±58% 188k ±12%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=3-8 1.17M ± 0% 0.11M ± 5% 0.11M ± 2%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=4-8 1.81M ±44% 0.10M ± 4% 0.10M ± 2%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=5-8 2.28M ±66% 0.09M ±13% 0.08M ±21%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=6-8 2.84M ±85% 0.07M ±15% 0.07M ±18%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=7-8 2.13M ±27% 0.06M ± 4% 0.06M ± 9%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=8-8 2.63M ± 6% 0.06M ±11% 0.06M ± 9%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=9-8 3.32M ±17% 0.06M ±16% 0.07M ±14%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=10-8 8.46M ±20% 4.37M ±21% 5.03M ±23%
StaggeredTickerLatency/work-dur=2ms/tickers-per-P=1-8 1.02M ± 1% 0.20M ± 2% 0.20M ± 2%
name \ max-late-ns go14.time.bench go15.time.bench fix.time.bench
ParallelTimerLatency-8 18.3M ± 1% 8.2M ± 0% 0.5M ±12%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=1-8 141k ±19% 127k ±19% 1129k ± 3%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=2-8 2.78M ± 4% 1.23M ±15% 1.26M ± 5%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=3-8 6.05M ± 5% 0.67M ±56% 0.81M ±33%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=4-8 7.93M ±20% 0.71M ±46% 0.76M ±41%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=5-8 9.41M ±30% 0.92M ±23% 0.81M ±44%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=6-8 10.8M ±42% 0.8M ±41% 0.8M ±30%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=7-8 9.62M ±24% 0.77M ±38% 0.88M ±27%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=8-8 10.6M ±10% 0.8M ±32% 0.7M ±27%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=9-8 11.9M ±36% 0.6M ±46% 0.8M ±38%
StaggeredTickerLatency/work-dur=300µs/tickers-per-P=10-8 36.8M ±21% 24.7M ±21% 27.5M ±16%
StaggeredTickerLatency/work-dur=2ms/tickers-per-P=1-8 2.12M ± 2% 1.02M ±11% 1.03M ± 7%
Other time benchmarks:
name \ time/op go14.time.bench go15.time.bench fix.time.bench
AfterFunc-8 137µs ± 4% 123µs ± 4% 131µs ± 2%
After-8 212µs ± 3% 195µs ± 4% 204µs ± 7%
Stop-8 165µs ± 6% 156µs ± 2% 151µs ±12%
SimultaneousAfterFunc-8 260µs ± 3% 248µs ± 3% 284µs ± 2%
StartStop-8 65.8µs ± 9% 64.4µs ± 7% 67.3µs ±15%
Reset-8 13.6µs ± 2% 9.6µs ± 2% 9.1µs ± 4%
Sleep-8 307µs ± 4% 306µs ± 3% 320µs ± 2%
Ticker-8 53.0µs ± 5% 54.5µs ± 5% 57.0µs ±11%
TickerReset-8 9.24µs ± 2% 9.51µs ± 3%
TickerResetNaive-8 149µs ± 5% 145µs ± 5%
Fixes#38860
Updates #25471
Updates #27707
Change-Id: If52680509b0f3b66dbd1d0c13fa574bd2d0bbd57
Reviewed-on: https://go-review.googlesource.com/c/go/+/232298
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Trust: Ian Lance Taylor <iant@golang.org>
misc/cgo/test fails in 'dist test' on arm64 if the C compiler is of GCC-9.4 or
above and its 'outline atomics' feature is enabled, since the internal linking
hasn't yet supported "__attribute__((constructor))" and also mis-handles hidden
visibility.
This change addresses the problem by skipping the internal linking cases of
misc/cgo/test on linux/arm64. It fixes 'dist test' failure only, user is expected to
pass a GCC option '-mno-outline-atomics' via CGO_CFLAGS if running into the same
problem when building cgo programs using internal linking.
Updates #39466
Change-Id: I57f9e85fca881e5fd2dae6c1b4446bce9e0c1975
Reviewed-on: https://go-review.googlesource.com/c/go/+/262357
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Previously, 2 goroutines were created in OpenDB and a comment in the DB.close() field indicated that they were canceled. Later, session Resetter () was removed, but the comment remained the same. This commit just fixes this message
Change-Id: Ie81026f51d7770e9cf8004818154021f626fb2e8
GitHub-Last-Rev: 38b338a0d1
GitHub-Pull-Request: golang/go#42191
Reviewed-on: https://go-review.googlesource.com/c/go/+/264838
Reviewed-by: Daniel Theophanes <kardianos@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Trust: Brad Fitzpatrick <bradfitz@golang.org>
Add additional rules to eliminate unnecessary sign/zero extension for riscv64.
Also where possible, replace an extension following a load with a different typed
load. This removes almost another 8,000 instructions from the go binary.
Of particular note, change Eq16/Eq8/Neq16/Neq8 to zero extend each value before
subtraction, rather than zero extending after subtraction. While this appears to
double the number of zero extensions, it often lets us completely eliminate them
as the load can already be performed in a properly typed manner.
As an example, prior to this change runtime.memequal16 was:
0000000000013028 <runtime.memequal16>:
13028: 00813183 ld gp,8(sp)
1302c: 00019183 lh gp,0(gp)
13030: 01013283 ld t0,16(sp)
13034: 00029283 lh t0,0(t0)
13038: 405181b3 sub gp,gp,t0
1303c: 03019193 slli gp,gp,0x30
13040: 0301d193 srli gp,gp,0x30
13044: 0011b193 seqz gp,gp
13048: 00310c23 sb gp,24(sp)
1304c: 00008067 ret
Whereas it now becomes:
0000000000012fa8 <runtime.memequal16>:
12fa8: 00813183 ld gp,8(sp)
12fac: 0001d183 lhu gp,0(gp)
12fb0: 01013283 ld t0,16(sp)
12fb4: 0002d283 lhu t0,0(t0)
12fb8: 405181b3 sub gp,gp,t0
12fbc: 0011b193 seqz gp,gp
12fc0: 00310c23 sb gp,24(sp)
12fc4: 00008067 ret
Change-Id: I16321feb18381241cab121c0097a126104c56c2c
Reviewed-on: https://go-review.googlesource.com/c/go/+/264659
Trust: Joel Sing <joel@sing.id.au>
Run-TryBot: Joel Sing <joel@sing.id.au>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
CryptGenRandom appears to be unfavorable these days, whereas the classic
RtlGenRandom is still going strong.
This commit also moves the warnBlocked function into rand_unix, rather
than rand, because it's now only used on unix.
Fixes#33542
Change-Id: I5c02a5917572f54079d627972401efb6e1ce4057
Reviewed-on: https://go-review.googlesource.com/c/go/+/210057
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Rather than handling sign and zero extension via rules, defer to the assembler
and use MOV pseudo-instructions. The instruction can also be omitted where the
type and size is already correct. This change results in more than 6,000
instructions being removed from the go binary (in part due to omitted
instructions, in part due to MOVBU having a more efficient implementation in
the assembler than what is used in the current ZeroExt8to{16,32,64} rules).
This will also allow for further rewriting to remove redundant sign/zero
extension.
Change-Id: I05e42fd9f09f40a69948be7de772cce8946c8744
Reviewed-on: https://go-review.googlesource.com/c/go/+/264658
Trust: Joel Sing <joel@sing.id.au>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Instead of duplicating PackageOpts fields in the loaderParams struct,
embed the PackageOpts directly. Many of the fields are duplicated, and
further fields that would also be duplicated will be added in
subsequent changes.
For #36460
Change-Id: I3b0770d162e901d23ec1643183eb07c413d51e0a
Reviewed-on: https://go-review.googlesource.com/c/go/+/263138
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
Currently, "x &^ y" gets rewriten into "x & ^y" during walk. It adds
unnecessary complexity to other parts, which must aware about this.
Instead, we can just implement "&^" in the conversion to SSA, so "&^"
can be handled like other binary operators.
However, this CL does not pass toolstash-check. It seems that implements
"&^" in the conversion to SSA causes registers allocation change.
With the parent:
obj: 00212 (.../src/runtime/complex.go:47) MOVQ X0, AX
obj: 00213 (.../src/runtime/complex.go:47) BTRQ $63, AX
obj: 00214 (.../src/runtime/complex.go:47) MOVQ "".n(SP), CX
obj: 00215 (.../src/runtime/complex.go:47) MOVQ $-9223372036854775808, DX
obj: 00216 (.../src/runtime/complex.go:47) ANDQ DX, CX
obj: 00217 (.../src/runtime/complex.go:47) ORQ AX, CX
With this CL:
obj: 00212 (.../src/runtime/complex.go:47) MOVQ X0, AX
obj: 00213 (.../src/runtime/complex.go:47) BTRQ $63, AX
obj: 00214 (.../src/runtime/complex.go:47) MOVQ $-9223372036854775808, CX
obj: 00215 (.../src/runtime/complex.go:47) MOVQ "".n(SP), DX
obj: 00216 (.../src/runtime/complex.go:47) ANDQ CX, DX
obj: 00217 (.../src/runtime/complex.go:47) ORQ AX, DX
Change-Id: I80acf8496a91be4804fb7ef3df04c19baae2754c
Reviewed-on: https://go-review.googlesource.com/c/go/+/264660
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
This change modifies addrRanges.findSucc to more efficiently find the
successor range in an addrRanges by using a binary search to narrow down
large addrRanges and iterate over no more than 8 addrRanges.
This change makes the runtime more robust against systems that may
aggressively randomize the address space mappings it gives the runtime
(e.g. Fuchsia).
For #40191.
Change-Id: If529df2abd2edb1b1496d8690ddd284ecd7138c2
Reviewed-on: https://go-review.googlesource.com/c/go/+/242679
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Since MemStats is now populated directly and some values are derived,
avoid duplicating the logic by instead populating the heap dump directly
from MemStats (external version) instead of memstats (runtime internal
version).
Change-Id: I0bec96bfa02d2ffd1b56475779c124a760e64238
Reviewed-on: https://go-review.googlesource.com/c/go/+/255817
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
CL 253748 introduced a special case in cmd/go to allow sync to import
runtime/internal/atomic. Besides introducing unnecessary complexity
into cmd/go, this breaks other packages (like gopls) that understand
how imports work, but don't understand this special case.
Fix this by using the more standard linkname-based approach to pull
the necessary functions from runtime/internal/atomic into sync. Since
these are compiler intrinsics, we also have to tell the compiler that
the linknamed symbols are intrinsics to get this optimization in sync.
Fixes#42196.
Change-Id: I1f91498c255c91583950886a89c3c9adc39a32f0
Reviewed-on: https://go-review.googlesource.com/c/go/+/265124
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Paul Murphy <murp@ibm.com>
TryBot-Result: Go Bot <gobot@golang.org>
When a function with non-zero frame size makes a return jump
(RET target), it assembles to, conceptually,
MOV (SP), LR
ADD $framesize, SP
JMP target
We did not clear some fields in the first instruction's Prog.To,
causing it printed like (on ARM)
MOVW.P 4(R13), (R14)(R14)(REG)
Clear the fields to make it print nicer.
Change-Id: I180901aeea41f1ff287d7c6034a6d69005927744
Reviewed-on: https://go-review.googlesource.com/c/go/+/264343
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Joel Sing <joel@sing.id.au>
This change adds a new benchmark to the runtime tests for measuring the
latency of the new metrics implementation, based on the
ReadMemStats latency benchmark. readMetrics will have more metrics added
to it in the future, and this benchmark will serve as a way to measure
the cost of adding additional metrics.
Change-Id: Ib05e3ed4afa49a70863fc0c418eab35b72263e24
Reviewed-on: https://go-review.googlesource.com/c/go/+/247042
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change adds support for a variety of runtime memory metrics and
contains the base implementation of Read for the runtime/metrics
package, which lives in the runtime.
It also adds testing infrastructure for the metrics package, and a bunch
of format and documentation tests.
For #37112.
Change-Id: I16a2c4781eeeb2de0abcb045c15105f1210e2d8a
Reviewed-on: https://go-review.googlesource.com/c/go/+/247041
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Trust: Michael Knyszek <mknyszek@google.com>
This change moves the mcache-local malloc stats into the
consistentHeapStats structure so the malloc stats can be managed
consistently with the memory stats. The one exception here is
tinyAllocs for which moving that into the global stats would incur
several atomic writes on the fast path. Microbenchmarks for just one CPU
core have shown a 50% loss in throughput. Since tiny allocation counnt
isn't exposed anyway and is always blindly added to both allocs and
frees, let that stay inconsistent and flush the tiny allocation count
every so often.
Change-Id: I2a4b75f209c0e659b9c0db081a3287bf227c10ca
Reviewed-on: https://go-review.googlesource.com/c/go/+/247039
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change replaces stacks_inuse, gcWorkBufInUse and
gcProgPtrScalarBitsInUse with their corresponding consistent stats. It
also adds checks to make sure the rest of the sharded stats line up with
existing stats in updatememstats.
Change-Id: I17d0bd181aedb5c55e09c8dff18cef5b2a3a14e3
Reviewed-on: https://go-review.googlesource.com/c/go/+/247038
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change adds a global set of heap statistics which are similar
to existing memory statistics. The purpose of these new statistics
is to be able to read them and get a consistent result without stopping
the world. The goal is to eventually replace as many of the existing
memstats statistics with the sharded ones as possible.
The consistent memory statistics use a tailor-made synchronization
mechanism to allow writers (allocators) to proceed with minimal
synchronization by using a sequence counter and a global generation
counter to determine which set of statistics to update. Readers
increment the global generation counter to effectively grab a snapshot
of the statistics, and then iterate over all Ps using the sequence
counter to ensure that they may safely read the snapshotted statistics.
To keep statistics fresh, the reader also has a responsibility to merge
sets of statistics.
These consistent statistics are computed, but otherwise unused for now.
Upcoming changes will integrate them with the rest of the codebase and
will begin to phase out existing statistics.
Change-Id: I637a11f2439e2049d7dccb8650c5d82500733ca5
Reviewed-on: https://go-review.googlesource.com/c/go/+/247037
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change adds a function getMCache which returns the current P's
mcache if it's available, and otherwise tries to get mcache0 if we're
bootstrapping. This function will come in handy as we need to replicate
this behavior in multiple places in future changes.
Change-Id: I536073d6f6dc6c6390269e613ead9f8bcb6e7f98
Reviewed-on: https://go-review.googlesource.com/c/go/+/246976
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This statistic is updated in many places but for MemStats may be
computed from existing statistics. Specifically by definition
heap_idle = heap_sys - heap_inuse since heap_sys is all memory allocated
from the OS for use in the heap minus memory used for non-heap purposes.
heap_idle is almost the same (since it explicitly includes memory that
*could* be used for non-heap purposes) but also doesn't include memory
that's actually used to hold heap objects.
Although it has some utility as a sanity check, it complicates
accounting and we want fewer, orthogonal statistics for upcoming metrics
changes, so just drop it.
Change-Id: I40af54a38e335f43249f6e218f35088bfd4380d1
Reviewed-on: https://go-review.googlesource.com/c/go/+/246974
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change breaks apart gc_sys into three distinct pieces. Two of those
pieces are pieces which come from heap_sys since they're allocated from
the page heap. The rest comes from memory mapped from e.g.
persistentalloc which better fits the purpose of a sysMemStat. Also,
rename gc_sys to gcMiscSys.
Change-Id: I098789170052511e7b31edbcdc9a53e5c24573f7
Reviewed-on: https://go-review.googlesource.com/c/go/+/246973
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Currently MemStats is populated via an unsafe memmove from memstats, but
this places unnecessary structural restrictions on memstats, is annoying
to reason about, and tightly couples the two. Instead, just populate the
fields of MemStats explicitly.
Change-Id: I96f6a64326b1a91d4084e7b30169a4bbe6a331f9
Reviewed-on: https://go-review.googlesource.com/c/go/+/246972
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change modifies the type of several mstats fields to be a new type:
sysMemStat. This type has the same structure as the fields used to have.
The purpose of this change is to make it very clear which stats may be
used in various functions for accounting (usually the platform-specific
sys* functions, but there are others). Currently there's an implicit
understanding that the *uint64 value passed to these functions is some
kind of statistic whose value is atomically managed. This understanding
isn't inherently problematic, but we're about to change how some stats
(which currently use mSysStatInc and mSysStatDec) work, so we want to
make it very clear what the various requirements are around "sysStat".
This change also removes mSysStatInc and mSysStatDec in favor of a
method on sysMemStat. Note that those two functions were originally
written the way they were because atomic 64-bit adds required a valid G
on ARM, but this hasn't been the case for a very long time (since
golang.org/cl/14204, but even before then it wasn't clear if mutexes
required a valid G anymore). Today we implement 64-bit adds on ARM with
a spinlock table.
Change-Id: I4e9b37cf14afc2ae20cf736e874eb0064af086d7
Reviewed-on: https://go-review.googlesource.com/c/go/+/246971
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change modifies mheap's span allocation API to have each caller
declare a purpose, defined as a new enum called spanAllocType.
The purpose behind this change is two-fold:
1. Tight control over who gets to allocate heap memory is, generally
speaking, a good thing. Every codepath that allocates heap memory
places additional implicit restrictions on the allocator. A notable
example of a restriction is work bufs coming from heap memory: write
barriers are not allowed in allocation paths because then we could
have a situation where the allocator calls into the allocator.
2. Memory statistic updating is explicit. Instead of passing an opaque
pointer for statistic updating, which places restrictions on how that
statistic may be updated, we use the spanAllocType to determine which
statistic to update and how.
We also take this opportunity to group all the statistic updating code
together, which should make the accounting code a little easier to
follow.
Change-Id: Ic0b0898959ba2a776f67122f0e36c9d7d60e3085
Reviewed-on: https://go-review.googlesource.com/c/go/+/246970
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change renames a bunch of malloc statistics stored in the mcache
that are all named with the "local_" prefix. It also renames largeAlloc
to allocLarge to prevent a naming conflict, and next_sample because it
would be the last mcache field with the old C naming style.
Change-Id: I29695cb83b397a435ede7e9ad5c3c9be72767ea3
Reviewed-on: https://go-review.googlesource.com/c/go/+/246969
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Now that local_scan is the last mcache-based statistic that is flushed
by purgecachedstats, and heap_scan and gcController.revise may be
interacted with concurrently, we don't need to flush heap_scan at
arbitrary locations where the heap is locked, and we don't need
purgecachedstats and cachestats anymore. Instead, we can flush
local_scan at the same time we update heap_live in refill, so the two
updates may share the same revise call.
Clean up unused functions, remove code that would cause the heap to get
locked in the allocSpan when it didn't need to (other than to flush
local_scan), and flush local_scan explicitly in a few important places.
Notably we need to flush local_scan whenever we flush the other stats,
but it doesn't need to be donated anywhere, so have releaseAll do the
flushing. Also, we need to flush local_scan before we set heap_scan at
the end of a GC, which was previously handled by cachestats. Just do so
explicitly -- it's not much code and it becomes a lot more clear why we
need to do so.
Change-Id: I35ac081784df7744d515479896a41d530653692d
Reviewed-on: https://go-review.googlesource.com/c/go/+/246968
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Trust: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change removes mcentral.nmalloc and adds mcache.local_nsmallalloc
which fulfills the same role but may be accessed non-atomically. It also
moves responsibility for updating heap_live and local_nsmallalloc into
mcache functions.
As a result of this change, mcache is now the sole source-of-truth for
malloc stats. It is also solely responsible for updating heap_live and
performing the various operations required as a result of updating
heap_live. The overall improvement here is in code organization:
previously malloc stats were fairly scattered, and now they have one
single home, and nearly all the required manipulations exist in a single
file.
Change-Id: I7e93fa297c1debf17e3f2a0d68aeed28a9c6af00
Reviewed-on: https://go-review.googlesource.com/c/go/+/246966
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change makes nlargealloc and largealloc into mcache fields just
like nlargefree and largefree. These local fields become the new
source-of-truth. This change also moves the accounting for these fields
out of allocSpan (which is an inappropriate place for it -- this
accounting generally happens much closer to the point of allocation) and
into largeAlloc. This move is partially possible now that we can call
gcController.revise at that point.
Furthermore, this change moves largeAlloc into mcache.go and makes it a
method of mcache. While there's a little bit of a mismatch here because
largeAlloc barely interacts with the mcache, it helps solidify the
mcache as the first allocation layer and provides a clear place to
aggregate and manage statistics.
Change-Id: I37b5e648710733bb4c04430b71e96700e438587a
Reviewed-on: https://go-review.googlesource.com/c/go/+/246965
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change makes it so that various local malloc stats (excluding
heap_scan and local_tinyallocs) are no longer written first to mheap
fields but are instead accessed directly from each mcache.
This change is part of a move toward having stats be distributed, and
cleaning up some old code related to the stats.
Note that because there's no central source-of-truth, when an mcache
dies, it must donate its stats to another mcache. It's always safe to
donate to the mcache for the 0th P, so do that.
Change-Id: I2556093dbc27357cb9621c9b97671f3c00aa1173
Reviewed-on: https://go-review.googlesource.com/c/go/+/246964
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
This change makes it so that the GC assist ratio (the pair of
gcControllerState fields assistBytesPerWork and assistWorkPerByte) is
updated atomically. Note that the pair of fields are not updated
together atomically, but that's OK. The code here was already racy for
some time and in practice the assist ratio moves very slowly.
The purpose of this change is so that we can document
gcController.revise to be safe for concurrent use, which will be useful
in further changes.
Change-Id: Ie25d630207c88e4f85f2b8953f6a0051ebf1b4ea
Reviewed-on: https://go-review.googlesource.com/c/go/+/246963
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
next_gc is mostly updated only during a STW, but may occasionally be
updated by calls to e.g. debug.SetGCPercent. In this case the update is
supposed to be protected by the heap lock, but in reality it's accessed
by gcController.revise which may be called without the heap lock held
(despite its documentation, which will be updated in a later change).
Change the synchronization policy on next_gc so that it's atomically
accessed when the world is not stopped to aid in making revise safe for
concurrent use.
Change-Id: I79657a72f91563f3241aaeda66e8a7757d399529
Reviewed-on: https://go-review.googlesource.com/c/go/+/246962
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
gcControllerState.scanWork's docs state that it must be accessed
atomically during a GC cycle, but gcControllerState.revise does not do
this (even when called with the heap lock held).
This change makes it so that gcControllerState.revise accesses scanWork
atomically and explicitly.
Note that we don't update gcControllerState.revise's erroneous doc
comment here because this change isn't about revise's guarantees, just
about heap_scan. The comment is updated in a later change.
Change-Id: Iafc3ad214e517190bfd8a219896d23da19f7659d
Reviewed-on: https://go-review.googlesource.com/c/go/+/246961
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Currently heap_scan is mostly protected by the heap lock, but
gcControllerState.revise sometimes accesses it without a lock. In an
effort to make gcControllerState.revise callable from more contexts (and
have its synchronization guarantees actually respected), make heap_scan
atomically read from and written to, unless the world is stopped.
Note that we don't update gcControllerState.revise's erroneous doc
comment here because this change isn't about revise's guarantees, just
about heap_scan. The comment is updated in a later change.
Change-Id: Iddbbeb954767c704c2bd1d221f36e6c4fc9948a6
Reviewed-on: https://go-review.googlesource.com/c/go/+/246960
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
The Windows callback support accepts Go functions with arguments that
are uintptr-sized or smaller. However, it doesn't implement smaller
arguments correctly. It assumes the Windows arguments layout is
equivalent to the Go argument layout. This is often true, but because
Windows C ABIs pad arguments to word size, while Go packs arguments,
the layout is different if there are multiple sub-word-size arguments
in a row. For example, a function with two uint16 arguments will have
a two-word C argument frame, but only a 4 byte Go argument frame.
There are also subtleties surrounding floating-point register
arguments that it doesn't handle correctly.
To fix this, when constructing a callback, we examine the Go
function's signature to construct a mapping between the C argument
frame and the Go argument frame. When the callback is invoked, we use
this mapping to build the Go argument frame and copy the result back.
This adds several test cases to TestStdcallAndCDeclCallbacks that
exercise more complex function signatures. These all fail with the
current code, but work with this CL.
In addition to fixing these callback types, this is also a step toward
the Go register ABI (#40724), which is going to make the ABI
translation more complex.
Change-Id: I19fb1681b659d9fd528ffd5e88912bebb95da052
Reviewed-on: https://go-review.googlesource.com/c/go/+/263271
Trust: Austin Clements <austin@google.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
This simplifies the systematic test of Windows callbacks with
different signatures and prepares it for expanded coverage of function
signatures.
It now returns a result from the Go function and threads it back
through C. This simplifies things, but also previously the code could
have succeeded by simply not calling the callbacks at all (though
other tests would have caught that).
It bundles together the C function description and the Go function
it's intended to call. Now the test source generation and the test
running both loop over a single slice of test functions.
Since the C function and Go function are now bundled, it generates the
C function by reflectively inspecting the signature of the Go
function. For the moment, we keep the same test suite, which is
entirely functions with "uintptr" arguments, but we'll expand this
shortly.
It now use sub-tests. This way tests automatically get useful
diagnostic labels in failures and the tests don't have to catch panics
on their own.
It eliminates the DLL function argument. I honestly couldn't figure
out what the point of this was, and it added what appeared to be an
unnecessary loop level to the tests.
Change-Id: I120dfd4785057cc2c392bd2c821302f276bd128e
Reviewed-on: https://go-review.googlesource.com/c/go/+/263270
Trust: Austin Clements <austin@google.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
On amd64 and 386, we have a very roundabout way of remembering that we
need to dropm on return that currently involves saving a zero to
needm's argument slot and later bringing it back. Just store the zero.
This also makes amd64 and 386 more consistent with cgocallback on all
other platforms: rather than saving the old M to the G stack, they now
save it to a named slot on the G0 stack.
The needm function no longer needs a dummy argument to get the SP, so
we drop that.
Change-Id: I7e84bb4a5ff9552de70dcf41d8accf02310535e7
Reviewed-on: https://go-review.googlesource.com/c/go/+/263268
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This redesigns the way calls work from C to exported Go functions. It
removes several steps from the call path, makes cmd/cgo no longer
sensitive to the Go calling convention, and eliminates the use of
reflectcall from cgo.
In order to avoid generating a large amount of FFI glue between the C
and Go ABIs, the cgo tool has long depended on generating a C function
that marshals the arguments into a struct, and then the actual ABI
switch happens in functions with fixed signatures that simply take a
pointer to this struct. In a way, this CL simply pushes this idea
further.
Currently, the cgo tool generates this argument struct in the exact
layout of the Go stack frame and depends on reflectcall to unpack it
into the appropriate Go call (even though it's actually
reflectcall'ing a function generated by cgo).
In this CL, we decouple this struct from the Go stack layout. Instead,
cgo generates a Go function that takes the struct, unpacks it, and
calls the exported function. Since this generated function has a
generic signature (like the rest of the call path), we don't need
reflectcall and can instead depend on the Go compiler itself to
implement the call to the exported Go function.
One complication is that syscall.NewCallback on Windows, which
converts a Go function into a C function pointer, depends on
cgocallback's current dynamic calling approach since the signatures of
the callbacks aren't known statically. For this specific case, we
continue to depend on reflectcall. Really, the current approach makes
some overly simplistic assumptions about translating the C ABI to the
Go ABI. Now we're at least in a much better position to do a proper
ABI translation.
For comparison, the current cgo call path looks like:
GoF (generated C function) ->
crosscall2 (in cgo/asm_*.s) ->
_cgoexp_GoF (generated Go function) ->
cgocallback (in asm_*.s) ->
cgocallback_gofunc (in asm_*.s) ->
cgocallbackg (in cgocall.go) ->
cgocallbackg1 (in cgocall.go) ->
reflectcall (in asm_*.s) ->
_cgoexpwrap_GoF (generated Go function) ->
p.GoF
Now the call path looks like:
GoF (generated C function) ->
crosscall2 (in cgo/asm_*.s) ->
cgocallback (in asm_*.s) ->
cgocallbackg (in cgocall.go) ->
cgocallbackg1 (in cgocall.go) ->
_cgoexp_GoF (generated Go function) ->
p.GoF
Notably:
1. We combine _cgoexp_GoF and _cgoexpwrap_GoF and move the combined
operation to the end of the sequence. This combined function also
handles reflectcall's previous role.
2. We combined cgocallback and cgocallback_gofunc since the only
purpose of having both was to convert a raw PC into a Go function
value. We instead construct the Go function value in cgocallbackg1.
3. cgocallbackg1 no longer reaches backwards through the stack to get
the arguments to cgocallback_gofunc. Instead, we just pass the
arguments down.
4. Currently, we need an explicit msanwrite to mark the results struct
as written because reflectcall doesn't do this. Now, the results are
written by regular Go assignments, so the Go compiler generates the
necessary MSAN annotations. This also means we no longer need to track
the size of the arguments frame.
Updates #40724, since now we don't need to teach cgo about the
register ABI or change how it uses reflectcall.
Change-Id: I7840489a2597962aeb670e0c1798a16a7359c94f
Reviewed-on: https://go-review.googlesource.com/c/go/+/258938
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Stackmap symbols are content-addressable, so it may be dedup'd
with another symbol with same content. We want stackmap symbols
4-byte aligned. But if it dedup's with another symbol with larger
alignment, preserve that alignment.
Fixes#42071.
Change-Id: I1616dd2b0c175b2aac8f68782a5c7a62053c0b57
Reviewed-on: https://go-review.googlesource.com/c/go/+/264897
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Joel Sing <joel@sing.id.au>
Reviewed-by: Than McIntosh <thanm@google.com>
Do not mimic glibc behavior if /etc/nsswitch.conf is missing. This will
will likely be missing on musl libc systems and glibc systems will likely
always have it, resulting in localhost lookup being done over DNS rather
than from /etc/hosts.
Do what makes most sense rather than making any assumption about the
libc.
Fixes#35305
Change-Id: I20bd7e24131bba8eaa39a20c8950fe552364784d
GitHub-Last-Rev: 119409839d
GitHub-Pull-Request: golang/go#39685
Reviewed-on: https://go-review.googlesource.com/c/go/+/238629
Run-TryBot: Dan Peterson <dpiddy@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Peterson <dpiddy@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Fixes#42079
Previously, EvalSymlinks returned an error when called with the root of
a UNC share (e.g. \\server\share). This was due to Windows's
FindFirstFile function not supporting a share root path.
To resolve this, now return early from toNorm in the case where the path
after the volume name is empty. Skipping the later path component
resolution shouldn't have any negative impact in this case, as if the
path is empty, there aren't any path components to resolve anyways.
The test case uses the localhost admin share (c$), as it should be
present in most situations. This allows testing without setting up an
external file share. However, this fix applies to all UNC share root
paths.
Change-Id: I05035bd86be93662d7bea34fab4b75fc8e918206
GitHub-Last-Rev: bd3db2cda6
GitHub-Pull-Request: golang/go#42096
Reviewed-on: https://go-review.googlesource.com/c/go/+/263917
Trust: Alex Brainman <alex.brainman@gmail.com>
Trust: Giovanni Bajo <rasky@develer.com>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
It might not be obvious from reading the code why we consider the test
as failed on incomplete ports even though it passed. Add a comment
documenting this behavior, as suggested by Dmitri in CL 155839.
Change-Id: I3eb7db27d01d63db277172381e5fa51577dad941
Reviewed-on: https://go-review.googlesource.com/c/go/+/264682
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Add support for signed and unsigned register to register moves of various
sizes. This makes it easier to handle zero and sign extension and will allow
for further changes that improve the compiler optimisations for riscv64.
While here, change the existing register to register moves from obj.Prog
rewriting to instruction generation.
Change-Id: Id21911019b76922367a134da13c3449a84a1fb08
Reviewed-on: https://go-review.googlesource.com/c/go/+/264657
Trust: Joel Sing <joel@sing.id.au>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Updates the (*Data).Ranges method to work with DWARFv5 which uses the
new debug_rnglists section instead of debug_ranges.
This does not include supporting DW_FORM_rnglistx.
General support for DWARFv5 was added by CL 175138.
Change-Id: I01f919a865616a3ff12f5bf649c2c9abf89fcf52
Reviewed-on: https://go-review.googlesource.com/c/go/+/236657
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
For some reason, currently unknown, this test case fails exclusively
on the linux-ppc64 platform. Until such time as it can be made to
work, we'll disable this test case on that platform.
The same issue causes TestSetuidEtc to fail too, so disable that
on this platform.
Updates #42178
Change-Id: Idd3f6c2ee9f2fba2eb8ce4de69de7f316858bb15
Reviewed-on: https://go-review.googlesource.com/c/go/+/264719
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
This prevents the custom-built version of openssl prefering the system
libraries over the ones compiled with the specified (weak crypto)
options necessary to generate the updates. This difference can lead to
confusing failures when updating the tests.
Fixes#31809
Change-Id: I2dd257f3121d6c6c62c6aeba52e1c74046b3c584
GitHub-Last-Rev: 6d4eeafadf
GitHub-Pull-Request: golang/go#41630
Reviewed-on: https://go-review.googlesource.com/c/go/+/257517
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
addrRanges represents a set of addresses. Currently, passing in a
zero-sized range will cause that range to be added to the list, even
though it doesn't represent any address (addrRanges.contains will still
always return false, and findSucc will give surprising results).
We could ignore this input, but it's almost always a bug for the calling
code to pass in a zero-sized range, so just throw.
Change-Id: I8ed09e15b79a3a33e2d0cf5ed55f9e497388e7a5
Reviewed-on: https://go-review.googlesource.com/c/go/+/242817
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Austin Clements <austin@google.com>
To allow passing around the package level *Logger, it is now exposed to
callers of the Default function. We considered exposing std, however at
this time there is no need to allow callers to replace std only pass and
call methods directly.
Fixes#39057
Change-Id: I710b16a3aa5e4e878870561dbf59560f98d8d09a
Reviewed-on: https://go-review.googlesource.com/c/go/+/264460
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
The existing usage of grave accent (`) and apostrophe (')
at some places made godoc to ignore them and show it as it is.
So, use both of the characters twice (consecutively) so that
godoc can convert it to {left,right} double quotation mark.
Fixes#41958
Change-Id: I64fd9b5fa34f416ad595009d09f5482e10bd8b4f
Reviewed-on: https://go-review.googlesource.com/c/go/+/262397
Reviewed-by: Russ Cox <rsc@golang.org>
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
This reset of gcBlackenEnabled is a no-op because it was already reset
almost immediately before in gcMarkDone, which is the only caller of
gcMarkTermination.
Adjust the comment to clarify setGCPhase a bit more. We are coming from
_GCmark, so write barriers are already enabled.
Change-Id: Ieac2dadf33c3c5a44e8a25a499dea8cfe03b8d73
Reviewed-on: https://go-review.googlesource.com/c/go/+/241357
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
In readonly mode, if a package is not provided by any module in the
build list, and there is an unused replacement that contains the
package, we now recommend a 'go get' command to add a requirement on
the highest replaced version.
Fixes#41416
Change-Id: Iedf3539292c70ea6ba6857433fd184454d9325da
Reviewed-on: https://go-review.googlesource.com/c/go/+/263146
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Previously, we would retain entries in go.sum for .mod files in the
module graph (reachable from the main module) and for .zip files
of modules providing packages.
This isn't quite enough: when we load a package, we need the content
of each module in the build list that *could* provide the package
(that is, each module whose path is a prefix of the package's path) so
we can diagnose ambiguous imports.
For #33008
Change-Id: I0b4d9d68c1f4ca382f0983a3a7e537764f35c3aa
Reviewed-on: https://go-review.googlesource.com/c/go/+/262781
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
If the go command needs a .mod or .zip file in -mod=readonly mode
(now the default), and that file doesn't have a hash in the main
module's go.sum file, the go command will now report an error before
fetching the file, rather than at the end when failing to update
go.sum. The error says specifically which entry is missing.
If this error is encountered when loading the build list, it will
suggest 'go mod tidy'.
If this error is encountered when loading a specific package (an
import or command line argument), the error will mention that package
and will suggest 'go mod tidy' or 'go get -d'.
Fixes#41934Fixes#41935
Change-Id: I96ec2ef9258bd4bade9915c43d47e6243c376a81
Reviewed-on: https://go-review.googlesource.com/c/go/+/262341
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
When 'go mod init' is run in an existing project, it may import
requirements from a vendor configuration file, but the requirements
may not be complete, and go.sum won't contain sums for module
zips. With -mod=readonly, the next build command is likely to fail.
'go mod init' will now suggest running 'go mod tidy' if there are .go
files or subdirectories in the current directory.
We could potentially run 'go mod tidy' automatically within
'go mod init', but it seems better to guide users to using 'go mod tidy'
as a separate command to fix missing dependencies.
For #41712
Updates #40278
Change-Id: Iaece607f291244588a732ef4c5d576108965ca91
Reviewed-on: https://go-review.googlesource.com/c/go/+/264622
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
InitMod is split into two functions. LoadModFile parses an existing
go.mod file and loads the build list (or checks vendor/modules.txt for
consistency in vendor mode). CreateModFile creates a new go.mod file,
possibly inferring the module path and importing a vendor
configuration file.
Some logic is moved from runInit to CreateModFile. init-specific logic
is removed from other functions.
This CL shouldn't cause substantial differences in behavior, though
some error messages are slightly different.
For #41712
Change-Id: Ia684945cfcf5beca30bbb81e7144fc246c4f27ed
Reviewed-on: https://go-review.googlesource.com/c/go/+/264621
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Previously, if a go.mod file was present, and it only contained a
module directive, any module-aware command would attempt to import
requirements from a vendor configuration file like Gopkg.lock.
This CL removes that functionality. It was undocumented and untested,
and it can cause problems with -mod=readonly. It should never come up
for go.mod files created with 'go mod init', since they have a "go"
directive.
For #40278
Change-Id: I64c0d67d204560aa5c775d29553883d094fd3b72
Reviewed-on: https://go-review.googlesource.com/c/go/+/264620
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
This change adds two new methods for invoking system calls
under Linux: syscall.AllThreadsSyscall() and
syscall.AllThreadsSyscall6().
These system call wrappers ensure that all OSThreads mirror
a common system call. The wrappers serialize execution of the
runtime to ensure no race conditions where any Go code observes
a non-atomic OS state change. As such, the syscalls have
higher runtime overhead than regular system calls, and only
need to be used where such thread (or 'm' in the parlance
of the runtime sources) consistency is required.
The new support is used to enable these functions under Linux:
syscall.Setegid(), syscall.Seteuid(), syscall.Setgroups(),
syscall.Setgid(), syscall.Setregid(), syscall.Setreuid(),
syscall.Setresgid(), syscall.Setresuid() and syscall.Setuid().
They work identically to their glibc counterparts.
Extensive discussion of the background issue addressed in this
patch can be found here:
https://github.com/golang/go/issues/1435
In the case where cgo is used, the C runtime can launch pthreads that
are not managed by the Go runtime. As such, the added
syscall.AllThreadsSyscall*() return ENOTSUP when cgo is enabled.
However, for the 9 syscall.Set*() functions listed above, when cgo is
active, these functions redirect to invoke their C.set*() equivalents
in glibc, which wraps the raw system calls with a nptl:setxid fixup
mechanism. This achieves POSIX semantics for these functions in the
combined Go and C runtime.
As a side note, the glibc/nptl:setxid support (2019-11-30) does not
extend to all security related system calls under Linux so using
native Go (CGO_ENABLED=0) and these AllThreadsSyscall*()s, where
needed, will yield more well defined/consistent behavior over all
threads of a Go program. That is, using the
syscall.AllThreadsSyscall*() wrappers for things like setting state
through SYS_PRCTL and SYS_CAPSET etc.
Fixes#1435
Change-Id: Ib1a3e16b9180f64223196a32fc0f9dce14d9105c
Reviewed-on: https://go-review.googlesource.com/c/go/+/210639
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
Trust: Ian Lance Taylor <iant@golang.org>
Trust: Michael Pratt <mpratt@google.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Austin Clements <austin@google.com>
This adds a pass to detect common selection operations,
to avoid generating duplicates. Duplicate offsets are
also detected.
All aggregate types are now handled; there is some freedom in where
expand_calls is run, though it must run before softfloat.
Debug-name-maintenance is now incremental both in decompose builtin
and in expand_calls; it might be good to push this into all the
decompose passes.
(this is a smash of 5 CLs that rewrote some of the same code several
times to deal with phase-ordering problems, and included an abandoned
attempt.)
For #40724.
Change-Id: I2a0c32f20660bf8b99e2bcecd33545d97d2bd3c6
Reviewed-on: https://go-review.googlesource.com/c/go/+/249458
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The history of pageAlloc using 's' as a receiver are lost to the depths
of time (perhaps it used to be called summary?), but it doesn't make
much sense anymore. Rename it to 'p'.
Generated with:
$ cd src/runtime
$ grep -R -b "func (s \*pageAlloc" . | awk -F : '{ print $1 ":#" $2+6 }' | xargs -n 1 -I {} env GOROOT=$(pwd)/../../ gorename -offset {} -to p -v
$ grep -R -b "func (s \*pageAlloc" . | awk -F : '{ print $1 ":#" $2+6 }' | xargs -n 1 -I {} env GOROOT=$(pwd)/../../ GOARCH=386 gorename -offset {} -to p -v
$ GOROOT=$(pwd)/../../ gorename -offset mpagecache.go:#2397 -to p -v
($2+6 to advance past "func (".)
Plus manual comment fixups.
Change-Id: I2d521a1cbf6ebe2ef6aae92e654bfc33c63d1aa9
Reviewed-on: https://go-review.googlesource.com/c/go/+/250517
Trust: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
According to #28614, proposal review agreed in December 2018 that
Match should return an error for failed matches where the unmatched
part of the pattern has a syntax error. (The failed match has to date
caused the scan of the pattern to stop early.)
This change implements that behavior: the match loop continues
scanning to the end of the pattern, even after a confirmed mismatch,
to check whether the pattern is even well-formed.
The change applies to both path.Match and filepath.Match.
Then filepath.Glob and fs.Glob make a single validity-checking
call to Match before beginning their usual processing.
Also update fstest.TestFS to check for correct validation in custom
Glob implementations.
Fixes#28614.
Change-Id: Ic1d35a4bb9c3565184ae83dbefc425c5c96318e7
Reviewed-on: https://go-review.googlesource.com/c/go/+/264397
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Work stealing is a scalability bottleneck in the scheduler. Since each P
has a work queue, work stealing must look at every P to determine if
there is any work. The number of Ps scales linearly with GOMAXPROCS
(i.e., the number of Ps _is_ GOMAXPROCS), thus this work scales linearly
with GOMAXPROCS.
Work stealing is a later attempt by a P to find work before it goes
idle. Since the P has no work of its own, extra costs here tend not to
directly affect application-level benchmarks. Where they show up is
extra CPU usage by the process as a whole. These costs get particularly
expensive for applications that transition between blocked and running
frequently.
Long term, we need a more scalable approach in general, but for now we
can make a simple observation: idle Ps ([1]) cannot possibly have
anything in their runq, so we need not bother checking at all.
We track idle Ps via a new global bitmap, updated in pidleput/pidleget.
This is already a slow path (requires sched.lock), so we don't expect
high contention there.
Using a single bitmap avoids the need to touch every P to read p.status.
Currently, the bitmap approach is not significantly better than reading
p.status. However, in a future CL I'd like to apply a similiar
optimization to timers. Once done, findrunnable would not touch most Ps
at all (in mostly idle programs), which will avoid memory latency to
pull those Ps into cache.
When reading this bitmap, we are racing with Ps going in and out of
idle, so there are a few cases to consider:
1. _Prunning -> _Pidle: Running P goes idle after we check the bitmap.
In this case, we will try to steal (and find nothing) so there is no
harm.
2. _Pidle -> _Prunning while spinning: A P that starts running may queue
new work that we miss. This is OK: (a) that P cannot go back to sleep
without completing its work, and (b) more fundamentally, we will recheck
after we drop our P.
3. _Pidle -> _Prunning after spinning: After spinning, we really can
miss work from a newly woken P. (a) above still applies here as well,
but this is also the same delicate dance case described in findrunnable:
if nothing is spinning anymore, the other P will unpark a thread to run
the work it submits.
Benchmark results from WakeupParallel/syscall/pair/race/1ms (see
golang.org/cl/228577):
name old msec new msec delta
Perf-task-clock-8 250 ± 1% 247 ± 4% ~ (p=0.690 n=5+5)
Perf-task-clock-16 258 ± 2% 259 ± 2% ~ (p=0.841 n=5+5)
Perf-task-clock-32 284 ± 2% 270 ± 4% -4.94% (p=0.032 n=5+5)
Perf-task-clock-64 326 ± 3% 303 ± 2% -6.92% (p=0.008 n=5+5)
Perf-task-clock-128 407 ± 2% 363 ± 5% -10.69% (p=0.008 n=5+5)
Perf-task-clock-256 561 ± 1% 481 ± 1% -14.20% (p=0.016 n=4+5)
Perf-task-clock-512 840 ± 5% 683 ± 2% -18.70% (p=0.008 n=5+5)
Perf-task-clock-1024 1.38k ±14% 1.07k ± 2% -21.85% (p=0.008 n=5+5)
[1] "Idle Ps" here refers to _Pidle Ps in the sched.pidle list. In other
contexts, Ps may temporarily transition through _Pidle (e.g., in
handoffp); those Ps may have work.
Updates #28808
Updates #18237
Change-Id: Ieeb958bd72e7d8fb375b0b1f414e8d7378b14e29
Reviewed-on: https://go-review.googlesource.com/c/go/+/259578
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Trust: Michael Pratt <mpratt@google.com>
This commit contains the compiler support for //go:embed lines.
The go command passes to the compiler an "embed config"
that maps literal patterns like *.txt to the set of files to embed.
The compiler then lays out the content of those files as static data
in the form of an embed.Files or string or []byte in the final object file.
The test for this code is the end-to-end test hooking up the
embed, cmd/compile, and cmd/go changes, in the next CL.
For #41191.
Change-Id: I916e57f8cc65871dc0044c13d3f90c252a3fe1bf
Reviewed-on: https://go-review.googlesource.com/c/go/+/243944
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
embed.FS is the implementation of embedded file trees, providing
an fs.FS for each embed.FS variable.
Tests are in a follow-up CL, in the package embed/internal/embedtest.
(They can only be written once the toolchain can initialize one of these,
which requires changes to cmd/compile and cmd/go.)
For #41191.
Change-Id: Ieb0ead1d305cdac3d5d4e11772dca75740a72730
Reviewed-on: https://go-review.googlesource.com/c/go/+/243942
Trust: Russ Cox <rsc@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This is somewhat academic (and no tests failed before this commit),
since func atof64 only calls func eiselLemire when func atof64exact
fails, and func atof64exact doesn't fail when parsing positive or
negative zeroes. But it's still worth fixing.
Change-Id: Ibe6ef4c8fd96827673b711d5456003fbc447e39c
Reviewed-on: https://go-review.googlesource.com/c/go/+/264140
Trust: Nigel Tao <nigeltao@golang.org>
Trust: Robert Griesemer <gri@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
In https://golang.org/cl/263357, I wasn't thinking that the assembly
file without a GOARCH suffix would be built for all architectures. Only
build assembly for amd64, and update the stderr matching line.
I manually verified that this works on 386; since the only Go file in
that package is a stub, and no assembly files match GOARCH=386, no
assembly is built at all.
Change-Id: Ief3c6c9bdc223f342821b0ec27f00098fc25246a
Reviewed-on: https://go-review.googlesource.com/c/go/+/264457
Trust: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
The Go toolchain has supported the simulator for years, but always in
buildmode=c-archive which is intrinsically externally linked and PIE.
This CL moves that support from GOOS=darwin GOARCH=amd64 -tags=ios to
just GOOS=ios GOARCH=amd64 to match the change for iOS devices.
This change also forces external linking and defaults to buildmode=pie
to support Go binaries in the default buildmode to run on the simulator.
CL 255257 added the necessary support to the exec wrapper.
Updates #38485Fixes#42100
Change-Id: I6e6ee0e8d421be53b31e3d403880e5b9b880d031
Reviewed-on: https://go-review.googlesource.com/c/go/+/263798
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Trust: Elias Naur <mail@eliasnaur.com>
This way, a -toolexec tool can tell precisely what package is being
built when it's run. This was very hard to do before, because the tool
had to piece together that information given the build action's
arguments or flags.
Since there wasn't a good set of tests for -toolexec, add one in the
form of a test script. It builds a simple set of packages with a variety
of build tools, to ensure that all the cases behave as expected.
Like other recent master changes, include the changelog item for this
user-facing change too.
Fixes#15677.
Change-Id: I0a5a1d9485840323ec138b2e64b7e7dd803fdf90
Reviewed-on: https://go-review.googlesource.com/c/go/+/263357
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Daniel Martí <mvdan@mvdan.cc>
Current linkmode checking in determining package dependencies doesn't
take multiple -linkmode options into consideration, may lead to redundant
dependency on 'runtime/cgo'.
Fixes the problem and adds a testcase.
Change-Id: Iac5ea9fb3ca5ef931201afd0f3441f41f946c919
Reviewed-on: https://go-review.googlesource.com/c/go/+/263497
Trust: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
gc debug flags are currently stored in a 256-long array, that is then
addressed using the ASCII numeric value of the flag itself (a quirk
inherited from the old C compiler). It is also a little wasteful,
since we only define 16 flags, and the other 240 array elements are
always empty.
This change makes Debug a struct, which also provides static checking
that we're not referencing flags that does not exist.
Change-Id: I2f0dfef2529325514b3398cf78635543cdf48fe0
Reviewed-on: https://go-review.googlesource.com/c/go/+/263539
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
The code in this CL does the work of looking for the "embed" import
and then finding and recording the arguments to //go:embed lines
in Go source files. The go command will use this information to prepare
information about embedded files to pass to the compiler.
The tests of the Package fields end up being end-to-end via the
go command (added in the CL with the go command changes),
like all the other Package fields.
For #41191.
Change-Id: I0c87b71ca809c0031603cc403c030d3088299e6b
Reviewed-on: https://go-review.googlesource.com/c/go/+/243941
Trust: Russ Cox <rsc@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Fixes the check for the reserved namespace prefix
"xml" to be case insensitive, so as to match all variants of:
(('X'|'x')('M'|'m')('L'|'l'))
as mandated by Section 2.3 of https://www.w3.org/TR/REC-xml/
This is a roll forward of CL 203417, which was rolled back by CL 240179.
We've decided that the roll back was incorrect, and any broken tests
should be fixed.
The original CL 203417 was by Tamás Gulácsi.
Fixes#35151
For #39876
Change-Id: I2e6daa7aeb252531fba0b8a56086613e13059528
Reviewed-on: https://go-review.googlesource.com/c/go/+/264024
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Tests that if the combination of:
* HTTP multipart file payload size
* ParseMultipartForm's maxMemory parameter
* the internal leeway buffer size of 10MiB
overflows, then we'll report an overflow instead of silently
passing.
Reapplies and fixes CL 254977, which was reverted in CL 263658.
The prior test lacked a res.Body.Close(), so fixed that and
added a leaked Transport check to verify correctness.
Updates 40430.
Change-Id: I3c0f7ef43d621f6eb00f07755f04f9f36c51f98f
Reviewed-on: https://go-review.googlesource.com/c/go/+/263817
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Damien Neil <dneil@google.com>
The other named errors - UnmarshalTypeError, etc - in this package do
the same, so we should prepend the package prefix to error messages
for consistency.
Add a note to the release docs in case this is interpreted as
a breaking change.
Fixes#36221.
Change-Id: Ie24b532bbf9812e108c259fa377e2a6b64319ed4
Reviewed-on: https://go-review.googlesource.com/c/go/+/263619
Run-TryBot: Kevin Burke <kev@inburke.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Kevin Burke <kev@inburke.com>
Trust: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Bypasses the signature verification check we previously added if the
signature algorithm is MD5WithRSA, as we only support this algorithm
for signing and not verification.
Change-Id: Idba6dbba8b365d6199d467526746b88a5f734af1
Reviewed-on: https://go-review.googlesource.com/c/go/+/264019
Run-TryBot: Roland Shoemaker <roland@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Trust: Roland Shoemaker <roland@golang.org>
The original riscv64 port used the thread pointer (TP aka X4) register for
the g pointer, however this register is also used when TLS support is
required, resulting in a conflict (for example, when a signal is received
we have no way of readily knowing if X4 contains a pointer to the TCB or
a pointer to a g).
In order to support cgo, free up the X4 register by moving g to X27.
This unfortunately means that the X4 register is unused in non-cgo mode,
however the alternative is to not support cgo on this platform.
Update #36641
Change-Id: Idcaf3e8ccbe42972a1b8943aeefde7149d9c960a
Reviewed-on: https://go-review.googlesource.com/c/go/+/263477
Trust: Joel Sing <joel@sing.id.au>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This happens with 'go env' and 'go bug'.
If GOFLAGS variable is set to something like '=value',
running `go env` panics with this error message:
goroutine 1 [running]:
cmd/go/internal/base.SetFromGOFLAGS(0xd96838)
cmd/go/internal/base/goflags.go:101 +0x9a7
main.main()
cmd/go/main.go:188 +0x755
This happens when the 'name' of the flag is not
specified ('=' or '=value'), with any combination of other flags.
Other commands show this error message:
go: parsing $GOFLAGS: non-flag
This happens only with 'env' and 'bug' because we have this:
https://go.googlesource.com/go/+/refs/heads/master/src/cmd/go/internal/base/goflags.go#40
New behaviour: ignore the bad flag, since we don't want to report
that with `go env` or `go bug`.
Fixes: #42013
Change-Id: I72602840ca00293d2a92ea28451b75b9799e3d6c
Reviewed-on: https://go-review.googlesource.com/c/go/+/263098
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Matloob <matloob@golang.org>
wasm_exec.js expects that either "require" is available or that the
globals "crypto", "TextEncoder" and "TextDecoder" are already defined.
Report a better error message if this is not the case, suggesting the
use of a polyfill.
Updates #41482
Change-Id: I5473cae15c98ae42e39f5928245b7762e7a5a8bf
Reviewed-on: https://go-review.googlesource.com/c/go/+/261357
Trust: Richard Musiol <neelance@gmail.com>
Run-TryBot: Richard Musiol <neelance@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The Dbg dumping of complex types was broken, because (I think) of a
recent change to handle recursive types correctly. Before this fix,
the Dump output of a closure node (where the last thing on the line is
the type of the node) was:
. . CLOSURE l(8) esc(h) tc(1) FUNC-@0
after this change it is:
. . CLOSURE l(8) esc(h) tc(1) FUNC-func(int) int
The problem is that that the 'mode == Fdbg' code was immediately
aborting the descent into tconv2, since it was calling down with the
same node that was just entered into the hash table.
Change-Id: Iee106b967cea1856dd92d4350681401dd34a23b3
Reviewed-on: https://go-review.googlesource.com/c/go/+/264025
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Supports 386 and amd64 architectures on all operating systems.
Example output:
$ go test -bench=.*
goos: darwin
goarch: amd64
pkg: strconv
cpu: Intel(R) Core(TM) i7-3520M CPU @ 2.90GHz
BenchmarkAtof64Decimal-4 24431032 46.8 ns/op
...
As the displayed CPU information is only used for information
purposes it is lazily initialized when needed using the new
internal/sysinfo package.
This allows internal/cpu to stay without dependencies and avoid
initialization costs when the CPU information is not needed as
the new code to query the CPU name in internal/cpu can be
dead code eliminated if not used.
Fixes#39214
Change-Id: I77ae5c5d2fed6b28fa78dd45075f9f0a6a7f1bfd
Reviewed-on: https://go-review.googlesource.com/c/go/+/263804
Trust: Martin Möhrmann <moehrmann@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Make code cleaner and a bit more adaptable:
instead of an ever-growing list of arguments and results for readImports,
put everything in a fileInfo struct, and rename function to readGoInfo.
(Not a goInfo struct because it gets used for non-Go source files as well,
but that processing is much simpler.)
The refactoring simplifies the embed work in the next CL,
but this CL makes no semantic changes.
For #41191.
Change-Id: Id2de2a3b8d351adc1c919dcf79dfbe79fc3d5301
Reviewed-on: https://go-review.googlesource.com/c/go/+/243940
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
The main missing piece here was supporting Stat in the overlay
filesystem, in the parts of the package code that determines whether
an command line argument is a file on disk or a directory. so this
change adds a Stat function to the fsys package. It's implemented the
same way as the already existing fsys.lstat function, but instead of
os.Lstat, it calls os.Stat on disk files.
Then, the change changes parts of the package code to use the overlay
Stat instead of the os package's Stat.
For #39958
Change-Id: I8e478ae386f05b48d7dd71bd7e47584f090623df
Reviewed-on: https://go-review.googlesource.com/c/go/+/262617
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Previously, when running cmd/go tests, if the module root directory is
requested when modules are explicitly disabled, we printed a stack
trace in addition to the error message that's normally printed. The
stack trace isn't that useful, and it makes the actual error hard to
find.
Change-Id: I8230d668f3f16659f08d0d685124c41b4055c5b9
Reviewed-on: https://go-review.googlesource.com/c/go/+/263659
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The old ioutil references are still valid, but update our code
to reflect best practices and get used to the new locations.
Code compiled with the bootstrap toolchain
(cmd/asm, cmd/dist, cmd/compile, debug/elf)
must remain Go 1.4-compatible and is excluded.
Also excluded vendored code.
For #41190.
Change-Id: I6d86f2bf7bc37a9d904b6cee3fe0c7af6d94d5b1
Reviewed-on: https://go-review.googlesource.com/c/go/+/263142
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
As proposed and approved in #40025, Discard, NopCloser, and ReadAll
do not really fit into io/ioutil, which exists mainly to hold things that
would cause an import cycle if implemented in io itself, which is to say
things that import "os".
These three do not import "os" - they are generic io helpers like
many of the things in io itself, so it makes sense for them to be there.
Fixes#40025.
Change-Id: I77f47e9b2a72839edf7446997936631980047b67
Reviewed-on: https://go-review.googlesource.com/c/go/+/263141
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Two different functions in the http API expect a FileSystem:
http.FileSystem and http.NewFileTransport.
Add a general converter http.FS to turn an fs.FS into an http.FileSystem
for use with either of these functions.
(The original plan was to add http.HandlerFS taking an fs.FS directly,
but that doesn't help with NewFileTransport.)
For #41190.
Change-Id: I5f242eafe9b963f4387419a2615bdb487c358f16
Reviewed-on: https://go-review.googlesource.com/c/go/+/243939
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Add Stat helper function, StatFS interface, and test.
Add Stat method to fstest.MapFS.
Add testing of Stat method to fstest.TestFS.
For #41190.
Change-Id: Icf8b6eb1c3fa6f93a9be8405ec5a9468fb1da97b
Reviewed-on: https://go-review.googlesource.com/c/go/+/243913
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
It will inevitably be important to be able to pass an operating system
directory to code written to expect an fs.FS.
os.DirFS provides the conversion.
For #41190.
Change-Id: Id1a8fcbe4c7a30de2c47dea0504e9481a88b1b39
Reviewed-on: https://go-review.googlesource.com/c/go/+/243911
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
This change adds basic test helpers for file system code.
The type MapFS is a simple map-based file system for use when
exercising general file system code.
The func TestFS tests a file system implementation.
For #41190.
Change-Id: I5a2036f57e733915ad508651ad7317749794423c
Reviewed-on: https://go-review.googlesource.com/c/go/+/243910
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Because the "ios" build constraint implies "darwin", it is already
included in the general "darwin" flakiness workaround in
cmd/go/internal/robustio. We just need to relax the renameio test
to avoid false-positives there.
I do not expect this change to drive the rate of false-positives down
to zero, but it should at least reduce noise on the build dashboard.
For #42066
Change-Id: Ia33dbd33295fce5b3261b4831f2807ce29b82e65
Reviewed-on: https://go-review.googlesource.com/c/go/+/263777
Trust: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Currently the scavenging tests implicitly assume that the system huge
page size is always strictly less than 4 MiB, or pallocChunkBytes. This
leads to failures on systems with huge pages of this size, and larger.
Filter out those tests on such platforms and add a test for the 4 MiB
case. The scavenger is already equipped to handle this case.
Huge page sizes > 4 MiB are effectively ignored, so also add a test case
to ensure that happens.
Unfortunately we can't actually run these tests in our CI because they
require the platform to provide the right huge page size, but we really
should just parameterize this value so we can test it (there's a TODO
about this already).
Fixes#42053.
Change-Id: Ia576cbf67e178a14a178a893967efbed27d6eb17
Reviewed-on: https://go-review.googlesource.com/c/go/+/263837
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
That is, the following two pieces of shell code are now equivalent:
$ go tool buildid $(go list -export -f {{.Export}} strings)
v_0VqA6yzwuMg2dn4u57/PXcIR2Pb8Mi9yRdcdkwe
$ go list -export -f {{.BuildID}} strings
v_0VqA6yzwuMg2dn4u57/PXcIR2Pb8Mi9yRdcdkwe
This does not expose any information that wasn't available before, but
makes this workflow simpler and faster. In the first example, we have to
execute two programs, and 'go tool buildid' has to re-open the export
data file to read the build ID.
With the new mechanism, 'go list -export' already has the build ID
ready, so we can simply print it out. Moreover, when listing lots of
related packages like './...', we can now obtain all their build IDs at
once.
Fixes#37281.
Change-Id: I8e2f65a08391b3df1a628c6e06e708b8c8cb7865
Reviewed-on: https://go-review.googlesource.com/c/go/+/263542
Trust: Daniel Martí <mvdan@mvdan.cc>
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Jay Conrod <jayconrod@google.com>
When cpu_no_init.go was created most architectures did not have
code in the doinit function. Currently only mips(le), riscv64 and
wasm do not have empty doinit functions.
Keeping cpu_no_init.go around does not reduce the work to satisfy
the build process when adding support for new architectures.
To support a new architecture a new file or build directive has to
be added to an existing file at any rate to define the constant
CacheLinePadSize. A new empty doinit can then be created in the
new file or the existing doinit can be reused when adding the
additional build directive.
Change-Id: I58a97f8cdf1cf1be85c37f4550c40750358aa031
Reviewed-on: https://go-review.googlesource.com/c/go/+/263801
Trust: Martin Möhrmann <moehrmann@google.com>
Reviewed-by: Tobias Klauser <tobias.klauser@gmail.com>
There are many reader behaviors that are subtle and
worth testing, and it's nice to have one complete tester
instead of many incomplete ones.
For #41190, which will use this as part of a larger
file system implementation tester.
Change-Id: Ib4cc7fae94b0d9b45dfacadc52baa77ad3761322
Reviewed-on: https://go-review.googlesource.com/c/go/+/243909
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
The old os references are still valid, but update our code
to reflect best practices and get used to the new locations.
Code compiled with the bootstrap toolchain
(cmd/asm, cmd/dist, cmd/compile, debug/elf)
must remain Go 1.4-compatible and is excluded.
For #41190.
Change-Id: I8f9526977867c10a221e2f392f78d7dec073f1bd
Reviewed-on: https://go-review.googlesource.com/c/go/+/243907
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
ReadDir provides a portable, efficient way to read a directory
and discover the type of directory entries.
This enables a more efficient file system walk, yet to be added.
See #41467 for the proposal review for the API.
Fixes#41467.
Change-Id: I461a526793ae46df48821aa448b04f1705546739
Reviewed-on: https://go-review.googlesource.com/c/go/+/261540
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
OCALLPART is exported in its original form, which is as an OXDOT.
The body of the method value wrapper created in makepartialcall() was
not being typechecked, and that was causing a problem during escape
analysis, so I added code to typecheck the body.
The go executable got slightly bigger with this change (13598111 ->
13598905), because of extra exported methods with OCALLPART (I
believe), while the text size got slightly smaller (9686964 ->
9686643).
This is mainly part of the work to make sure all function bodies can
be exported (for purposes of generics), but might as well fix the
OCALLPART inlining bug as well.
Fixes#18493
Change-Id: If7aa055ff78ed7a6330c6a1e22f836ec567d04fd
Reviewed-on: https://go-review.googlesource.com/c/go/+/263620
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
finlock may be held across a write barrier, which could then acquire the
mheap lock. Notably, this occurs in the mp.unlockf write in gopark where
finlock is held by the finalizer goroutines and is going to sleep.
Fixes#42062.
Change-Id: Icf76637ae6fc12795436272633dca3d473780875
Reviewed-on: https://go-review.googlesource.com/c/go/+/263678
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
asNode(t.Nod).Name.Param will be nil for builtin types (i.e., the
universal predeclared types and unsafe.Pointer). These types can't be
part of a cycle anyway, so we can just skip them.
Fixes#42075.
Change-Id: Ic7a44de65c6bfd16936545dee25e36de8850acf3
Reviewed-on: https://go-review.googlesource.com/c/go/+/263717
Trust: Matthew Dempsky <mdempsky@google.com>
Trust: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
This extends a change made in https://golang.org/cl/252258 to the go
command (to define an asm macro when GOEXPERIMENT=regabi is in
effect); we need this same macro during the bootstrap build in order
to build the runtime correctly.
In addition, expand the set of packages where the macro is applied to
{runtime, reflect, syscall, runtime/internal/*}, and move the logic
for deciding when something is a "runtime package" out of the
assembler and into cmd/{go,dist}, introducing a new assembler command
line flag instead.
Updates #27539, #40724.
Change-Id: Ifcc7f029f56873584de1e543c55b0d3e54ad6c49
Reviewed-on: https://go-review.googlesource.com/c/go/+/262317
Trust: Than McIntosh <thanm@google.com>
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The fastest compression mode can pick up a false match for every 2GB
of input data resulting in incorrectly decompressed data.
Since matches are allowed to be up to and including at maxMatchOffset
we must offset the buffer by an additional element to prevent the first
4 bytes to match after an out-of-reach value after shiftOffsets has
been called.
We offset by `maxMatchOffset + 1` so offset 0 in the table will now
fail the `if offset > maxMatchOffset` in all cases.
Fixes#41420
Change-Id: If1fbe01728e132b8a207e3f3f439edd832dcc710
GitHub-Last-Rev: 50fabab0da
GitHub-Pull-Request: golang/go#41477
Reviewed-on: https://go-review.googlesource.com/c/go/+/255879
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Joe Tsai <thebrokentoaster@gmail.com>
Trust: Matthew Dempsky <mdempsky@google.com>
Like we do for applying relocations, for generator symbols, run
the generator function along with symbol writing. This will
probably have better locality and parallelism.
Linking cmd/compile,
Asmb 29.9ms ± 5% 19.1ms ±12% -36.18% (p=0.000 n=10+9)
TotalTime 351ms ± 3% 339ms ± 2% -3.51% (p=0.000 n=11+10)
Change-Id: I9cda6718bf70b3bcf1b7a501a845d6136234d2ee
Reviewed-on: https://go-review.googlesource.com/c/go/+/263640
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Use GOOS=ios.
Run "xcodebuild -showsdks" correctly.
Wrap testDWARF in subtests, as it calls t.Parallel, otherwise
the two calls of testDWARF will cause t.Parallel being called
twice, which panics.
Updates #38485.
Change-Id: I614c8daa99c83cbfd05a4cfa041968d49279f16a
Reviewed-on: https://go-review.googlesource.com/c/go/+/263639
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Also put Reset in the correct place for the other
benchmarks.
name old time/op new time/op delta
NewWriteSum-8 1.01µs ± 0% 1.01µs ± 1% ~ (p=0.945 n=9+9)
name old speed new speed delta
NewWriteSum-8 31.7MB/s ± 0% 31.6MB/s ± 1% ~ (p=0.948 n=9+9)
name old alloc/op new alloc/op delta
NewWriteSum-8 544B ± 0% 544B ± 0% ~ (all equal)
name old allocs/op new allocs/op delta
NewWriteSum-8 7.00 ± 0% 7.00 ± 0% ~ (all equal)
Fixes#41089
Change-Id: I3dae660adbe4993963130bf3c2636bd53899164b
Reviewed-on: https://go-review.googlesource.com/c/go/+/261960
Trust: Katie Hockman <katie@golang.org>
Trust: Roland Shoemaker <roland@golang.org>
Run-TryBot: Katie Hockman <katie@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
We don't put Go static symbols in the symbol table, as they are
always compiler-generated (there is no way to define a static
symbol in user code in Go). We retain static symbols in assembly
code, as it may be user-defined. Also retain static symbols in C.
Change-Id: Ie0425bb4df33f183857b1fd5ba4b2bdfdc497571
Reviewed-on: https://go-review.googlesource.com/c/go/+/263259
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Function symbols defined and referenced by assembly source currently
always default to ABI0; this patch adds preliminary support for
accepting an explicit ABI selector clause for func defs/refs. This
functionality is currently only enabled when compiling runtime-related
packages (runtime, syscall, reflect). Examples:
TEXT ·DefinedAbi0Symbol<ABI0>(SB),NOSPLIT,$0
RET
TEXT ·DefinedAbi1Symbol<ABIInternal>(SB),NOSPLIT,$0
CALL ·AbiZerolSym<ABI0>(SB)
...
JMP ·AbiInternalSym<ABIInternal>(SB)
RET
Also included is a small change to the code in the compiler that reads
the symabis file emitted by the assembler.
New behavior is currently gated under GOEXPERIMENT=regabi.
Updates #27539, #40724.
Change-Id: Ia22221fe26df0fa002191cfb13bdfaaa38d7df38
Reviewed-on: https://go-review.googlesource.com/c/go/+/260477
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Trust: Than McIntosh <thanm@google.com>
@jstarks suggested that recent versions of Windows provide access to high resolution timers. See
https://github.com/golang/go/issues/8687#issuecomment-656259353
for details.
I tried to run this C program on my Windows 10 computer
```
#include <stdio.h>
#include <Windows.h>
#pragma comment(lib, "Winmm.lib")
// Apparently this is already defined when I use msvc cl.
//#define CREATE_WAITABLE_TIMER_HIGH_RESOLUTION = 0x00000002;
int usleep(HANDLE timer, LONGLONG d) {
LARGE_INTEGER liDueTime;
DWORD ret;
LARGE_INTEGER StartingTime, EndingTime, ElapsedMicroseconds;
LARGE_INTEGER Frequency;
QueryPerformanceFrequency(&Frequency);
QueryPerformanceCounter(&StartingTime);
liDueTime.QuadPart = d;
liDueTime.QuadPart = liDueTime.QuadPart * 10; // us into 100 of ns units
liDueTime.QuadPart = -liDueTime.QuadPart; // negative for relative dure time
if (!SetWaitableTimer(timer, &liDueTime, 0, NULL, NULL, 0)) {
printf("SetWaitableTimer failed: errno=%d\n", GetLastError());
return 1;
}
ret = WaitForSingleObject(timer, INFINITE);
if (ret != WAIT_OBJECT_0) {
printf("WaitForSingleObject failed: ret=%d errno=%d\n", ret, GetLastError());
return 1;
}
QueryPerformanceCounter(&EndingTime);
ElapsedMicroseconds.QuadPart = EndingTime.QuadPart - StartingTime.QuadPart;
ElapsedMicroseconds.QuadPart *= 1000000;
ElapsedMicroseconds.QuadPart /= Frequency.QuadPart;
printf("delay is %lld us - slept for %lld us\n", d, ElapsedMicroseconds.QuadPart);
return 0;
}
int testTimer(DWORD createFlag)
{
HANDLE timer;
timer = CreateWaitableTimerEx(NULL, NULL, createFlag, TIMER_ALL_ACCESS);
if (timer == NULL) {
printf("CreateWaitableTimerEx failed: errno=%d\n", GetLastError());
return 1;
}
usleep(timer, 1000LL);
usleep(timer, 100LL);
usleep(timer, 10LL);
usleep(timer, 1LL);
CloseHandle(timer);
return 0;
}
int main()
{
printf("\n1. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is off\n");
testTimer(0);
printf("\n2. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is off\n");
testTimer(CREATE_WAITABLE_TIMER_HIGH_RESOLUTION);
timeBeginPeriod(1);
printf("\n3. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is on\n");
testTimer(0);
printf("\n4. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is on\n");
testTimer(CREATE_WAITABLE_TIMER_HIGH_RESOLUTION);
}
```
and I see this output
```
1. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is off
delay is 1000 us - slept for 4045 us
delay is 100 us - slept for 3915 us
delay is 10 us - slept for 3291 us
delay is 1 us - slept for 2234 us
2. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is off
delay is 1000 us - slept for 1076 us
delay is 100 us - slept for 569 us
delay is 10 us - slept for 585 us
delay is 1 us - slept for 17 us
3. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is on
delay is 1000 us - slept for 742 us
delay is 100 us - slept for 893 us
delay is 10 us - slept for 414 us
delay is 1 us - slept for 920 us
4. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is on
delay is 1000 us - slept for 1466 us
delay is 100 us - slept for 559 us
delay is 10 us - slept for 535 us
delay is 1 us - slept for 5 us
```
That shows, that indeed using CREATE_WAITABLE_TIMER_HIGH_RESOLUTION
will provide sleeps as low as about 500 microseconds, while our
current approach provides about 1 millisecond sleep.
New approach also does not require for timeBeginPeriod to be on,
so this change solves long standing problem with go programs draining
laptop battery, because it calls timeBeginPeriod.
This change will only run on systems where
CREATE_WAITABLE_TIMER_HIGH_RESOLUTION flag is available. If not
available, the runtime will fallback to original code that uses
timeBeginPeriod.
This is how this change affects benchmark reported in issue #14790
name old time/op new time/op delta
ChanToSyscallPing 1.05ms ± 2% 0.68ms ±11% -35.43% (p=0.000 n=10+10)
The benchmark was run with GOMAXPROCS set to 1.
Fixes#8687
Updates #14790
Change-Id: I5b97ba58289c088c17c05292e12e45285c467eae
Reviewed-on: https://go-review.googlesource.com/c/go/+/248699
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Austin Clements <austin@google.com>
In CL 258220 I added replacement versions to the repo versions used in
the modload.Query functions. The versions are computed from a map in
the modfile index, which has a nondeterministic iteration order.
I added a short-circuit condition to skip sorting in the (vastly
common) case where no replacement versions are added. However, while
cleaning up the change I accidentally deleted the line of code that
sets that condition. As a result, the test of that functionality
(mod_get_replaced) has been failing nondeterministically.
This change fixes the condition by comparing the slices before and
after adding versions, rather than by setting a separate variable.
The test now passes reliably (tested with -count=200).
Updates #41577
Updates #41416
Updates #37438
Updates #26241
Change-Id: I49a66a3a5510da00ef42b47f20a168de66100db6
Reviewed-on: https://go-review.googlesource.com/c/go/+/263266
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Also fix BenchmarkAtof64Random* to initialize the test data when none
of the TestAtof* tests are run.
Passing "go test -test.count=5 -test.run=xxx -test.bench=Atof64" on to
benchstat:
name old time/op new time/op delta
Atof64Decimal-4 47.9ns ± 0% 48.3ns ± 1% ~ (p=0.238 n=4+5)
Atof64Float-4 58.3ns ± 3% 57.7ns ± 0% ~ (p=0.151 n=5+5)
Atof64FloatExp-4 107ns ± 0% 71ns ± 1% -33.89% (p=0.016 n=4+5)
Atof64Big-4 163ns ± 0% 166ns ± 2% ~ (p=0.159 n=4+5)
Atof64RandomBits-4 299ns ± 1% 166ns ± 1% -44.41% (p=0.008 n=5+5)
Atof64RandomFloats-4 188ns ± 1% 144ns ± 0% -23.03% (p=0.008 n=5+5)
The canada.json file from github.com/miloyip/nativejson-benchmark is
full of geospatial coordinates (i.e. numbers). With this program:
src, _ := ioutil.ReadFile("canada.json")
for i := 0; i < 5; i++ {
now := time.Now()
for j := 0; j < 10; j++ {
dst := interface{}(nil)
if err := json.Unmarshal(src, &dst); err != nil {
log.Fatal(err)
}
}
fmt.Println(time.Since(now))
}
Median of the 5 printed numbers, lower is better.
Before: 760.819549ms
After: 702.651646ms
Ratio: 1.08x
The new detailedPowersOfTen table weighs in at 596 * 16 = 9536 bytes,
but some of that weight gain can be clawed back, in a follow-up commit,
that folds in the existing powersOfTen table in extfloat.go.
RELNOTE=yes
Change-Id: I3953110deaa1f5f6941e88e8417c4665b649ed80
Reviewed-on: https://go-review.googlesource.com/c/go/+/260858
Run-TryBot: Nigel Tao <nigeltao@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Trust: Nigel Tao <nigeltao@golang.org>
Illumos supports the accept4 syscall, use it in internal/poll.accept
like on other platforms.
Add Accept4 to package syscall despite the package being frozen. The
other option would have been to add this to internal/syscall/unix, but
adding it to syscall avoids duplicating a lot of code in internal/poll
and net/internal/socktest. Also, all other platforms supporting the
accept4 syscall already export Accept4.
Follow CL 97196, CL 40895 and CL 94295
Change-Id: I13b32f0163a683840c02b16722730d9dfdb98f56
Reviewed-on: https://go-review.googlesource.com/c/go/+/256101
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The Query function allows the caller to specify the current version of
the requested module, but the QueryPattern function is missing that
parameter: instead, it always assumes that the current version is the
one selected from the global build list.
This change removes that assumption, instead adding a callback
function to determine the current version. (The callback is currently
invoked once per candidate module, regardless of whether that module
exists, but in a future change we can refactor it to invoke the
callback only when needed.)
For #36460
For #40775
Change-Id: I001a4a8ab24f5b4fcc66a670d9bd305b47e948ba
Reviewed-on: https://go-review.googlesource.com/c/go/+/261640
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
'go mod tidy' has been able to use replaced versions since CL 152739,
but 'go get' failed for many of the same paths. Now that we are
recommending 'go get' more aggressively due to #40728, we should make
that work too.
In the future, we might consider factoring out the new replacementRepo
type so that 'go list' can report the new versions as well.
For #41577
For #41416
For #37438
Updates #26241
Change-Id: I9140c556424b584fdd9bdd0a747842774664a7d8
Reviewed-on: https://go-review.googlesource.com/c/go/+/258220
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
It was added in CL 240740 to fix#39978
but without any discussion of the exported API.
The error can still be returned to fix the issue,
without adding new public API to package io.
Also fix the error message to refer to lower-case write
like the other errors in the package.
Change-Id: I134de5eaf3ac903d73913c5cadcde904c5255d79
Reviewed-on: https://go-review.googlesource.com/c/go/+/262877
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
On 386 and arm, unaligned 64-bit atomics aren't safe, so we check for
this and panic. Currently, we panic by dereferencing nil, which may be
expedient but is pretty user-hostile since it gives no hint of what
the actual problem was.
This CL replaces this with an actual panic. The only subtlety here is
now the atomic assembly implementations are calling back into Go, so
they have to play nicely with stack maps and stack scanning. On 386,
this just requires declaring NO_LOCAL_POINTERS. On arm, this is
somewhat more complicated: first, we have to move the alignment check
into the functions that have Go signatures. Then we have to support
both the tail call from these functions to the underlying
implementation (which requires that they have no frame) and the call
into Go to panic (which requires that they have a frame). We resolve
this by forcing them to have no frame and setting up the frame
manually just before the panic call.
Change-Id: I19f1e860045df64088013db37a18acea47342c69
Reviewed-on: https://go-review.googlesource.com/c/go/+/262778
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
This drops package prefixes from the assembly code on 386 and arm. In
addition to just being nicer, this allows the assembler to
automatically pick up the argument stack map from the Go signatures of
these functions. This doesn't matter right now because these functions
never call back out to Go, but prepares us for the next CL.
Change-Id: I90fed7d4dd63ad49274529c62804211b6390e2e9
Reviewed-on: https://go-review.googlesource.com/c/go/+/262777
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
This is primarily aimed at client requests where the user can supply
their own io.ReadCloser, but also clarifies server request behavior.
A server request body can be one of:
- *body
- *http2RequestBody
- *expectContinueReader
- *maxBytesReader
Of those, *expectContinueReader did not meet these expectations, so this
change also removes the data race.
Change-Id: Id4f1ae573d938347b1123a7b612b271aabb045a4
Reviewed-on: https://go-review.googlesource.com/c/go/+/251087
Reviewed-by: Damien Neil <dneil@google.com>
Reviewed-by: Russ Cox <rsc@golang.org>
Trust: Damien Neil <dneil@google.com>
Run-TryBot: Damien Neil <dneil@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
In preparation for moving os.FileInfo into io/fs.
Also keep syscall from depending on io again.
We want to keep them separated, in case io ever
needs to start depending on time.
For #41190.
Change-Id: I98350fa03accf4a20c75ddebb0e961aa1ccccd2c
Reviewed-on: https://go-review.googlesource.com/c/go/+/243905
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Apparently I never actually understood the new file table in Go
object files. The PC value stream actually encodes the file index
in the per-CU table. I thought it was indexing into a per-function
table, which then contains index to the per-CU table. Remove the
extra indirection.
Change-Id: I0aea5629f7b3888ebe3a04fea437aa15ce89519e
Reviewed-on: https://go-review.googlesource.com/c/go/+/262779
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
When a fix was made at the end of the last release related to
NOPs, it was discovered that the ppc64.s testcase was out of date
and contained comments that weren't being processed. Essentially the
instructions in that test were being assembled but there was no
verification that the encodings weres correct. The ppc64enc.s file
was mostly complete and included the valid encodings for verification.
This change moves ppc64enc.s to ppc64.s and adds the instructions
that were missing.
This also adds a minor fix to asm9.go on the assembly of the
addex that was discovered during this testing.
Change-Id: Iaada1563b137849ad195fa88f32ecc9ab3e1e95f
Reviewed-on: https://go-review.googlesource.com/c/go/+/260217
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Lynn Boger <laboger@linux.vnet.ibm.com>
- Describe requirements on Less more precisely.
- Standardize on x for the variable name of the data being sorted
(was variously a, p, slice).
- Many other minor wording changes.
Fixes#41951.
Change-Id: Ic9e222a53ec035fcc3b5ddfc7f0eefbe1bb2890d
Reviewed-on: https://go-review.googlesource.com/c/go/+/262657
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
This creates space for a different kind of extension field
in LSym without making the struct any larger.
(There are many LSym, so we care about keeping the struct small.)
Change-Id: Ib16edb9e15f54c2a7351c8b875e19684058711e5
Reviewed-on: https://go-review.googlesource.com/c/go/+/243943
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Pass the trimpath flag to cmd/compile to use the correct file paths
for files that are overlaid: that is, the "destination" path in the
overlay's Replace mapping rather than the "source" path.
Also fix paths to go source files provided to the gccgo compiler.
For #39958
Change-Id: I3741aeb2272bd0d5aa32cb28133b61e58264fd39
Reviewed-on: https://go-review.googlesource.com/c/go/+/257198
Trust: Michael Matloob <matloob@golang.org>
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This CL replaces the ad hoc and duplicated logic for detecting
inlinable calls with a single "inlCallee" function, which uses the
"staticValue" helper function introduced in an earlier commit.
Updates #41474.
Change-Id: I103d4091b10366fce1344ef2501222b7df68f21d
Reviewed-on: https://go-review.googlesource.com/c/go/+/256460
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Trust: Matthew Dempsky <mdempsky@google.com>
On current macOS versions a program that receives a signal during an
execve can fail with a SIGILL signal. This appears to be a macOS
kernel bug. It has been reported to Apple.
This CL partially works around the problem by using execLock to not
send preemption signals during execve. Of course some other stray
signal could occur, but at least we can avoid exacerbating the problem.
We can't simply disable signals, as that would mean that the exec'ed
process would start with all signals blocked, which it likely does not
expect.
Fixes#41702
Change-Id: I91b0add967b315671ddcf73269c4d30136e579b4
Reviewed-on: https://go-review.googlesource.com/c/go/+/262438
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
We already allow inlining "if" and "goto" statements, so we might as
well allow "for" loops too. The majority of frontend support is
already there too.
The critical missing feature at the moment is that inlining doesn't
properly reassociate OLABEL nodes with their control statement (e.g.,
OFOR) after inlining. This eventually causes SSA construction to fail.
As a workaround, this CL only enables inlining for unlabeled "for"
loops. It's left to a (yet unplanned) future CL to add support for
labeled "for" loops.
The increased opportunity for inlining leads to a small growth in
binary size. For example:
$ size go.old go.new
text data bss dec hex filename
9740163 320064 230656 10290883 9d06c3 go.old
9793399 320064 230656 10344119 9dd6b7 go.new
Updates #14768.
Fixes#41474.
Change-Id: I827db0b2b9d9fa2934db05caf6baa463f0cd032a
Reviewed-on: https://go-review.googlesource.com/c/go/+/256459
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Escape analysis is currently very naive about identifying calls to
known functions: it only recognizes direct calls to a declared
function, or direct calls to a closure.
This CL adds a new "staticValue" helper function that can trace back
through local variables that were initialized and never reassigned
based on a similar optimization already used by inlining. (And to be
used by inlining in a followup CL.)
Updates #41474.
Change-Id: I8204fd3b1e150ab77a27f583985cf099a8572b2e
Reviewed-on: https://go-review.googlesource.com/c/go/+/256458
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: David Chase <drchase@google.com>
Normally, when variables are declared and initialized using ":=", we
set the variable's n.Name.Defn to point to the initialization
assignment node (i.e., OAS or OAS2). Further, some frontend
optimizations look for variables that are initialized but never
reassigned.
However, when inl.go inlines calls, it was declaring the inlined
variables, and then separately assigning to them. This CL changes
inl.go tweaks the AST to fit the combined declaration+initialization
pattern.
This isn't terribly useful by itself, but it allows further followup
optimizations.
Updates #41474.
Change-Id: I62a9752c60414305679e0ed15a6563baa0224efa
Reviewed-on: https://go-review.googlesource.com/c/go/+/256457
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
On a negative answer, the assignop and convertop functions write the
reason why to a string pointer passed as an argument, likely a C-ism
leftover since the compiler's machine assisted translation to Go.
This change makes why a return parameter.
It also fixes a few places where the assignop/convertop result was
compared to 0. While OXXX's value may be zero now, using the named
constant is more robust.
Change-Id: Id9147ed4c1b97d658d30a2f778f876b7867006b4
Reviewed-on: https://go-review.googlesource.com/c/go/+/261857
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
debugCachedWork and all of its dependent fields and code were added to
aid in debugging issue #27993. Now that the source of the problem is
known and mitigated (via the extra work check after STW in gcMarkDone),
these extra checks are no longer required and simply make the code more
difficult to follow.
Remove it all.
Updates #27993
Change-Id: I594beedd5ca61733ba9cc9eaad8f80ea92df1a0d
Reviewed-on: https://go-review.googlesource.com/c/go/+/262350
Trust: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Currently, there are several panics in token that simply say "illegal!". This CL
adds the values.
This is valuable when the token call is wrapped under several layers and you
can't easily see which value is being passed to token.
Change-Id: Ib04b55cafcd9b9ec6820dcf416fc4d49afaea15f
Reviewed-on: https://go-review.googlesource.com/c/go/+/262017
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org>
Trust: Robert Griesemer <gri@golang.org>
Trust: Jean de Klerk <deklerk@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Add support of internal linking on darwin/arm64 (macOS).
Still incomplete. Pure Go binaries work. Cgo doesn't. TLS is not
set up when cgo is not used (as before) (so asynchronous
preemption is not enabled).
Internal linking is not enabled by default but can be requested
via -ldflags=-linkmode=internal.
Updates #38485.
Change-Id: I1e0c81b6028edcb1ac26dcdafeb9bb3f788cf732
Reviewed-on: https://go-review.googlesource.com/c/go/+/261643
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
This changes checks the signature generated during CreateCertificate
and returns an error if the verification fails. A benchmark is also
added. For RSA keys the delta looks to be insignificant, but for
ECDSA keys it introduces a much larger delta which is not ideal.
name old time/op new time/op delta
RSA_2048-8 1.38ms ± 6% 1.41ms ± 2% ~ (p=0.182 n=10)
ECDSA_P256-8 42.6µs ± 4% 116.8µs ± 4% +174.00% (p=0.000 n=1
Fixes#40458
Change-Id: I22827795bb9bb6868b4fa47391927db1d3bc19a1
Reviewed-on: https://go-review.googlesource.com/c/go/+/259697
Run-TryBot: Roland Shoemaker <roland@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
Trust: Roland Shoemaker <roland@golang.org>
Follow-up for CL 261363 which had to retain the fat tz data format due
to failing test.
The reason for the failed tests was that when caching location data, the
extended time format past the end of zone transitions was not
considered. The respective change was introduced in (*Location).lookup
by CL 215539.
This slims down zoneinfo.zip (and thus also the embedded copy in
time/tzdata) by ~350KB.
Change-Id: I412f79de98ba45358b8696aca784999b3479135e
Reviewed-on: https://go-review.googlesource.com/c/go/+/261877
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
The Node type has shortcuts to access bool and int Values:
func (n *Node) Int64() int64
for n.Val().U.(*Mpint).Int64()
func (n *Node) Bool() bool
for n.Val().U.(bool)
I was convinced we didn't have one for string literal nodes, until I
noticed that we do, it's just called strlit, it's not a method, and
it's later in the file:
func strlit(n *Node) string
This change, for consistency:
- Renames strlit to StringVal and makes it a *Node method
- Renames Bool and Int64 to BoolVal and Int64Val
- Moves StringVal near the other two
Change-Id: I18e635384c35eb3a238fd52b1ccd322b1a74d733
Reviewed-on: https://go-review.googlesource.com/c/go/+/261361
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Setting inittrace=1 causes the runtime to emit a single line to standard error for
each package with init work, summarizing the execution time and memory allocation.
The emitted debug information for init functions can be used to find bottlenecks
or regressions in Go startup performance.
Packages with no init function work (user defined or compiler generated) are omitted.
Tracing plugin inits is not supported as they can execute concurrently. This would
make the implementation of tracing more complex while adding support for a very rare
use case. Plugin inits can be traced separately by testing a main package importing
the plugins package imports explicitly.
$ GODEBUG=inittrace=1 go test
init internal/bytealg @0.008 ms, 0 ms clock, 0 bytes, 0 allocs
init runtime @0.059 ms, 0.026 ms clock, 0 bytes, 0 allocs
init math @0.19 ms, 0.001 ms clock, 0 bytes, 0 allocs
init errors @0.22 ms, 0.004 ms clock, 0 bytes, 0 allocs
init strconv @0.24 ms, 0.002 ms clock, 32 bytes, 2 allocs
init sync @0.28 ms, 0.003 ms clock, 16 bytes, 1 allocs
init unicode @0.44 ms, 0.11 ms clock, 23328 bytes, 24 allocs
...
Inspired by stapelberg@google.com who instrumented doInit
in a prototype to measure init times with GDB.
Fixes#41378
Change-Id: Ic37c6a0cfc95488de9e737f5e346b8dbb39174e1
Reviewed-on: https://go-review.googlesource.com/c/go/+/254659
Trust: Martin Möhrmann <moehrmann@google.com>
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This CL adds support of PIE internal linking on darwin/amd64.
This is also preparation for supporting internal linking on
darwin/arm64 (macOS), which requires PIE for everything.
Updates #38485.
Change-Id: I2ed58583dcc102f5e0521982491fc7ba6f2754ed
Reviewed-on: https://go-review.googlesource.com/c/go/+/261642
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
"aliased" is the function responsible for detecting whether we can
turn "a, b = x, y" into just "a = x; b = y", or we need to pre-compute
y and save it in a temporary variable because it might depend on a.
It currently has two issues:
1. It suboptimally treats assignments to blank as writes to heap
memory. Users generally won't write "_, b = x, y" directly, but it
comes up a lot in generated code within the compiler.
This CL changes it to ignore blank assignments.
2. When deciding whether the assigned variable might be referenced by
pointers, it mistakenly checks Class() and Name.Addrtaken() on "n"
(the *value* expression being assigned) rather than "a" (the
destination expression).
It doesn't appear to result in correctness issues (i.e.,
incorrectly reporting no aliasing when there is potential aliasing),
due to all the (overly conservative) rewrite passes before code
reaches here. But it generates unnecessary code and could have
correctness issues if we improve those other passes to be more
aggressive.
This CL fixes the misuse of "n" for "a" by renaming the variables
to "r" and "l", respectively, to make their meaning clearer.
Improving these two cases shaves 4.6kB of text from cmd/go, and 93kB
from k8s.io/kubernetes/cmd/kubelet:
text data bss dec hex filename
9732136 290072 231552 10253760 9c75c0 go.before
9727542 290072 231552 10249166 9c63ce go.after
97977637 1007051 301344 99286032 5eafc10 kubelet.before
97884549 1007051 301344 99192944 5e99070 kubelet.after
While here, this CL also collapses "memwrite" and "varwrite" into a
single variable. Logically, they're detecting the same thing: are we
assigning to a memory location that a pointer might alias. There's no
need for two variables.
Updates #6853.
Updates #23017.
Change-Id: I5a307b8e20bcd2196e85c55eb025d3f01e303008
Reviewed-on: https://go-review.googlesource.com/c/go/+/261677
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
We generally don't care about errors in resolving a repo if the result
we're looking for is already in the module cache. Moreover, we can
avoid some expense in initializing the repo if all of the methods we
plan to call on it hit in the cache — especially when using
GOPROXY=direct.
This also incidentally fixes a possible (but rare) bug in Download:
we had forgotten to reset the downloaded file in case the Zip method
returned an error after writing a nonzero number of bytes.
For #37438
Change-Id: Ib64f10f763f6d1936536b8e1f7d31ed1b463e955
Reviewed-on: https://go-review.googlesource.com/c/go/+/259158
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Previously, ImportMissingError said
"cannot find module providing package …"
even when we didn't even attempt to find such a module.
Now, we write "no module requirement provides package …"
when we did not attempt to identify a suitable module,
and suggest either 'go mod tidy' or 'go get -d' as appropriate.
Fixes#41576
Change-Id: I979bb999da4066828c54d99a310ea66bb31032ec
Reviewed-on: https://go-review.googlesource.com/c/go/+/258298
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Formatter is mentioned further down, but it's helpful
to add it amongst the verbs and flags.
Background: I spent a while puzzling how "%+v" prints
a stack trace for github.com/pkg/errors when this isn't
documented under 'flags'.
Change-Id: Ic70145902a36780147dedca568b3cf482974fc38
GitHub-Last-Rev: 6571b499f2
GitHub-Pull-Request: golang/go#39860
Reviewed-on: https://go-review.googlesource.com/c/go/+/240000
Reviewed-by: Rob Pike <r@golang.org>
Trust: Rob Pike <r@golang.org>
Trust: Ian Lance Taylor <iant@golang.org>
In CL 240065 we changed it to set to FlagTextAddr. Normally it
is the address of the first function, except on plan9/amd64
where, as FlagTextAddr is not aligned, it is rounded up. Set it
to the actual text start address.
Fixes#41137.
Change-Id: I1bba67f5eb4e24d9f745a11350fc999ff48bff45
Reviewed-on: https://go-review.googlesource.com/c/go/+/261644
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: David du Colombier <0intro@gmail.com>
Reviewed-by: Than McIntosh <thanm@google.com>
We are converting from using error-prone ad-hoc syntax // +build lines
to less error-prone, standard boolean syntax //go:build lines.
The timeline is:
Go 1.16: prepare for transition
- Builds still use // +build for file selection.
- Source files may not contain //go:build without // +build.
- Builds fail when a source file contains //go:build lines without // +build lines. <<<
Go 1.17: start transition
- Builds prefer //go:build for file selection, falling back to // +build
for files containing only // +build.
- Source files may contain //go:build without // +build (but they won't build with Go 1.16).
- Gofmt moves //go:build and // +build lines to proper file locations.
- Gofmt introduces //go:build lines into files with only // +build lines.
- Go vet rejects files with mismatched //go:build and // +build lines.
Go 1.18: complete transition
- Go fix removes // +build lines, leaving behind equivalent // +build lines.
This CL provides part of the <<< marked line above in the Go 1.16 step:
rejecting files containing //go:build but not // +build.
For #41184.
Change-Id: I29b8a789ab1526ab5057f613d5533bd2060ba9cd
Reviewed-on: https://go-review.googlesource.com/c/go/+/240600
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
We are converting from using error-prone ad-hoc syntax // +build lines
to less error-prone, standard boolean syntax //go:build lines.
The timeline is:
Go 1.16: prepare for transition
- Builds still use // +build for file selection.
- Source files may not contain //go:build without // +build.
- Builds fail when a source file contains //go:build lines without // +build lines. <<<
Go 1.17: start transition
- Builds prefer //go:build for file selection, falling back to // +build
for files containing only // +build.
- Source files may contain //go:build without // +build (but they won't build with Go 1.16).
- Gofmt moves //go:build and // +build lines to proper file locations.
- Gofmt introduces //go:build lines into files with only // +build lines.
- Go vet rejects files with mismatched //go:build and // +build lines.
Go 1.18: complete transition
- Go fix removes // +build lines, leaving behind equivalent // +build lines.
This CL provides part of the <<< marked line above in the Go 1.16 step:
rejecting files containing //go:build but not // +build.
Reject any //go:build comments found after actual assembler code
(include #include etc directives), because the go command itself
doesn't read that far.
For #41184.
Change-Id: Ib460bfd380cce4239993980dd208afd07deff3f1
Reviewed-on: https://go-review.googlesource.com/c/go/+/240602
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
We are converting from using error-prone ad-hoc syntax // +build lines
to less error-prone, standard boolean syntax //go:build lines.
The timeline is:
Go 1.16: prepare for transition
- Builds still use // +build for file selection.
- Source files may not contain //go:build without // +build.
- Builds fail when a source file contains //go:build lines without // +build lines. <<<
Go 1.17: start transition
- Builds prefer //go:build for file selection, falling back to // +build
for files containing only // +build.
- Source files may contain //go:build without // +build (but they won't build with Go 1.16).
- Gofmt moves //go:build and // +build lines to proper file locations.
- Gofmt introduces //go:build lines into files with only // +build lines.
- Go vet rejects files with mismatched //go:build and // +build lines.
Go 1.18: complete transition
- Go fix removes // +build lines, leaving behind equivalent // +build lines.
This CL provides part of the <<< marked line above in the Go 1.16 step:
rejecting files containing //go:build but not // +build.
The standard go command checks only consider the top of the file.
This compiler check, along with a separate go vet check for ignored files,
handles the remainder of the file.
For #41184.
Change-Id: I014006eebfc84ab5943de18bc90449e534f150a2
Reviewed-on: https://go-review.googlesource.com/c/go/+/240601
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Make ctxt.shouldBuild return multiple values
instead of modifying *sawBinaryOnly in place.
Also give it a table-driven test.
Cleanup in preparation for boolean expressions,
but nice even if those don't end up happening.
Change-Id: Ibb78b0080070deafac7299a6de87ab8bebeb702d
Reviewed-on: https://go-review.googlesource.com/c/go/+/240598
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This cgo warning shows up when running "go test" in cmd/fix:
warning: no cgo types: exit status 1
(It may show up at other times too.)
Warnings belong on standard error, and while we're here,
make it say where it came from.
Change-Id: I3eaba592a219cc6abf2d584762374f9d6ec5135e
Reviewed-on: https://go-review.googlesource.com/c/go/+/240597
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This makes the changes to the file easier to explain.
Not all the changes may come from the fixers directly,
if the file is not gofmt-ed already.
Change-Id: I81776da446a34a1239a3130317d2aae1437d61a5
Reviewed-on: https://go-review.googlesource.com/c/go/+/240555
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Show constraint-ignored non-.go files in go list, as Package.IgnoredOtherFiles
(same as go/build's IgnoredOtherFiles).
Pass full list of ignored files to vet, to help buildtag checker.
For #41184.
Change-Id: I749868de9082cbbc1efbc59370783c8c82fe735f
Reviewed-on: https://go-review.googlesource.com/c/go/+/240553
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Just like it is useful to report ignored .go files,
it is also useful to report ignored non-.go files.
Also, remove _* and .* files from IgnoredGoFiles.
The fact that they were there at all was a mistake.
This list is for files being ignored due to build constraints.
Change-Id: Ie08252dde9f56bda08647e7ebeab7906e9271c54
Reviewed-on: https://go-review.googlesource.com/c/go/+/240551
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
- For "if" statements without a condition, provide a BadExpr rather than nil
(clients expect IfStmt.Cond != nil since the parser is taking care of
reporting a missing condition).
- For 3-index slice expressions, also provide BadExpr where an index is
required but missing.
- Declare a parser-local error method to hide the embedded error method
so we don't use it by mistake.
Accidentally found while adjusting prototype parser to work for generics.
Change-Id: Iacc211cc60869be05efe9ae630d65dff1dac00a0
Reviewed-on: https://go-review.googlesource.com/c/go/+/261218
Trust: Robert Griesemer <gri@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
go test runs a limited number of vet checks by default.
In the standard library, we run more, both to get additional checking
for the standard library and to get experience with whether to enable
any others by default.
One that experience has shown us should not be enabled by default
is go vet -unreachable. When you are testing, it is common to want to
put an early return or a panic into code to bypass a section of code.
That often causes unreachable code. It's incredibly frustrating if the
result is an "unreachable code" error that keeps your test from completing.
Change-Id: Ib194e87759eb65f5a193d771a9880b38d2fd3ba9
Reviewed-on: https://go-review.googlesource.com/c/go/+/240550
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Currently on darwin, when a symbol needs to be exported, we
export it both statically and dynamically. The dynamic export is
unnecessary for some symbols. Only export the necessary ones.
For special runtime C symbols (e.g. crosscall2), they used to be
exported dynamically, and we had a special case for pclntab to
not include those symbols (otherwise, when the dynamic linker
dedup them, the pclntab entries end up pointing out of the
module's address space). This CL changes it to not export those
symbols, and remove the special case.
Change-Id: I2ab40630742d48a09b86ee150aa5f1f7002b134d
Reviewed-on: https://go-review.googlesource.com/c/go/+/261497
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Add note about using per-use seeds.
Delete "collision-resistant but" in:
> The hash functions are collision-resistant but not cryptographically secure.
"Collision-resistant" has a precise cryptographic meaning that is
incompatible with "not cryptographically secure".
All that is really meant by it here here is "it's a good hash function",
which should be established already.
Also delete:
> The hash value of a given byte sequence is consistent within a
> single process, but will be different in different processes.
This was added for its final clause in response to #37040,
but "The hash value of a given byte sequence" is by design not a
concept in this package. Only "... of a given seed and byte sequence".
And seeds cannot be shared between processes, so again by design
you can't even set up the appropriate first half of the sentence
to say the second half.
Change-Id: I2c02bee0e804ef3b120cb4752bf89e60f3f5ff5d
Reviewed-on: https://go-review.googlesource.com/c/go/+/255968
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
This allows multiline constructs like:
{{"hello" |
printf}}
Now that unclosed actions can span multiple lines,
track and report the start of the action when reporting errors.
Also clean up a few "unexpected <error message>" to be just "<error message>".
Fixes#29770.
Change-Id: I54c6c016029a8328b7902a4b6d85eab713ec3285
Reviewed-on: https://go-review.googlesource.com/c/go/+/254257
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
When a module author retracts a version, the go command should act as
if it doesn't exist unless it's specifically requested.
When converting a revision to a version, we should ignore tags for
retracted versions. For example, if the tag v1.0.0 is retracted, and
branch B points to the same revision, we should convert B to a
pseudo-version, not v1.0.0. Similarly, if B points to a commit after
v1.0.0, we should not use v1.0.0 as the base; we can use an earlier
non-retracted tag or no base.
Fixes#41700
Change-Id: Ia596b05b0780e5acfe6616a04e94d24bd342fbae
Reviewed-on: https://go-review.googlesource.com/c/go/+/261079
Run-TryBot: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Currently, mhdr/methods is emitted with the same len/cap. There's no way
to distinguish between exported and non-exported methods statically.
This CL splits mhdr/methods into two parts, use "len" for number of
exported methods, and "cap" for all methods. This fixes the bug in
issue #22075, which intends to return the number of exported methods but
currently return all methods.
Note that with this encoding, we still can access either
all/exported-only/non-exported-only methods:
mhdr[:cap(mhdr)] // all methods
mhdr // exported methods
mhdr[len(mhdr):cap(mhdr)] // non-exported methods
Thank to Matthew Dempsky (@mdempsky) for suggesting this encoding.
Fixes#22075
Change-Id: If662adb03ccff27407d55a5578a0ed05a15e7cdd
Reviewed-on: https://go-review.googlesource.com/c/go/+/259237
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
On macOS, we can do "go build", can exec, and have the source
tree available, so we can enable more tests.
Skip ones that don't work. Most of them are due to that it
requires external linking (for now) and some tests don't work
with external linking (e.g. runtime deadlock detection). For
them, helper functions CanInternalLink/MustInternalLink are
introduced. I still want to have internal linking implemented,
but it is still a good idea to identify which tests don't work
with external linking.
Updates #38485.
Change-Id: I6b14697573cf3f371daf54b9ddd792acf232f2f2
Reviewed-on: https://go-review.googlesource.com/c/go/+/260719
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
We lost a sign extension that was necessary. The nonnegative comparison
didn't have the correct extension on it. If the larger constant is
positive, but its shorter sign extension is negative, the rule breaks.
Fixes#41872
Change-Id: I6592ef103f840fbb786bf8cb94fd8804c760c976
Reviewed-on: https://go-review.googlesource.com/c/go/+/260701
Trust: Keith Randall <khr@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alberto Donizetti <alb.donizetti@gmail.com>
This brings over the architectures that the gofrontend knows about.
This permits using the main cgo tool for those architectures,
as cgo can be used with -godefs without gc support.
This will help add golang.org/x/sys/unix support for other architectures.
For #37443
Change-Id: I63632b9c5139e71b9ccab8edcc7acdb464229b74
Reviewed-on: https://go-review.googlesource.com/c/go/+/260657
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Updated cmd/pprof.objTool.Disasm to accept
an additional bool param introduced in
https://github.com/google/pprof/pull/520 to support
intel syntax in the assembly report.
Returns an error if the intelSyntax param is set. We use
src/cmd/internal/objfile to disassemble and print assembly
so I am not sure if it is relevant, and if so, how.
Fixes#38802
Updates #36905
Change-Id: Iae2b4322404f232196705f05210f00e2495588d9
Reviewed-on: https://go-review.googlesource.com/c/go/+/248499
Trust: Hyang-Ah Hana Kim <hyangah@gmail.com>
Run-TryBot: Hyang-Ah Hana Kim <hyangah@gmail.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Prior to this commit, the profiling code had a sleep() function that
waits and unblocks on either time.After() or a channel provided by an
http.CloseNotifier derived from a supplied http.ResponseWriter.
According to the documentation, http.CloseNotifier is deprecated:
Deprecated: the CloseNotifier interface predates Go's context package.
New code should use Request.Context instead.
This patch does just that -- sleep() now takes an *http.Request and uses
http.Request.Context() to signal when a request has been cancelled.
Change-Id: I98702314addf494f5743a4f99172dc607389dbb8
GitHub-Last-Rev: c1e37a03ca
GitHub-Pull-Request: golang/go#41756
Reviewed-on: https://go-review.googlesource.com/c/go/+/259157
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
This change makes go env -w check if GOTMPDIR is an absolute path.
If GOTMPDIR is not an absolute and not existing path there will be an
error at every `work.Builder.Init()`. If `go env` has `-u/-w` as
argument `work.Builder.Init()` is not called.
`go env -w GOTMPDIR=` work in the same way as `go env -u GOTMPDIR`.
Fixes#40932
Change-Id: I6b0662302eeace7f20460b6d26c6e59af1111da2
Reviewed-on: https://go-review.googlesource.com/c/go/+/250198
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
The implementation has been ported from reflect, but to avoid
introducing a dependency on strconv, Kind.String() falls back to
"invalid" if the Kind is unknown rather than "kind" + strconv.Itoa(int(k))
Fixes#39286
Change-Id: I82277242a6c41d0146dabd9d20339fe72d562500
Reviewed-on: https://go-review.googlesource.com/c/go/+/235522
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
On iOS, when running under lldb, we install xx_cgo_panicmem as
EXC_BAD_ACCESS handler so we can get a proper Go panic for
SIGSEGV. Only build it on iOS.
Updates #38485.
Change-Id: I801c477439e05920a4bb8fdf5eae6f4923ab8274
Reviewed-on: https://go-review.googlesource.com/c/go/+/259440
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Currently we don't use sigaltstack on darwin/arm64, as is not
supported on iOS. However, it is supported on macOS. Use it.
(iOS remains unchanged.)
Change-Id: Icc154c5e2edf2dbdc8ca68741ad9157fc15a72ee
Reviewed-on: https://go-review.googlesource.com/c/go/+/256917
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
There is no correctness reason to zero out the w.bytes array in (w
*huffmanBitWriter).reset, since w.nbytes is correctly set to zero. The elements of
the bytes array are always written sequentially, with nbytes indicating how many
elements have been written, and are only read up to the current value of nybytes.
We have a pprof profile of a web server that compresses its request/responses, and
the zeroing in reset() is taking up 2.6% of the CPU time of the server (and could
be causing more slowdowns elsewhere due to its effects on the cache). This
overhead may be showing up especially because there are many request/responses
that are all fairly small.
I'm not sure if the zeroing of the bytes array was intended as extra protection of
data across reset uses in the same program, but no protection is needed as long as
the huffman_bit_writer code remains correct.
Change-Id: I67f2b2f56cff9dcc38d8fc0aea885bb010aeedbf
Reviewed-on: https://go-review.googlesource.com/c/go/+/258577
Run-TryBot: Dan Scales <danscales@google.com>
Run-TryBot: Joe Tsai <joetsai@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Klaus Post <klauspost@gmail.com>
Reviewed-by: Joe Tsai <joetsai@google.com>
Trust: Joe Tsai <joetsai@google.com>
Trust: Dan Scales <danscales@google.com>
This CL adds basic support for listing packages with overlays.
The new cmd/go/internal/fs package adds an abstraction for communicating
with the file system that will open files according to their overlaid paths,
and provides functions to override those in the build context to open
overlaid files. There is also some support for executing builds on packages
with overlays. In cmd/go/internal/work.(*Builder).build, paths are mapped
to their overlaid paths before they are given as arguments to tools.
For #39958
Change-Id: I5ec0eb9ebbca303e2f1e7dbe22ec32613bc1fd17
Reviewed-on: https://go-review.googlesource.com/c/go/+/253747
Trust: Michael Matloob <matloob@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Two part fix:
1) bring the type "correction" forward from a later CL in the expand calls series
2) when a leaf-selwect is rewritten in place, update the type (it might have been
changed by the type correction in 1).
Fixes#41736.
Change-Id: Id097efd10481bf0ad92aaead81a7207221c144b5
Reviewed-on: https://go-review.googlesource.com/c/go/+/259203
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Support for GO386=387 is being dropped in Go 1.16. There
is no need for the target to be available for testing on
the master branch (where Go 1.16 development is ongoing).
For #40255.
Change-Id: I4a4ee80b0c0a535b6b0b246fe991f26964eb07ca
Reviewed-on: https://go-review.googlesource.com/c/go/+/257963
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
Mobile targets are not supported by misc-compile trybots, as tracked in
golang.org/issue/25963, and need to be filtered out. The buildall.bash
script was created in CL 9438, back when it was a single all-compile
builder, and it was easier to filter out mobile targets in the script
than to come up with a pattern that matches all non-mobile targets.
As of CL 254740, all mobile targets (Android and iOS) have unique GOOS
values. That makes it it easy to filter them out in x/build/dashboard.
This was done in CL 258057. As a result, it's now viable to simplify
this script and perform all misc-compile target selection in x/build,
rather than having it spread it across two places.
Also, as of CL 10750, the all-compile builder has turned into multiple
misc-compile builders, so update the script description accordingly.
Updates #41610.
Change-Id: I1e33260ac18cf0a70bb68cd8e3db5587100c7e87
Reviewed-on: https://go-review.googlesource.com/c/go/+/257962
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alexander Rakoczy <alex@golang.org>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
Most of the cmd/go tests build the cmd/go binary and run that binary to
test it, but TestAbsolutePath used the GOROOT's cmd/go instead, which
makes debugging confusing and means that make.bash has to be run in each
iteration cycle. Update TestAbsolutePath to use the same go binary as
the rest of the cmd/go tests.
Change-Id: Ib4e8ae707b66f1f75ceb346b98358f5604fd28c1
Reviewed-on: https://go-review.googlesource.com/c/go/+/256979
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
ARM64 used to require that all assembly frame sizes were of the form
16*N+8 because ARM64 requires 16-byte SP alignment and the assembler
added an 8 byte LR slot. This made all of the runtime.call* frame
sizes wonky. The assembler now rounds up the frame size appropriately
after adding any additional slots it needs, so this is no longer
necessary.
This CL cleans up the frame sizes of these functions so they look the
way you'd expect and match all other architectures.
Change-Id: I47819092296b8983c43eadf2e66c7c1e0d518555
Reviewed-on: https://go-review.googlesource.com/c/go/+/259448
Trust: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Currently, runtime.call16 is defined and used only on 32-bit
architectures, while 64-bit architectures all start at call32 and go
up from there. This led to unnecessary complexity because call16's
prototype needed to be in a different file, separate from all of the
other call* prototypes, which in turn led to it getting out of sync
with the other call* prototypes. This CL adds call16 on 64-bit
architectures, bringing them all into sync, and moves the call16
prototype to live with the others.
Prior to CL 31655 (in 2016), call16 couldn't be implemented on 64-bit
architectures because it needed at least four words of argument space
to invoke "callwritebarrier" after copying back the results. CL 31655
changed the way call* invoked the write barrier in preparation for the
hybrid barrier; since the hybrid barrier had to be invoked prior to
copying back results, it needed a different solution that didn't reuse
call*'s stack space. At this point, call16 was no longer a problem on
64-bit, but we never added it. Until now.
Change-Id: Id10ade0e4f75c6ea76afa6229ddaee2b994c27dd
Reviewed-on: https://go-review.googlesource.com/c/go/+/259339
Trust: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The signature of call16 is currently missing the "typ" parameter. This
CL fixes this. This wasn't caught by vet because call16 is defined by
macro expansion (see #17544), and we didn't notice the mismatch with
the other call* functions because call16 is defined only on 32-bit
architectures and lives alone in stubs32.go.
Unfortunately, this means its GC signature is also wrong: the "arg"
parameter is treated as a scalar rather than a pointer, so GC won't
trace it and stack copying won't adjust it. This turns out to matter
in exactly one case right now: on 32-bit architectures (which are the
only architectures where call16 is defined), a stack-allocated defer
of a function with a 16-byte or smaller argument frame including a
non-empty result area can corrupt memory if the deferred function
grows the stack and is invoked during a panic. Whew. All other current
uses of reflectcall pass a heap-allocated "arg" frame (which happens
to be reachable from other stack roots, so tracing isn't a problem).
Curiously, in 2016, the signatures of all call* functions were wrong
in exactly this way. CL 31654 fixed all of them in stubs.go, but
missed the one in stubs32.go.
Fixes#41795.
Change-Id: I31e3c0df201f79ee5707eeb8dc4ff0d13fc10ada
Reviewed-on: https://go-review.googlesource.com/c/go/+/259338
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This change clarifies the usage of the SET type name suffix. Previously
the documentation was somewhat confusing about where the suffix should
be used, and when used what it applied to. For instance the previous
language could be interpreted such that []exampleSET would be parsed as
a SEQUENCE OF SET, which is incorrect as the SET suffix only applies to
slice types, such as type exampleSET []struct{} which is parsed as a
SET OF SEQUENCE.
Change-Id: I74201d9969f931f69391c236559f66cb460569ec
GitHub-Last-Rev: d0d2ddc587
GitHub-Pull-Request: golang/go#38543
Reviewed-on: https://go-review.googlesource.com/c/go/+/229078
Trust: Roland Shoemaker <roland@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
This change renames mustHeapAlloc to heapAllocReason, and changes it
to return the reason why the argument must escape, so we don't have to
re-deduce it in its callers just to print the escape reason. It also
embeds isSmallMakeSlice body in heapAllocReason, since the former was
only used by the latter, and deletes isSmallMakeSlice.
An outdated TODO to remove smallintconst, which the TODO claimed was
only used in one place, was also removed, since grepping shows we
currently call smallintconst in 11 different places.
Change-Id: I0bd11bf29b92c4126f5bb455877ff73217d5a155
Reviewed-on: https://go-review.googlesource.com/c/go/+/258678
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Handling ast.GenDecls while typechecking is repetitive, and a source of
diffs compared to typechecking using the cmd/compile/internal/syntax
package, which unpacks declaration groups into individual declarations.
Refactor to extract the logic for walking declarations. This introduces
a new AST abstraction: types.decl, which comes at some minor performance
cost. However, if we are to fully abstract the AST we will be paying
this cost anyway, and benchmarking suggests that the cost is negligible.
Change-Id: If73c30c3d08053ccf7bf21ef886f0452fdbf142e
Reviewed-on: https://go-review.googlesource.com/c/go/+/256298
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Trust: Robert Findley <rfindley@google.com>
Previously by default, we extracted modules to a temporary directory,
then renamed it into place. This failed with ERROR_ACCESS_DENIED on
Windows if another process (usually an anti-virus scanner) opened
files in the temporary directory.
Since Go 1.15, users have been able to set
GODEBUG=modcacheunzipinplace=1 to opt into new behavior: we extract
modules at their final location, and we create and later delete a
.partial file to prevent the directory from being used if we crash.
.partial files are recognized by Go 1.14.2 and later.
With this change, the new behavior is the only behavior.
modcacheunzipinplace is no longer recognized.
Fixes#36568
Change-Id: Iff19fca5cd6eaa3597975a69fa05c4cb1b834bd6
Reviewed-on: https://go-review.googlesource.com/c/go/+/258798
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This was useful for debugging failures occurring during make.bash.
The added flush also ensures that any hints in the GOSSAFUNC output
are flushed before fatal exit.
The environment variable GOSSADIR specifies where the SSA html debugging
files should be placed. To avoid collisions, each one is written into
the [package].[functionOrMethod].html, where [package] is the filepath
separator separated package name, function is the function name, and method
is either (*Type).Method, or Type.Method, as appropriate. Directories
are created as necessary to make this work.
Change-Id: I420927426b618b633bb1ffc51cf0f223b8f6d49c
Reviewed-on: https://go-review.googlesource.com/c/go/+/252338
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Currently on 32-bit systems 8-byte fields in a struct have an alignment
of 4 bytes, which means that atomic instructions may fault. This issue
is tracked in #36606.
Our current workaround is to allocate memory and put any such atomically
accessed fields at the beginning of the object. This workaround fails
because the tiny allocator might not align the object right. This case
specifically only happens with 12-byte objects because a type's size is
rounded up to its alignment. So if e.g. we have a type like:
type obj struct {
a uint64
b byte
}
then its size will be 12 bytes, because "a" will require a 4 byte
alignment. This argument may be extended to all objects of size 9-15
bytes.
So, make this workaround work by specifically aligning such objects to 8
bytes on 32-bit systems. This change leaves a TODO to remove the code
once #36606 gets resolved. It also adds a test which will presumably no
longer be necessary (the compiler should enforce the right alignment)
when it gets resolved as well.
Fixes#37262.
Change-Id: I3a34e5b014b3c37ed2e5e75e62d71d8640aa42bc
Reviewed-on: https://go-review.googlesource.com/c/go/+/254057
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
A recent change to improve shifts was generating some
invalid cases when the rule was based on an AND. The
extended mnemonics CLRLSLDI and CLRLSLWI only allow
certain values for the operands and in the mask case
those values were not being checked properly. This
adds a check to those rules to verify that the
'b' and 'n' values used when an AND was part of the rule
have correct values.
There was a bug in some diag messages in asm9. The
message expected 3 values but only provided 2. Those are
corrected here also.
The test/codegen/shift.go was updated to add a few more
cases to check for the case mentioned here.
Some of the comments that mention the order of operands
in these extended mnemonics were wrong and those have been
corrected.
Fixes#41683.
Change-Id: If5bb860acaa5051b9e0cd80784b2868b85898c31
Reviewed-on: https://go-review.googlesource.com/c/go/+/258138
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Paul Murphy <murp@ibm.com>
Reviewed-by: Carlos Eduardo Seo <carlos.seo@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Lynn Boger <laboger@linux.vnet.ibm.com>
Includes a few tweaks to Value.copyOf(a) (make it a no-op for
a self-copy) and new pattern hack "___" (3 underscores) is
like ellipsis, except the replacement doesn't need to have
matching ellipsis/underscores.
Moved the arg-length check in generated pattern-matching code
BEFORE the args are probed, because not all instances of
variable length OpFoo will have all the args mentioned in
some rule for OpFoo, and when that happens, the compiler
panics without the early check.
Change-Id: I66de40672b3794a6427890ff96c805a488d783f4
Reviewed-on: https://go-review.googlesource.com/c/go/+/247537
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This change incorporates the decision that it should be possible to
run call expansion relatively late in the optimization chain, so that
(1) calls themselves can be exposed to useful optimizations
(2) the effect of selectors on aggregates is seen at the rewrite,
so that assignment of parts into registers is less complicated
(at least I hope it works that way).
That means that selectors feeding into SelectN need to be processed,
and Make* feeding into call parameters need to be processed.
This does however require that call expansion run before decompose
builtins.
This doesn't yet handle rewrites of strings, slices, interfaces,
and complex numbers.
Passes run.bash and race.bash
Change-Id: I71ff23d3c491043beb30e926949970c4f63ef1a4
Reviewed-on: https://go-review.googlesource.com/c/go/+/245133
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
TestBenchmark doesn't use the -bench flag, so that it can format custom
output -- the number of checked lines per second. This is a barrier both
to discoverability, and to piping benchmark output into analysis tools
such as benchstat.
Using testing.B.ReportMetric and a bit of manual timing, we can achieve
similar results while conforming to normal benchmark output. Do this,
and rename the test func to BenchmarkCheck (for symmetry with
TestCheck).
Change-Id: Ie8f2259c1ca9e6986f0137287acf8eb2843f96b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/257958
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Robert Findley <rfindley@google.com>
Reviewed-by: Robert Griesemer <gri@golang.org>
2 conflicts, that make sense.
src/cmd/internal/obj/objfile.go
src/cmd/link/internal/loader/loader.go
Change-Id: Ib224e2d248cb568fa1e888af79dd908b2f5e05ff
1.Revise ambiguous comments: "all current buckets" means buckets in hmap.buckets, actually current bucket and all the overflow buckets connected to it are full
2.All the pointer address add use src/runtime/stubs.go:add, keep the code style uniform
Change-Id: Idc7224dbe6c391e1b03bf5d009c3734bc75187ce
Reviewed-on: https://go-review.googlesource.com/c/go/+/257979
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Extends the IA5String encoding restrictions that are currently applied
to name constraints to dNSName, rfc822Name, and
uniformResourceIdentifier elements of the SAN. The utility function
isIA5String is updated to use unicode.MaxASCII rather than utf8.RuneSelf
as it is somewhat more readable.
Certificates that include these badly encoded names do exist, but are
exceedingly rare. zlint and other linters enforce this encoding and
searching censys.io reveals only three currently trusted certificates
with this particular encoding issue.
Fixes#26362
Change-Id: I7a4f3e165a1754e5b4bfaeabc03e01eb7367f3c9
Reviewed-on: https://go-review.googlesource.com/c/go/+/235078
Run-TryBot: Roland Shoemaker <roland@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Roland Shoemaker <roland@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Originally, zeroExtendAndCopy is used to pad src with leading zeros and
copy the padded src into the destination. It is no longer needed after
CL 230397 introduced FillBytes. We can simply use that and remove the
zeroExtendAndCopy function. It is cleaner and reduces some allocation.
In addition, this patch tries to avoid calling hashToInt function in
both Sign and Verify function so some allocation is reduced.
Benchmarks:
name old alloc/op new alloc/op delta
SignP256-8 1.60kB ± 0% 1.49kB ± 0% -7.23% (p=0.000 n=20+20)
SignP384-8 1.74kB ± 0% 1.59kB ± 0% -8.50% (p=0.000 n=20+18)
VerifyP256-8 176B ± 0% 0B -100.00% (p=0.000 n=20+20)
KeyGeneration-8 640B ± 0% 640B ± 0% ~ (all equal)
name old allocs/op new allocs/op delta
SignP256-8 22.0 ± 0% 17.0 ± 0% -22.73% (p=0.000 n=20+20)
SignP384-8 22.0 ± 0% 17.0 ± 0% -22.73% (p=0.000 n=20+20)
VerifyP256-8 7.00 ± 0% 0.00 -100.00% (p=0.000 n=20+20)
KeyGeneration-8 13.0 ± 0% 13.0 ± 0% ~ (all equal)
Change-Id: Ic4c95191eded55deb3420d97db501689f3b173c9
Reviewed-on: https://go-review.googlesource.com/c/go/+/232297
Reviewed-by: Michael Munday <mike.munday@ibm.com>
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Filippo Valsorda <filippo@golang.org>
'go install pkg@version' runs without a main module or a module root
directory. The -modfile flag cannot be used to set the module root
directory or to substitute a different go.mod file.
This error won't be reported if -modfile is set in
GOFLAGS. Unsupported flags there are generally ignored.
For #40276
Change-Id: I0b39b1fa9184c15c6e863b647d43c328710920f4
Reviewed-on: https://go-review.googlesource.com/c/go/+/258297
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
If vdso is disabled, the goruntime calls gettimeofday from vsyscall,
but if vsyscall is disabled too, all golang binaries crash:
SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0xffffffffff600000} ---
killed by SIGSEGV (core dumped) ++
vsyscall doesn't work as it was designed for a long time due to security
reasons and now vsyscall is a little more expensive than real syscalls:
https://github.com/torvalds/linux/commit/5cec93c216db
This patch reworks the code to call syscalls if the vdso library isn't
available.
Change-Id: I16cbf3f49871bea91e26af1f49aa0ae2fbd3a01d
GitHub-Last-Rev: 1d133cd30a
GitHub-Pull-Request: golang/go#41681
Reviewed-on: https://go-review.googlesource.com/c/go/+/257982
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Trust: Michael Pratt <mpratt@google.com>
Similar to how we report initialization loops in initorder.go and type
alias loops in typecheck.go, this CL updates align.go to warn about
invalid recursive types. The code is based on the loop code from
initorder.go, with minimal changes to adapt from detecting
variable/function initialization loops to detecting type declaration
loops.
Thanks to Cuong Manh Le for investigating this, helping come up with
test cases, and exploring solutions.
Fixes#41575
Updates #41669.
Change-Id: Idb2cb8c5e1d645e62900e178fcb50af33e1700a1
Reviewed-on: https://go-review.googlesource.com/c/go/+/258177
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Trust: Matthew Dempsky <mdempsky@google.com>
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
When building a x509 chain the algorithm currently looks for parents
that have a subject key identifier (SKID) that matches the child
authority key identifier (AKID), if it is present, and returns all
matches. If the child doesn't have an AKID, or there are no parents
with matching SKID it will instead return all parents that have a
subject DN matching the child's issuer DN. Prioritizing AKID/SKID
matches over issuer/subject matches means that later in buildChains we
have to throw away any pairs where these DNs do not match. This also
prevents validation when a child has a SKID with two possible parents,
one with matching AKID but mismatching subject DN, and one with a
matching subject but missing AKID. In this case the former will be
chosen and the latter ignored, meaning a valid chain cannot be built.
This change alters how possible parents are chosen. Instead of doing a
two step search it instead only consults the CertPool.byName subject DN
map, avoiding issues where possible parents may be shadowed by parents
that have SKID but bad subject DNs. Additionally it orders the list of
possible parents by the likelihood that they are in fact a match. This
ordering follows this pattern:
* AKID and SKID match
* AKID present, SKID missing / AKID missing, SKID present
* AKID and SKID don't match
In an ideal world this should save a handful of cycles when there are
multiple possible matching parents by prioritizing parents that have
the highest likelihood. This does diverge from past behavior in that
it also means there are cases where _more_ parents will be considered
than in the past. Another version of this change could just retain the
past behavior, and only consider parents where both the subject and
issuer DNs match, and if both parent and child have SKID and AKID also
compare those, without any prioritization of the candidate parents.
This change removes an existing test case as it assumes that the
CertPool will return a possible candidate where the issuer/subject DNs
do not match.
Fixes#30079
Change-Id: I629f579cabb0b3d0c8cae5ad0429cc5a536b3e58
Reviewed-on: https://go-review.googlesource.com/c/go/+/232993
Trust: Roland Shoemaker <roland@golang.org>
Run-TryBot: Roland Shoemaker <roland@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Currently, in the linker's deadcode pass, when an interface type
is live, the linker thinks all its methods are live, and uses
them to match methods on concrete types. The interface method may
never be used, though.
This CL changes it to only keep used interface methods, for
matching concrete type methods. To do that, when an interface
method is used, the compiler generates a mark relocation. The
linker uses the marker relocations to mark used interface
methods, and only the used ones.
binary size before after
cmd/compile 18887400 18812200
cmd/go 13470652 13470492
Change-Id: I3cfd9df4a53783330ba87735853f2a0ec3c42802
Reviewed-on: https://go-review.googlesource.com/c/go/+/256798
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
In the method table, the method's code pointer is stored as an
offset from the start of the text section. Currently, for an
unreachable method, the offset is left as 0, which resolves to
the start of the text section at run time. It is possible that
there is valid code there. If an unreachable method is ever
reached (due to a compiler or linker bug), the execution will
jump to a wrong location but may continue to run for a while,
until it fails with a seemingly unrelated error.
This CL changes it to use -1 for unreachable method instead. At
run time this will resolve to an invalid address, which makes it
fail immediately if it is ever reached.
Change-Id: Ied6ed7f1833c4f3b991fdf55d8810d70d307b2e6
Reviewed-on: https://go-review.googlesource.com/c/go/+/257203
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
A test that checks if "tls.(*Conn)" appears in any symbol's name.
tls.Conn is a type, so the string "tls.(*Conn)" can only appear
in the name of a method of Conn. But the test code doesn't use
any of the methods. Not sure why this needs to be live. In
particular, the linker is now able to prune all methods of Conn.
Remove this requirement. In fact, just drop the only_conn test
case, as simply allocating a type doesn't necessarily bring
anything live.
Change-Id: I754291b75d38e1465b5291b4dea20806615d21b3
Reviewed-on: https://go-review.googlesource.com/c/go/+/257973
Trust: Cherry Zhang <cherryyz@google.com>
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The function takes five 24-bit chunks from the hash, resulting in 120
bits. When base-64 encoded, this results in a 20-byte output string,
which is confirmed by "var dst [chunks * 4]byte".
It seems like the documented math could have been written for a previous
implementation with shorter hashes, using 4 chunks instead of 5, as then
the math checks out.
Since this code has been working correctly for over three years, let's
fix the documentation to reflect the code.
Change-Id: I9e908e6bafb5dcc1e3c23915e2b6c8843ed444d6
Reviewed-on: https://go-review.googlesource.com/c/go/+/257646
Trust: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The linker prunes methods that are not directly reachable if the
receiver type is never converted to interface. Currently, this
"never" is too strong: it is invalidated even if the interface
conversion is in an unreachable function. This CL improves it by
only considering interface conversions in reachable code. To do
that, we introduce a marker relocation R_USEIFACE, which marks
the target symbol as UsedInIface if the source symbol is reached.
binary size before after
cmd/compile 18897528 18887400
cmd/go 13607372 13470652
Change-Id: I66c6b69eeff9ae02d84d2e6f2bc7f1b29dd53910
Reviewed-on: https://go-review.googlesource.com/c/go/+/256797
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
As part of type checking make's arguments, we were converting untyped
float and complex constant arguments to integers. However, we were
doing this without concern for whether the argument was a declared
constant. Thus a call like "make([]T, n)" could change n from an
untyped float or untyped complex to an untyped integer.
The fix here is to simply change checkmake to not call SetVal, which
will be handled by defaultlit anyway. However, we also need to
properly return the defaultlit result value to the caller, so
checkmake's *Node parameter is also changed to **Node.
Fixes#41680.
Change-Id: I858927a052f384ec38684570d37b10a6906961f7
Reviewed-on: https://go-review.googlesource.com/c/go/+/257966
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
When explaining why the slice from a make() call escapes for the -m -m
message, we print "non-const size" if any one of Isconst(n.Left) and
Isconst(n.Right) return false; but for OMAKESLICE nodes with no cap,
n.Right is nil, so Isconst(n.Right, CTINT) will be always false.
Only call Isconst on n.Right if it's not nil.
Fixes#41635
Change-Id: I8729801a9b234b68ae40adad64d66fa7653adf09
Reviewed-on: https://go-review.googlesource.com/c/go/+/257641
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
ENOTSUP was used as a signaling error in the custom implementation of
syscall.Getwd to fall back to the slow algorithm. Since CL 257637 Getwd
directly calls the respective function from libSystem.dylib which can no
longer return ENOTSUP.
Change-Id: I8e65e42b3ea069bf78969a29f2af1c55552e2949
Reviewed-on: https://go-review.googlesource.com/c/go/+/257644
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
This CL revises the document of File.Fd that explicitly points
its user to runtime.SetFinalizer where contains the information
that a file descriptor could be closed in a finalizer and therefore
causes a failure in syscall.Write if runtime.KeepAlive is not invoked.
The CL also suggests an alternative of File.Fd towards File.SyscallConn.
Fixes#41505
Change-Id: I6816f0157add48b649bf1fb793cf19dcea6894b5
Reviewed-on: https://go-review.googlesource.com/c/go/+/256899
Reviewed-by: Rob Pike <r@golang.org>
Trust: Ian Lance Taylor <iant@golang.org>
Directly wrap the getcwd implementation provided by libSystem.dylib on
darwin and use it to implement Getwd like on the BSDs. This allows to
drop the custom implementation using getAttrList and to merge the
implementation of Getwd for darwin and the BSDs in syscall_bsd.go.
Same as CL 257497 did for golang.org/x/sys/unix
Change-Id: If30390c4c17cd463bb8fdcb5465f40d6fa11f391
Reviewed-on: https://go-review.googlesource.com/c/go/+/257637
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Pick up GOOS=ios changes.
This is done by
cd $GOROOT/src
go get -d golang.org/x/net@latest
go mod tidy
go mod vendor
go get -d golang.org/x/sys@latest
go mod tidy
go mod vendor
cd $GOROOT/src/cmd
go get -d golang.org/x/sys@latest
go mod tidy
go mod vendor
Updates #38485.
Change-Id: Ic2b54febb1f851814c9d76c4b55a8837ac4779f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/257618
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
acquireLockRank and releaseLockRank are called from nosplit context, and
thus must be nosplit.
lockWithRank, unlockWithRank, and lockWithRankMayAcquire are called from
spittable context, and thus don't strictly need to be nosplit.
The stated reasoning for making these functions nosplit is to avoid
re-entrant calls due to a stack split on function entry taking a lock.
There are two potential issues at play here:
1. A stack split on function entry adds a new lock ordering edge before
we (a) take lock l, or (b) release lock l.
2. A stack split in a child call (such as to lock2) introduces a new
lock ordering edge _in the wrong order_ because e.g., in the case of
lockWithRank, we've noted that l is taken, but the stack split in
lock2 actually takes stack split locks _before_ l is actually locked.
(1) is indeed avoided by marking these functions nosplit, but this is
really just a bit of duct tape that generally has no effect overall. Any
earlier call can have a stack split and introduce the same new edge.
This includes lock/unlock which are not nosplit!
I began this CL as a change to extend nosplit to lock and unlock to try
to make this mitigation more effective, but I've realized that as long
as there is a _single_ nosplit call between a lock and unlock, we can
end up with the edge. There seems to be few enough cases without any
calls that is does not seem worth the extra cognitive load to extend
nosplit throughout all of the locking functions.
(2) is a real issue which would cause incorrect ordering, but it is
already handled by switching to the system stack before recording the
lock ordering. Adding / removing nosplit has no effect on this issue.
Change-Id: I94fbd21b2bf928dbf1bf71aabb6788fc0a012829
Reviewed-on: https://go-review.googlesource.com/c/go/+/254367
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
Trust: Michael Pratt <mpratt@google.com>
The rules for go:notinheap were recently tweaked to disallow stack
allocation (CL 249917). This CL updates the documentation about
go:notinheap in runtime/HACKING.md.
Change-Id: Ibca5d9b9d02e1c22c6af1d303aa84c6303a86d92
Reviewed-on: https://go-review.googlesource.com/c/go/+/257357
Trust: Austin Clements <austin@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
We grow the backing store on append by 2x for small sizes and 1.25x
for large sizes. The threshold we use for picking the growth factor
used to depend on the old length, not the old capacity. That's kind of
unfortunate, because then doing append(s, 0, 0) and append(append(s,
0), 0) do different things. (If s has one more spot available, then
the former expression chooses its growth based on len(s) and the
latter on len(s)+1.) If we instead use the old capacity, we get more
consistent behavior. (Both expressions use len(s)+1 == cap(s) to
decide.)
Fixes#41239
Change-Id: I40686471d256edd72ec92aef973a89b52e235d4b
Reviewed-on: https://go-review.googlesource.com/c/go/+/257338
Trust: Keith Randall <khr@golang.org>
Trust: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
The CL 249758 added `FMOVQ $vcon, Vd` instruction and assembler used
128-bit simd literal-loading to load `$vcon` from pool into 128-bit vector
register `Vd`. Because Go does not have 128-bit integers for now, the
assembler will report an error of `immediate out of range` when
assembleing `FMOVQ $0x123456789abcdef0123456789abcdef, V0` instruction.
This patch lets 128-bit integers take two 64-bit operands, for the high
and low parts separately and adds `VMOVQ $hi, $lo, Vd` instruction to
move `$hi<<64+$lo' into 128-bit register `Vd`.
In addition, this patch renames `FMOVQ/FMOVD/FMOVS` ops to 'VMOVQ/VMOVD/VMOVS'
and uses them to move 128-bit, 64-bit and 32-bit constants into vector
registers, respectively
Update the go doc.
Fixes#40725
Change-Id: Ia3c83bb6463f104d2bee960905053a97299e0a3a
Reviewed-on: https://go-review.googlesource.com/c/go/+/255900
Trust: fannie zhang <Fannie.Zhang@arm.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The 387 port needs to load a floating-point control word from a
global location to implement float32 arithmetic.
When compiling with -pie, loading that control word clobbers an
integer register. If that register had something important in it, boom.
Fix by using LEAL to materialize the address of the global location
first. LEAL with -pie works because the destination register is
used as the scratch register.
387 support is about to go away (#40255), so this will need to be
backported to have any effect.
No test. I have one, but it requires building with -pie, which
requires cgo. Our testing infrastructure doesn't make that easy.
Not worth it for a port which is about to vanish.
Fixes#41503
Change-Id: I140f9fc8fdce4e74a52c2c046e2bd30ae476d295
Reviewed-on: https://go-review.googlesource.com/c/go/+/257277
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Keith Randall <khr@golang.org>
In CL 255718 the -S=2 assembly output was enhanced to dump symbol
ABIs. This patch fixes a bug in that CL: when dumping the relocations
on a symbol, we were dumping the symbol's ABI as opposed to the
relocation target symbol's ABI.
Change-Id: I134128687757f549fa37b998cff1290765889140
Reviewed-on: https://go-review.googlesource.com/c/go/+/257202
Trust: Than McIntosh <thanm@google.com>
Run-TryBot: Than McIntosh <thanm@google.com>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Other test flags passed to the test binary, such as -run or -count, are
equally pointless when -c or -i are used, since the test binary is never
run. However, custom flags in that scenario are far more likely to be
due to human error, such as:
# note the "ldflags" typo, which silently did nothing
go test -c -lflags=-w
Instead, make this scenario error. It seems unlikely that anyone is
using -c along with intended custom-defined test flags, and if they are,
removing those extra flags that do nothing is probably a good idea
anyway.
We don't add this restriction for the flags defined in 'go help
testflag', since they are far less likely to be typos or unintended
mistakes. Another reason not to do that change is that other commands
similarly silently ignore no-op flags, such as:
# -d disables the build, so -ldflags is never used
go get -d -ldflags=-w
Fixes#39484.
Change-Id: I6ba2f6866562fe8f8fceaf4cd862d874bf5cd978
Reviewed-on: https://go-review.googlesource.com/c/go/+/237697
Trust: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
If a signal is sent to t.signal before the panic is triggered,
a panicking test may end up with "warning: no tests to run" because
the tRunner that invokes the test in t.Run calls runtime.Goexit on
panic, which causes the panicking test not be recorded in runTests.
Send the signal if and only if there is no panic.
Fixes#41479
Change-Id: I812f1303bfe02c443a1902732e68d21620d6672e
Reviewed-on: https://go-review.googlesource.com/c/go/+/256098
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
Trust: Bryan C. Mills <bcmills@google.com>
CL 250357 exported net.ErrClosed to allow more reliable detection
of closed network connection errors. Use that error in crypto/tls
as well.
The error message is changed from "tls: use of closed connection"
to "use of closed network connection", so the code that detected such
errors by looking for that text in the error message will need to be
updated to use errors.Is(err, net.ErrClosed) instead.
Fixes#41066
Change-Id: Ic05c0ed6a4f57af2a0302d53b00851a59200be2e
Reviewed-on: https://go-review.googlesource.com/c/go/+/256897
Reviewed-by: Katie Hockman <katie@golang.org>
Trust: Katie Hockman <katie@golang.org>
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Katie Hockman <katie@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
The IndexString implementation in the bytealg package requires that
the string passed into it be in the range '2 <= len(s) <= MaxLen'
where MaxLen may be any value (including 0).
CL 156998 added calls to bytealg.IndexString where MaxLen was not
first checked. This led to an illegal instruction on s390x with
the vector facility disabled.
This CL guards the calls to bytealg.IndexString with a MaxLen check.
If the check fails then the code now falls back to the pre CL 156998
implementation (a loop over the runes in the string).
Since the MaxLen check is now in place the generic implementation is
no longer called so I have returned it to its original unimplemented
state.
In future we may want to drop MaxLen to prevent this kind of
confusion.
Fixes#41552.
Change-Id: Ibeb3f08720444a05c08d719ed97f6cef2423bbe9
Reviewed-on: https://go-review.googlesource.com/c/go/+/256717
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Munday <mike.munday@ibm.com>
Reviewed-by: Keith Randall <khr@golang.org>
Introduce GOOS=ios for iOS systems. GOOS=ios matches "darwin"
build tag, like GOOS=android matches "linux" and GOOS=illumos
matches "solaris". Only ios/arm64 is supported (ios/amd64 is
not).
GOOS=ios and GOOS=darwin remain essentially the same at this
point. They will diverge at later time, to differentiate macOS
and iOS.
Uses of GOOS=="darwin" are changed to (GOOS=="darwin" || GOOS=="ios"),
except if it clearly means macOS (e.g. GOOS=="darwin" && GOARCH=="amd64"),
it remains GOOS=="darwin".
Updates #38485.
Change-Id: I4faacdc1008f42434599efb3c3ad90763a83b67c
Reviewed-on: https://go-review.googlesource.com/c/go/+/254740
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
QueryPackage was a wrapper around QueryPattern with extra validation,
called only once from within the same package. Most of that validation
was already performed much earlier, in (*loader).Load. Inline the
remaining validation and remove the needless indirection.
For #36460
Change-Id: I108a01d416197db8f886889554e07b29f0c37f3f
Reviewed-on: https://go-review.googlesource.com/c/go/+/256057
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This change replaces ImportPaths, ImportPathsQuiet, LoadALL, and
LoadVendor with a single LoadPackages function, with a LoadOpts struct
that more clearly documents the variations in behavior.
It also eliminates the cmd/go/internal/load.ImportPaths function,
which was undocumented and had only one call site (within its own
package).
The modload.LoadTests global variable is subsumed by a field in the
new LoadOpts struct, and is no longer needed for callers that invoke
LoadPackages directly. It has been (temporarily) replaced with a
similar global variable, load.ModResolveTests, which can itself be
converted to an explicit, local argument.
For #37438
For #36460
Updates #40775Fixes#26977
Change-Id: I4fb6086c01b04de829d98875db19cf0118d40f8c
Reviewed-on: https://go-review.googlesource.com/c/go/+/255938
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Functions that require holding sched.lock now have an assertion.
A few places with missing locks have been fixed in this CL:
Additionally, locking is added around the call to procresize in
schedinit. This doesn't technically need a lock since the program is
still starting (thus no concurrency) when this is called, but lock held
checking doesn't know that.
Updates #40677
Change-Id: I198d3cbaa727f7088e4d55ba8fa989cf1ee8f9cf
Reviewed-on: https://go-review.googlesource.com/c/go/+/250261
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
When lock ranking is enabled, we can now assert that lock preconditions
are met by checking that the caller holds required locks on function
entry.
This change adds the infrastructure to add assertions. Actual assertions
will be added for various locks in subsequent changes.
Some functions are protected by locks that are not directly accessible
in the function. In that case, we can use assertRankHeld to check that
any lock with the rank is held. This is less precise, but it avoids
requiring passing the lock into the functions.
Updates #40677
Change-Id: I843c6874867f975e90a063f087b6e2ffc147877b
Reviewed-on: https://go-review.googlesource.com/c/go/+/245484
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
The Go 1.16 development cycle has started. This is the time to update
all golang.org/x/... module versions that contribute packages to the
std and cmd modules in the standard library to latest master versions.
Those versions have already gone through code review, and now they
will undergo additional testing during the development period.
If there are new issues in these dependencies discovered, we have
development period to deal with that. We will do this update once
more at the end of the development cycle, by the code freeze, and
so doing it now will make that update smaller and safer.
Overall, this change will help us build confidence that the
Go 1.16 release and its selected dependencies will be robust.
Also increment the Go language version to 1.16 in standard library
go.mod files.
This change was created with a program from CL 256357 patch set 1
(which updates golang.org/x modules only) and the bundle tool at
CL 255053 patch set 1:
$ updatestd -goroot=$HOME/gotip -branch=master
> go version
go version devel +eda1d40544 Mon Sep 21 16:50:07 2020 +0000 darwin/amd64
> go env GOROOT
/Users/dmitshur/gotip
> go version -m /Users/dmitshur/go/bin/bundle
/Users/dmitshur/go/bin/bundle: go1.15.2
path golang.org/x/tools/cmd/bundle
mod golang.org/x/tools (devel) # CL 255053 PS 1
dep golang.org/x/mod v0.3.0 h1:RM4zey1++hCTbCVQfnWeKs9/IEsaBLA8vTkd0WVtmH4=
dep golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1 h1:go1bK/D/BFZV2I8cIQd1NKEZ+0owSTG1fDTci4IqFcE=
updating module cmd in /Users/dmitshur/gotip/src/cmd
skipping github.com/chzyer/logex (out of scope, it's not a golang.org/x dependency)
skipping github.com/chzyer/readline (out of scope, it's not a golang.org/x dependency)
skipping github.com/chzyer/test (out of scope, it's not a golang.org/x dependency)
skipping github.com/google/pprof (out of scope, it's not a golang.org/x dependency)
skipping github.com/ianlancetaylor/demangle (out of scope, it's not a golang.org/x dependency)
skipping github.com/yuin/goldmark (out of scope, it's not a golang.org/x dependency)
skipping rsc.io/pdf (out of scope, it's not a golang.org/x dependency)
> go mod edit -go=1.16
> go get -d golang.org/x/arch@b19915210f009e139b20abfd6a6052c7acc1f445 golang.org/x/crypto@5c72a883971a4325f8c62bf07b6d38c20ea47a6a golang.org/x/mod@ce943fd02449f621243c9ea6e64098e84752b92b golang.org/x/net@62affa334b73ec65ed44a326519ac12c421905e3 golang.org/x/sync@6e8e738ad208923de99951fe0b48239bfd864f28 golang.org/x/sys@af09f7315aff1cbc48fb21d21aa55d67b4f914c5 golang.org/x/text@a8b4671254579a87fadf9f7fa577dc7368e9d009 golang.org/x/tools@d647fc2532668b2b75a92f468487b8085e6ed58b golang.org/x/xerrors@5ec99f83aff198f5fbd629d6c8d8eb38a04218ca
go: golang.org/x/sys af09f7315aff1cbc48fb21d21aa55d67b4f914c5 => v0.0.0-20200918174421-af09f7315aff
go: golang.org/x/text a8b4671254579a87fadf9f7fa577dc7368e9d009 => v0.3.4-0.20200826142016-a8b467125457
go: golang.org/x/tools d647fc2532 => v0.0.0-20200918232735-d647fc253266
go: golang.org/x/net 62affa334b73ec65ed44a326519ac12c421905e3 => v0.0.0-20200904194848-62affa334b73
go: golang.org/x/crypto 5c72a883971a4325f8c62bf07b6d38c20ea47a6a => v0.0.0-20200820211705-5c72a883971a
go: golang.org/x/arch b19915210f009e139b20abfd6a6052c7acc1f445 => v0.0.0-20200826200359-b19915210f00
go: golang.org/x/xerrors 5ec99f83aff198f5fbd629d6c8d8eb38a04218ca => v0.0.0-20200804184101-5ec99f83aff1
> go mod tidy
> go mod vendor
updating module std in /Users/dmitshur/gotip/src
> go mod edit -go=1.16
> go get -d golang.org/x/crypto@5c72a883971a4325f8c62bf07b6d38c20ea47a6a golang.org/x/net@62affa334b73ec65ed44a326519ac12c421905e3 golang.org/x/sys@af09f7315aff1cbc48fb21d21aa55d67b4f914c5 golang.org/x/text@a8b4671254579a87fadf9f7fa577dc7368e9d009 golang.org/x/tools@d647fc2532668b2b75a92f468487b8085e6ed58b
go: golang.org/x/crypto 5c72a883971a4325f8c62bf07b6d38c20ea47a6a => v0.0.0-20200820211705-5c72a883971a
go: golang.org/x/text a8b4671254579a87fadf9f7fa577dc7368e9d009 => v0.3.4-0.20200826142016-a8b467125457
go: golang.org/x/sys af09f7315aff1cbc48fb21d21aa55d67b4f914c5 => v0.0.0-20200918174421-af09f7315aff
go: golang.org/x/tools d647fc2532 => v0.0.0-20200918232735-d647fc253266
> go mod tidy
> go mod vendor
updating bundles in /Users/dmitshur/gotip/src
> go generate -run=bundle std cmd
Other non-golang.org/x module dependencies (pprof and demangle)
still need to be updated in a future CL.
For #36905.
Change-Id: I83a350bf8714ebc249284c0d69abe4941700565e
Reviewed-on: https://go-review.googlesource.com/c/go/+/255860
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
For a common pattern of iterating only over top-level compilation units (CU)
Reader.SkipChildren has decode and meterialize all CU subentries just
to skip them, because DW_TAG_compile_unit does not have DW_AT_sibling.
However, CUs have total size encoded before the unit and we already parse them
and know all unit sizes.
Optimize Reader.SkipChildren to use that size when skipping CUs children.
This speeds up iteration over a 1.3GB object file from 7.5s to 0.73s.
Change-Id: I2a8f00955159b4bd13571409f4817805f934cb69
Reviewed-on: https://go-review.googlesource.com/c/go/+/256217
Run-TryBot: Dmitry Vyukov <dvyukov@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Than McIntosh <thanm@google.com>
Currently, if make.bash is run with a GOOS or GOARCH that the
bootstrap toolchain doesn't support, it will print an ominous but
harmless warning like:
2020/09/21 09:05:27 unsupported GOARCH arm64
This comes from the invocation of "go version" to get the exact
bootstrap toolchain version.
Since the GOOS and GOARCH don't matter for this purpose, this CL
simply clears them on the invocation of the bootstrap toolchain's "go
version".
Fixes#41525.
Change-Id: I17d44eaafed9999b9fa7dcb9fb100b5fd5e554d0
Reviewed-on: https://go-review.googlesource.com/c/go/+/256297
Trust: Austin Clements <austin@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
unlockf is called after the G is put into _Gwaiting, meaning another G
may have readied this one before unlockf is called.
This is implied by the current doc, but add additional notes to call out
this behavior, as it can be quite surprising.
Updates #40641
Change-Id: I60b1ccc6a4dd9ced8ad2aa1f729cb2e973100b59
Reviewed-on: https://go-review.googlesource.com/c/go/+/256058
Trust: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Currently activeStackChans is set before a goroutine blocks on a channel
operation in an unlockf passed to gopark. The trouble is that the
unlockf is called *after* the G's status is changed, and the G's status
is what is used by a concurrent mark worker (calling suspendG) to
determine that a G has successfully been suspended. In this window
between the status change and unlockf, the mark worker could try to
shrink the G's stack, and in particular observe that activeStackChans is
false. This observation will cause the mark worker to *not* synchronize
with concurrent channel operations when it should, and so updating
pointers in the sudog for the blocked goroutine (which may point to the
goroutine's stack) races with channel operations which may also
manipulate the pointer (read it, dereference it, update it, etc.).
Fix the problem by adding a new atomically-updated flag to the g struct
called parkingOnChan, which is non-zero in the race window above. Then,
in isShrinkStackSafe, check if parkingOnChan is zero. The race is
resolved like so:
* Blocking G sets parkingOnChan, then changes status in gopark.
* Mark worker successfully suspends blocking G.
* If the mark worker observes parkingOnChan is non-zero when checking
isShrinkStackSafe, then it's not safe to shrink (we're in the race
window).
* If the mark worker observes parkingOnChan as zero, then because
the mark worker observed the G status change, it can be sure that
gopark's unlockf completed, and gp.activeStackChans will be correct.
The risk of this change is low, since although it reduces the number of
places that stack shrinking is allowed, the window here is incredibly
small. Essentially, every place that it might crash now is replaced with
no shrink.
This change adds a test, but the race window is so small that it's hard
to trigger without a well-placed sleep in park_m. Also, this change
fixes stackGrowRecursive in proc_test.go to actually allocate a 128-byte
stack frame. It turns out the compiler was destructuring the "pad" field
and only allocating one uint64 on the stack.
Fixes#40641.
Change-Id: I7dfbe7d460f6972b8956116b137bc13bc24464e8
Reviewed-on: https://go-review.googlesource.com/c/go/+/247050
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Trust: Michael Knyszek <mknyszek@google.com>
In the current implementation, we can observe crashes after calling
debug.SetMaxStack and allocating a stack larger than 4GB since
stackalloc works with 32-bit sizes. To avoid this, we define an upper
limit as the largest feasible point we can grow a stack to and provide a
better error message when we get a stack overflow.
Fixes#41228
Change-Id: I55fb0a824f47ed9fb1fcc2445a4dfd57da9ef8d4
Reviewed-on: https://go-review.googlesource.com/c/go/+/255997
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Giovanni Bajo <rasky@develer.com>
Reviewed-by: Keith Randall <khr@golang.org>
According to POSIX, there are three formats for TZ variable. When
it refers to timezone file, it should starts with a colon. This commit
removes the colon if it exists, so that it keeps compatible with both
the spec and the old behavior.
Change-Id: I30cfeaea530d24e174de309952338cb1146694a5
GitHub-Last-Rev: 11d83d11ca
GitHub-Pull-Request: golang/go#27570
Reviewed-on: https://go-review.googlesource.com/c/go/+/134217
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
In CL 231397, we stopped marking symbols' GoType reachable in
general, but not when -linkshared. It was left as a TODO. This CL
addresses it.
The problem was that the type names are mangled in the shared
library, so we need to mangle the name consistently in the
executable as well (regardless of whether the symbol is reachable
or not), so that the GCProg generation code can find the
corresponding symbol from the shared library.
Change-Id: I1040747402929a983ec581109f1681a77893682e
Reviewed-on: https://go-review.googlesource.com/c/go/+/255964
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
When we're building a panic that's triggered by a memory fault when
SetPanicOnFault has been called, include an Addr method. This
method reports the address at which the fault occurred.
Fixes#37023
RELNOTE=yes
Change-Id: Idff144587d6b75070fdc861a36efec76f4ec7384
Reviewed-on: https://go-review.googlesource.com/c/go/+/249677
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Keith Randall <khr@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
This is testing for the limited case of zero or one SSA-able results.
One regression is that the later expansion of "Dereference" into
Move into argument slots thwarts the
MOVE A -> B, MOVE B -> C
replaced-by
MOVE A -> B, MOVE A -> C
optimization; the second move is written instead as a Dereference at the
phase where the optimization occurs, and because the target of the
dereference is not visible in the dereference, it's not possible to verify
that A and B or A and C don't overlap in some peculiar way (and for results
fed to args, they can).
Regression is repaired in a later CL by changing when calls are expanded.
Change-Id: Ia0f48a9d483d5a54a99657a24b12b25b8edde55f
Reviewed-on: https://go-review.googlesource.com/c/go/+/242782
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
When we load module retractions from the latest version of a module,
it's possible that the latest version has a different module path than
the version we're loading. This can happen when a module is renamed or
a go.mod file is added for the first time.
We should not report an error in this case. Retractions should still
apply to old aliases of a module.
Fixes#41350
Change-Id: If1bc0b6b2b26fc7023e02fc211aa0cd8eb00796e
Reviewed-on: https://go-review.googlesource.com/c/go/+/255961
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
The linker prunes methods that are not directly reachable if the
receiver type is never converted to interface. A type can be
converted to interface using reflection through other types.
The linker already takes this into consideration but it missed
the case that the intermediate is a method descriptor. Handle
this case.
Change-Id: I590efc5da163c326db8d43583908a2ef67f65d9d
Reviewed-on: https://go-review.googlesource.com/c/go/+/255858
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
retsigerr was used to create error message for both wrong function
arguments and return arguments, so change its name to sigerr to reflect
that.
While at it, also add documentation for the wrong function arguments
case.
Passes toolstash-check.
Change-Id: I740c717ad38d4afab9e8c20f2e94579c8bca67ff
Reviewed-on: https://go-review.googlesource.com/c/go/+/255240
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
The revised test now checks that unsafe-uintptr correctly works for
variadic uintptr parameters too, and the CL corrects the code so this
code compiles again.
The pointers are still not kept alive properly. That will be fixed by
a followup CL. But this CL at least allows programs not affected by
that to build again.
Updates #24991.
Updates #41460.
Change-Id: If4c39167b6055e602213fb7522c4f527c43ebda9
Reviewed-on: https://go-review.googlesource.com/c/go/+/255877
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
When -S=2 is in effect for the compiler/assembler, include symbol ABI
values for defined symbols and relocations. This is intended to help
make it easier to distinguish between a symbol and its ABI wrapper.
Change-Id: Ifbf71372392075f15363b40e882b2132406b7d6d
Reviewed-on: https://go-review.googlesource.com/c/go/+/255718
Trust: Than McIntosh <thanm@google.com>
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: David Chase <drchase@google.com>
Previously, modload imported load, but it mainly just did so in order
to install callbacks to the modload API. This was important during vgo
development, but there's no longer a strong reason to do this. Nothing
modload imports strongly depends on load, so there's little danger of
a dependency cycle.
This change deletes the callbacks in load and instead, makes load call
exported functions in modload directly. In the future, these functions
may have different signatures than their GOPATH counterparts.
Change-Id: Ifde5c3ffebd190b5bd184924ec447d272b936f27
Reviewed-on: https://go-review.googlesource.com/c/go/+/255719
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
//go:notinheap
type T int
type U T
We already correctly propagate the notinheap-ness of T to U. But we
have an assertion in the typechecker that if there's no explicit
//go:notinheap associated with U, then report an error. Get rid of
that error so that implicit propagation is allowed.
Adjust the tests so that we make sure that uses of types like U
do correctly report an error when U is used in a context that might
cause a Go heap allocation.
Fixes#41451
Update #40954
Update #41432
Change-Id: I1692bc7cceff21ebb3f557f3748812a40887118d
Reviewed-on: https://go-review.googlesource.com/c/go/+/255637
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Set GO111MODULE=off explicitly in tests specific to GOPATH mode.
Added a go.mod file to other tests that assumed GOPATH mode.
Fixed an issue in the build metadata file generated in
modload/build.go, which did not end with a newline. This broke the
build_dash_x test, which expects to be able to run the script printed
by 'go build -x' to produce the same result. The script is broken if
the build metadata file doesn't end with a newline.
Change-Id: I59f2a492a9f5a66f6c4aa702f429909d5c5e815d
Reviewed-on: https://go-review.googlesource.com/c/go/+/255051
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
This change adds rules to find pairs of instructions that can
be combined into a single shifts. These instruction sequences
are common in array addressing within loops. Improvements can
be seen in many crypto packages and the hash packages.
These are based on the extended mnemonics found in the ISA
sections C.8.1 and C.8.2.
Some rules in PPC64.rules were moved because the ordering prevented
some matching.
The following results were generated on power9.
hash/crc32:
CRC32/poly=Koopman/size=40/align=0 195ns ± 0% 163ns ± 0% -16.41%
CRC32/poly=Koopman/size=40/align=1 200ns ± 0% 163ns ± 0% -18.50%
CRC32/poly=Koopman/size=512/align=0 1.98µs ± 0% 1.67µs ± 0% -15.46%
CRC32/poly=Koopman/size=512/align=1 1.98µs ± 0% 1.69µs ± 0% -14.80%
CRC32/poly=Koopman/size=1kB/align=0 3.90µs ± 0% 3.31µs ± 0% -15.27%
CRC32/poly=Koopman/size=1kB/align=1 3.85µs ± 0% 3.31µs ± 0% -14.15%
CRC32/poly=Koopman/size=4kB/align=0 15.3µs ± 0% 13.1µs ± 0% -14.22%
CRC32/poly=Koopman/size=4kB/align=1 15.4µs ± 0% 13.1µs ± 0% -14.79%
CRC32/poly=Koopman/size=32kB/align=0 137µs ± 0% 105µs ± 0% -23.56%
CRC32/poly=Koopman/size=32kB/align=1 137µs ± 0% 105µs ± 0% -23.53%
crypto/rc4:
RC4_128 733ns ± 0% 650ns ± 0% -11.32% (p=1.000 n=1+1)
RC4_1K 5.80µs ± 0% 5.17µs ± 0% -10.89% (p=1.000 n=1+1)
RC4_8K 45.7µs ± 0% 40.8µs ± 0% -10.73% (p=1.000 n=1+1)
crypto/sha1:
Hash8Bytes 635ns ± 0% 613ns ± 0% -3.46% (p=1.000 n=1+1)
Hash320Bytes 2.30µs ± 0% 2.18µs ± 0% -5.38% (p=1.000 n=1+1)
Hash1K 5.88µs ± 0% 5.38µs ± 0% -8.62% (p=1.000 n=1+1)
Hash8K 42.0µs ± 0% 37.9µs ± 0% -9.75% (p=1.000 n=1+1)
There are other improvements found in golang.org/x/crypto which are all in the
range of 5-15%.
Change-Id: I193471fbcf674151ffe2edab212799d9b08dfb8c
Reviewed-on: https://go-review.googlesource.com/c/go/+/252097
Trust: Lynn Boger <laboger@linux.vnet.ibm.com>
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Carlos Eduardo Seo <cseo@linux.vnet.ibm.com>
xtop holds package's top-level declaration statements, but OCLOSURE
only appears in expression contexts. xtop will instead hold the
synthetic ODCLFUNC representing OCLOSURE's function body.
This CL makes the loop consistent with the later phases that only look
for ODCLFUNC nodes in xtop.
Passes toolstash-check.
Change-Id: I852a10ef1bf75bb3351e3da0357ca8b2e26aec6e
Reviewed-on: https://go-review.googlesource.com/c/go/+/255340
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
The existing translation assumes an in-memory return values, thus it returns
the address of the result(s). Most consumers immediately load from the
address to get the value, and in late call expansion that is the favored idiom,
and it is also the favored idiom when arguments and results use registers
instead of memory.
Change-Id: Ie0ccc70f399682a42509d847b330ef3956462d56
Reviewed-on: https://go-review.googlesource.com/c/go/+/240186
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This is prerequisite to moving call expansion later into SSA,
and probably a good idea anyway. Passes tests.
This is the first minimal CL that does a 1-for-1 substitution
of *ssa.AuxCall for *obj.LSym. Next step (next CL) is to make
this change for all calls so that additional information can
be stored in AuxCall.
Change-Id: Ia3a7715648fd9fb1a176850767a726e6f5b959eb
Reviewed-on: https://go-review.googlesource.com/c/go/+/237680
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
"cannot assign to" compiler errors are very laconic: they never
explain why the lhs cannot be assigned to (with one exception, when
assigning to a struct field in a map).
This change makes them a little more specific, in two more cases: when
assigning to a string, or to a const; by giving a very brief reason
why the lhs cannot be assigned to.
Change-Id: I244cca7fc3c3814e00e0ccadeec62f747c293979
Reviewed-on: https://go-review.googlesource.com/c/go/+/255199
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
They can't reasonably be allocated on the heap. Not a huge deal, but
it has an interesting and useful side effect.
After CL 249917, the compiler and runtime treat pointers to
go:notinheap types as uintptrs instead of real pointers (no write
barrier, not processed during stack scanning, ...). That feature is
exactly what we want for cgo to fix#40954. All the cases we have of
pointers declared in C, but which might actually be filled with
non-pointer data, are of this form (JNI's jobject heirarch, Darwin's
CFType heirarchy, ...).
Fixes#40954
Change-Id: I44a3b9bc2513d4287107e39d0cbbd0efd46a3aae
Reviewed-on: https://go-review.googlesource.com/c/go/+/250940
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Keith Randall <khr@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
The alias doesn't need to be marked go:notinheap. It gets its
notinheap-ness from the target type.
Without this change, the type alias test in the notinheap.go file
generates these two errors:
notinheap.go:62: misplaced compiler directive
notinheap.go:63: type nih must be go:notinheap
The first is a result of go:notinheap pragmas not applying
to type alias declarations.
The second is the result of then trying to match the notinheap-ness
of the alias and the target type.
Add a few more go:notinheap tests while we are here.
Update #40954
Change-Id: I067ec47698df6e9e593e080d67796fd05a1d480f
Reviewed-on: https://go-review.googlesource.com/c/go/+/250939
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Trust: Keith Randall <khr@golang.org>
CL 249917 made the mspan in MSpanCountAlloc no longer stack-allocated
(for good reason), but then allocated an mspan on each call and did not
free it, resulting in a leak. That allocation was also not protected by
the heap lock, which could lead to data corruption of mheap fields and
the spanalloc.
To fix this, export some functions to allocate/free dummy mspans from
spanalloc (with proper locking) and allocate just one up-front for the
benchmark, freeing it at the end. Then, update MSpanCountAlloc to accept
a dummy mspan.
Note that we need to allocate the dummy mspan up-front otherwise we
measure things like heap locking and fixalloc performance instead of
what we actually want to measure: how fast we can do a popcount on the
mark bits.
Fixes#41391.
Change-Id: If6629a6ec1ece639c7fb78532045837a8c872c04
Reviewed-on: https://go-review.googlesource.com/c/go/+/255297
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Michael Knyszek <mknyszek@google.com>
Use a common runtime slicecopy function to copy strings or slices
into slices. This deduplicates similar code previously used in
reflect.slicecopy and runtime.stringslicecopy.
Change-Id: I09572ff0647a9e12bb5c6989689ce1c43f16b7f1
Reviewed-on: https://go-review.googlesource.com/c/go/+/254658
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Martin Möhrmann <moehrmann@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
In https://golang.org/cl/221397 we made commands like "go version -v"
error, since both of the command's flags only make sense when arguments
follow them. Without arguments, the command only reports Go's own
version, and the flags are most likely a mistake.
However, the script below is entirely reasonable:
export GOFLAGS=-v # make all Go commands verbose
go version
go build
After the previous CL, "go version" would error. Instead, only error if
the flag was passed explicitly, and not via GOFLAGS.
The patch does mean that we won't error on "GOFLAGS=-v go version -v",
but that very unlikely false negative is okay. The error is only meant
to help the user not misuse the flags, anyway - it's not a critical
error of any sort.
To reuse inGOFLAGS, we move it to the base package and export it there,
since it's where the rest of the GOFLAGS funcs are.
Fixes#41264.
Change-Id: I74003dd25d94bacf9ac507b5cad778fd65233321
Reviewed-on: https://go-review.googlesource.com/c/go/+/254157
Trust: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This makes error reporting a bit more consistent for 'go mod'
subcommands. Most of these commands only work in module mode when a
go.mod file is present.
Setting modload.ForceUseModules reports an error when GO111MODULE=off.
Setting modload.RootMode to modload.NeedRoot reports an error when no
go.mod file is present.
Change-Id: I1daa8d2971cb8658e0c804765839d903734a412e
Reviewed-on: https://go-review.googlesource.com/c/go/+/254369
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
Currently, there's awkward reentrancy issue with funccompile:
funccompile -> compile -> dtypesym -> geneq/genhash/genwrapper -> funccompile
Though it's not a problem at this moment, some attempts by @mdempsky to
move order/walk/instrument into buildssa was failed, due to SSA cache
corruption.
This commit fixes that reentrancy issue, by making generated functions
to be pumped through the same compile workqueue that normal functions
are compiled. We do this by adding them to xtop, instead of calling
funccompile directly in geneq/genhash/genwrapper. In dumpdata, we look
for uncompiled functions in xtop instead of compilequeue, then finish
compiling them.
Updates #38463Fixes#33485
Change-Id: Ic9f0ce45b56ae2ff3862f17fd979253ddc144bb5
Reviewed-on: https://go-review.googlesource.com/c/go/+/254617
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Keith Randall <khr@golang.org>
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
This allows the global initializers function to go through normal
mid-end optimizations (e.g., inlining, escape analysis) like any other
function.
Updates #33485.
Change-Id: I9bcfe98b8628d1aca09b4c238d8d3b74c69010a5
Reviewed-on: https://go-review.googlesource.com/c/go/+/254839
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
funchdr and funcbody currently assume that either (1) Curfn == nil &&
dclcontext == PEXTERN, or (2) Curfn != nil && dclcontext == PAUTO.
This is a reasonable assumption during parsing. However, these
functions end up getting used in other contexts, and not all callers
are so disciplined about Curfn/dclcontext handling.
This CL changes them to save/restore arbitrary Curfn/dclcontext pairs
instead. This is necessary for the followup CL, which pushes fninit
earlier. Otherwise, Curfn/dclcontext fall out of sync, and funchdr
panics.
Passes toolstash-check.
Updates #33485.
Change-Id: I19b1be23db1bad6475345ae5c81bbdc66291a3a7
Reviewed-on: https://go-review.googlesource.com/c/go/+/254838
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
This reverts CL 253037.
Reason for revert: The recommended way to check for a type of error is errors.As. API changes should also start with a proposal.
Change-Id: I62896717aa47ed491c2c4775d2b05d80e5e9cde3
Reviewed-on: https://go-review.googlesource.com/c/go/+/254837
Trust: Damien Neil <dneil@google.com>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
This reverts CL 254537.
Reason for revert: Reason for revert: The recommended way to check for a type of error is errors.As. API changes should also start with a proposal.
Change-Id: I07c37428575e99c80b17525833a61831d10963bb
Reviewed-on: https://go-review.googlesource.com/c/go/+/254857
Trust: Damien Neil <dneil@google.com>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Previously, it was impossible to call FailNow in a Cleanup.
Because it can terminate a panicking goroutine and cause its
parent hangs on t.signal channel. This CL sends the signal
in a deferred call to prevent the hang.
Fixes#41355
Change-Id: I4552d3a7ea763ef86817bf9b50c0e37fb34bf20f
Reviewed-on: https://go-review.googlesource.com/c/go/+/254637
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
As of CL 251159, when 'go list -deps std' is run within GOROOT/src, it
treats the vendored external dependencies as real module dependencies,
not standard-library "vendor/" packages (which still exist in that
case, but are treated as distinct packages outside the "std" module).
Fixes#41358
Updates #30241
Change-Id: Ic23eae9829d90e74a340d49ca9052e9191597410
Reviewed-on: https://go-review.googlesource.com/c/go/+/254738
Run-TryBot: Bryan C. Mills <bcmills@google.com>
Trust: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This CL ensures that responses served via CGI and FastCGI
have a Content-Type header based on the content of the
response if not explicitly set by handlers.
If the implementers of the handler did not explicitly
specify a Content-Type both CGI implementations would default
to "text/html", potentially causing cross-site scripting.
Thanks to RedTeam Pentesting GmbH for reporting this.
Fixes#40928
Fixes CVE-2020-24553
Change-Id: I82cfc396309b5ab2e8d6e9a87eda8ea7e3799473
Reviewed-on: https://team-review.git.corp.google.com/c/golang/go-private/+/823217
Reviewed-by: Russ Cox <rsc@google.com>
Reviewed-on: https://go-review.googlesource.com/c/go/+/252179
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Katie Hockman <katie@golang.org>
CL 96455 brings CL 57753 to Windows
However, a path comparison within it was left unquoted.
If the Go source directory resides in a path containing whitespace,
the interpreter will compare against the first portion of the path string,
and treat the remainder as an expression.
This patch amends that.
For example, consider the path
`C:\Users\Dominic Della Valle\Projects\Go\goroot\src`
Issuing `make.bat` will print out `'Della' is not recognized as an internal or external command, operable program or batch file.` before proceeding.
Change-Id: Ifcec159baeec940c29c61aa721c64c13c6fd8c14
GitHub-Last-Rev: 809ddbb4db
GitHub-Pull-Request: golang/go#41319
Reviewed-on: https://go-review.googlesource.com/c/go/+/253898
Run-TryBot: Giovanni Bajo <rasky@develer.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Giovanni Bajo <rasky@develer.com>
Trust: Giovanni Bajo <rasky@develer.com>
CL 254397 attached OVARLIVE nodes to OCALLxxx nodes Nbody.
The NeedsWrapper flag is now redundant with n.Nbody.Len() > 0
condition, so use that condition instead and remove the flag.
Passes toolstash-check.
Change-Id: Iebc3e674d3c0040a876ca4be05025943d2b4fb31
Reviewed-on: https://go-review.googlesource.com/c/go/+/254398
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
So we can insert theses OVARLIVE nodes right after OpStaticCall in SSA.
This helps fixing issue that unsafe-uintptr arguments are not kept alive
during return statement, or can be kept alive longer than expected.
Fixes#24491
Change-Id: Ic04a5d1bbb5c90dcfae65bd95cdd1da393a66800
Reviewed-on: https://go-review.googlesource.com/c/go/+/254397
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Allows users to check:
errors.Is(err, &UnmarshalTypeError{})
errors.Is(err, &UnmarshalFieldError{})
errors.Is(err, &InvalidUnmarshalError{})
errors.Is(err, &UnsupportedValueError{})
errors.Is(err, &MarshalerError{})
which is the recommended way of checking for kinds of errors.
SyntaxError.Is was implemented in CL 253037.
As and Unwrap relevant methods will be added in future CLs.
Change-Id: I1f8a503b8fdc0f3afdfe9669a91f3af8d960e028
GitHub-Last-Rev: 930cda5384
GitHub-Pull-Request: golang/go#41360
Reviewed-on: https://go-review.googlesource.com/c/go/+/254537
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Trust: Emmanuel Odeke <emm.odeke@gmail.com>
We want to error early if GOFLAGS contains any flag that isn't known to
any cmd/go command. Thus, at init time we would recursively use VisitAll
on each of the flagsets to populate a map of all registered flags.
This was unfortunate, as populating said map constituted a whole 5% of
the run-time of 'go env GOARCH'. This is because VisitAll is pretty
expensive; it copies all the maps from the flagset's map to a slice,
sorts the slice, then does one callback per flag.
First, this was a bit wasteful. We only ever needed to query the
knownFlag map if GOFLAGS wasn't empty. If it's empty, there's no work to
do, thus we can skip the map populating work.
Second and most important, we don't actually need the map at all. A
flag.FlagSet already has a Lookup method, so we can simply recursively
call those methods for each flag in GOFLAGS. Add a hasFlag func to make
that evident.
This mechanism is different; its upfront cost is none, but it will
likely mean a handful of map lookups for each flag in GOFLAGS. However,
that tradeoff is worth it; we don't expect GOFLAGS to contain thousands
of flags. The most likely scenario is less than a dozen flags, in which
case constructing a "unified" map is not at all a net win.
One possible reason the previous mechanism was that way could be
AddKnownFlag. Thankfully, the one and only use of that API was removed
last year when Bryan cleaned up flag parsing in cmd/go.
The wins for the existing benchmark with an empty GOFLAGS are
significant:
name old time/op new time/op delta
ExecGoEnv-8 575µs ± 1% 549µs ± 2% -4.44% (p=0.000 n=7+8)
name old sys-time/op new sys-time/op delta
ExecGoEnv-8 1.69ms ± 1% 1.68ms ± 2% ~ (p=0.281 n=7+8)
name old user-time/op new user-time/op delta
ExecGoEnv-8 1.80ms ± 1% 1.66ms ± 2% -8.09% (p=0.000 n=7+8)
To prove that a relatively large number of GOFLAGS isn't getting
noticeably slower, we measured that as well, via benchcmd and GOFLAGS
containing 50 valid flags:
GOFLAGS=$(yes -- -race | sed 50q) benchcmd -n 500 GoEnvGOFLAGS go env GOARCH
And the result, while noisy, shows no noticeable difference (note that
it measures 3ms instead of 0.6ms since it's sequential):
name old time/op new time/op delta
GoEnvGOFLAGS 3.04ms ±32% 3.03ms ±35% ~ (p=0.156 n=487+481)
Finally, we've improved the existing Go benchmark. Now it's parallel,
and it also reports sys-time and user-time, which are useful metrics.
Change-Id: I9b4551415cedf2f819eb184a02324b8bd919e2bd
Reviewed-on: https://go-review.googlesource.com/c/go/+/248757
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
'go mod download' calls WriteGoMod once via modload.ListModules when
it loads the build list. This saves sums for go.mod files needed by
MVS, but the write occurs before any zip files are downloaded.
With this change, 'go mod download' calls WriteGoMod again (and thus,
modfetch.WriteGoSum) after downloading and verifying module zip files,
so the sums of the zip files will be saved, too.
Fixes#41341
Change-Id: I7d56754aa255256ed45fd93cb154c2e6ea5f45a9
Reviewed-on: https://go-review.googlesource.com/c/go/+/254357
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
On darwin, we put read-only data in __TEXT segment on AMD64 in
exe (non-PIE) buildmode, and in __DATA on everywhere else. This
is not ideal: things in __DATA segment are not read-only, and
being mapped R/W may use more run-time resources.
In fact, newer darwin systems support a __DATA_CONST segment,
which the dynamic linker will map it read-only after applying
relocations. Use that.
Fixes#38830.
Change-Id: Ic281e6c6ca8ef5fec4bb7c5b71c50dd5393e78ae
Reviewed-on: https://go-review.googlesource.com/c/go/+/253919
Reviewed-by: Than McIntosh <thanm@google.com>
When -mod=readonly is set, Import will now allow imports from
replacements without explicit requirements. With -mod=mod, this would
add a new requirement but does not trigger a module lookup, so it's
determinisitic.
Before reporting an error for an unknown import with -mod=readonly,
check whether the import is valid. If there's a typo in the import,
that's more relevant.
For #40728
Change-Id: I05e138ff76ba3d0eb2e3010c15589fa363deb8d3
Reviewed-on: https://go-review.googlesource.com/c/go/+/253745
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Keep track of whether the -mod flag was set explicitly. When
-mod=readonly is the default, we'll want to adjust our error messages
if it's set explicitly.
Also, register the -mod, -modcacherw, and -modfile flags in functions
in internal/base instead of internal/work. 'go mod' commands that
don't load packages shouldn't depend on internal/work.
For #40728
Change-Id: I272aea9e19908ba37e151baac4ea8630e90f241f
Reviewed-on: https://go-review.googlesource.com/c/go/+/253744
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
In the deadcode pass, a type symbol may be marked twice, one
without UsedInIface, one with. For the second time, don't
update the Reachparent graph, so it only records the path of
the first time the symbol is reached. This ensures the
Reachparent graph is acyclic.
TODO: add a test. (This only affects GOEXPERIMENT=fieldtrack)
Change-Id: I68e8a1a69c3830bc8aee5df946151dc22dcb2b29
Reviewed-on: https://go-review.googlesource.com/c/go/+/254297
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
BenchmarkFullASCIIRune tests the performance of function utf8.FullRune,
which will be inlined in BenchmarkFullASCIIRune. Since the return value
of FullRune is not referenced, it will be removed as dead code.
This CL makes the FullRune functions return value referenced by a global
variable to avoid this point. In addition, this CL adds one more benchmark
to cover more code paths, and puts them together as sub benchmarks of
BenchmarkFullRune.
Change-Id: I6e79f4c087adf70e351498a4b58d7482dcd1ec4a
Reviewed-on: https://go-review.googlesource.com/c/go/+/233979
Run-TryBot: eric fang <eric.fang@arm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Enable VBSL, VBIT, VCMTST, VUXTL VUXTL2 and FMOVQ SIMD
instructions required by the issue #40725. And FMOVQ
instrucion is used to move a large constant to a Vn
register.
Add test cases.
Fixes#40725
Change-Id: I1cac1922a0a0165d698a4b73a41f7a5f0a0ad549
Reviewed-on: https://go-review.googlesource.com/c/go/+/249758
Reviewed-by: Cherry Zhang <cherryyz@google.com>
I found the control flow of this function a bit tricky to reason about
due to nesting and interaction between conditions and iteration. This
change factors out a helper function that can return early instead of
mixing conditionals and 'continue' statements.
Also remove the (unused) ModuleUsedDirectly function.
For #36460
Change-Id: I60a2a5a1b32989e5a17a14e1a8c858b280cda8f2
Reviewed-on: https://go-review.googlesource.com/c/go/+/251998
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
With lazy loading, the “build list” can be refined as packages are loaded.
Rename functions that return the build list to more precisely describe
the set of modules returned by the call.
Also eliminate a redundant call to LoadBuildList (right before
ListModules, which itself begins with the same call).
For #36460
Change-Id: I0fc4f9dd7602e0df5e166e329ee5d516d810ca53
Reviewed-on: https://go-review.googlesource.com/c/go/+/249878
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
The new semantics of the "all" package pattern can be implemented
without actually changing module loading per se. This change
implements those semantics, so that the change can be decoupled from
the changes to the module requirement graph.
For #36460
Change-Id: I0ee8b17afa8b728dc470a42a540fcc01764a4442
Reviewed-on: https://go-review.googlesource.com/c/go/+/240623
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
Due to a bug in CL 173017, if QueryPackages found multiple candidates
for the given package and *at least* one of those candidates was not
available to add, we would reject *all* such candidates — even those
that were still viable.
Now, we return the first viable candidate, and only return an error if
*no* candidate is viable given the current build list.
Fixes#41113
Change-Id: Idb2e77244be7c0f5dd511efb142c3059925d7336
Reviewed-on: https://go-review.googlesource.com/c/go/+/251446
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
Fix a data race for clients that mutate requests after receiving a
response error which is caused by the writeLoop goroutine left
running, this can be seen on cancelled requests.
Fixes#37669
Change-Id: Ia4743c6b8abde3a7503de362cc6a3782e19e7f60
Reviewed-on: https://go-review.googlesource.com/c/go/+/251858
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
modload.Import previously performed two otherwise-separable tasks:
1. Identify which module in the build list contains the requested
package.
2. If no such module exists, search available modules to try to find
the missing package.
This change splits those two tasks into two separate unexported
functions, and reports import-resolution errors by attaching them to
the package rather than emitting them directly to stderr. That allows
'list' to report the errors, but 'list -e' to ignore them.
With the two tasks now separate, it will be easier to avoid the
overhead of resolving missing packages during lazy loading if we
discover that some existing dependency needs to be promoted to the top
level (potentially altering the main module's selected versions, and
thus suppling packages that were previously missing).
For #36460
Updates #26909
Change-Id: I32bd853b266d7cd231d1f45f92b0650d95c4bcbd
Reviewed-on: https://go-review.googlesource.com/c/go/+/251445
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
forceStdVendor was a special-case mechanism to allow Go contributors
to use vendored dependencies by default when working in GOROOT/src.
As of Go 1.14,¹ the 'go' command uses vendored dependencies by default
within all modules, so the 'std' and 'cmd' modules no longer need to
be special cases, and we can remove this special-case code.
¹ https://golang.org/doc/go1.14#vendor
Updates #33848
Updates #30241
Change-Id: Ib2fb5841c253113b17fa86a086ce85a22ac3d121
Reviewed-on: https://go-review.googlesource.com/c/go/+/251159
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
If the user requests the 'all' pattern in addition to explicit roots
outside of 'all', we should not load the transitive dependencies of
those explicit roots as if they were *in* 'all'. Without the '-test'
flag, we should not load test dependencies of any package outside of
'all'. Even *with* the '-test' flag, we should only load test
dependencies of the requested roots, not test dependencies of other
packages that happen to be imported by those roots.
More precise tracking of membership in 'all' will be important when we
implement lazy loading, because membership in 'all' determines which
module dependencies we will record in the main module's go.mod file.
This change also reduces reliance on global state, factors out the
loading process into several smaller functions, and sets us up to
reuse the 'go mod vendor' version of the 'all' pattern for lazy
loading.
For #36460Fixes#40799
Change-Id: I5ca21c86a860daee1316f732cea131a331d8ddf9
Reviewed-on: https://go-review.googlesource.com/c/go/+/240505
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
This avoids a deadlock on prof.signalLock between setcpuprofilerate
and cpuprof.add if a SIGPROF is delivered to the thread between the
call to setThreadCPUProfiler and acquiring prof.signalLock.
Fixes#41014
Change-Id: Ie825e8594f93a19fb1a6320ed640f4e631553596
Reviewed-on: https://go-review.googlesource.com/c/go/+/253758
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Both ReadMemStatsSlow and CheckScavengedBits iterate over the page
allocator's chunks but don't actually check if they exist. During the
development process the chunks index became sparse, so now this was a
possibility. If the runtime tests' heap is sparse we might end up
segfaulting in either one of these functions, though this will generally
be very rare.
The pattern here to return nil for a nonexistent chunk is also useful
elsewhere, so this change introduces tryChunkOf which won't throw, but
might return nil. It also updates the documentation of chunkOf.
Fixes#41296.
Change-Id: Id5ae0ca3234480de1724fdf2e3677eeedcf76fa0
Reviewed-on: https://go-review.googlesource.com/c/go/+/253777
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Pre-resolve package index references, so it doesn't need to do a
map lookup in every cross-package symbol reference resolution. It
increases the memory usage very slightly (O(# imported packages)).
Change-Id: Ia76c97ac51f1c2c2d5ea7ae34853850ec69ef0a8
Reviewed-on: https://go-review.googlesource.com/c/go/+/253604
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This preserves zip sums when 'go get' is run on a module that does not
have a package in the root directory. The zip must be fetched to
determine whether the package should be loaded, so we already load and
verify the sum.
Note that 'go mod tidy' may still remove these sums, since they
aren't needed to load packages.
Fixes#41103
Change-Id: I78f10a25f0392461fdc98518a7c92a38ee3233c3
Reviewed-on: https://go-review.googlesource.com/c/go/+/251880
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Currently, the statement:
go g(uintptr(f()))
gets rewritten into:
tmp := f()
newproc(8, g, uintptr(tmp))
runtime.KeepAlive(tmp)
which doesn't guarantee that tmp is still alive by time the g call is
scheduled to run.
This CL fixes the issue, by wrapping g call in a closure:
go func(p unsafe.Pointer) {
g(uintptr(p))
}(f())
then this will be rewritten into:
tmp := f()
go func(p unsafe.Pointer) {
g(uintptr(p))
runtime.KeepAlive(p)
}(tmp)
runtime.KeepAlive(tmp) // superfluous, but harmless
So the unsafe.Pointer p will be kept alive at the time g call runs.
Updates #24491
Change-Id: Ic10821251cbb1b0073daec92b82a866c6ebaf567
Reviewed-on: https://go-review.googlesource.com/c/go/+/253457
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
It appears the machoCalcStart function is meant to align the
segment, but it doesn't. Replace it with an actual alignment
calculation. Also, use the alignment from the configuration,
instead of hardcode.
With this fix we could enable DWARF combining on macOS ARM64.
Change-Id: I19ec771b77d752b83a54c53b6ee65af78a31b8ae
Reviewed-on: https://go-review.googlesource.com/c/go/+/253558
Reviewed-by: Than McIntosh <thanm@google.com>
In Mach-O DWARF combining, some code was written using reflection,
so it could support both 32-bit and 64-bit Mach-O files without
duplicating code. We no longer support 32-bit darwin platforms
now. 32-bit support can go. Rewrite it with direct field access,
for 64-bit only.
Change-Id: If1338c3cd37cecf603f4df0c6eb0c890eaebfe5f
Reviewed-on: https://go-review.googlesource.com/c/go/+/253557
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
On darwin, with external linking, the system linker produces STAB
(symbolic debugging) symbols in the binary's symbol table. These
include paths of the intermediate object files, like
<tmpdir>/go.o, which changes from run to run, making the build
non-reproducible.
Since we run dsymutil to produce debug info and combine them
back into the binary, we don't need those STAB symbols anymore.
Strip them after running dsymutil.
If DWARF is not enabled, we don't run dsymutil. We can pass
"-Wl,-S" to let the system linker not generate those symbols.
While here, also make it more consistent about DWARF combining.
Currently we only do DWARF combining on macOS/AMD64, when DWARF
is enabled. On ARM64, we run dsymutil, but then throw the result
away. This CL changes it to not run dsymutil (and strip) on
ARM64.
TODO: add a test. We don't do it here as it fails on some
(non-darwin) platforms.
Fixes#40979.
Change-Id: If770f7828cdb858857d6079e0585bf067f8f7a92
Reviewed-on: https://go-review.googlesource.com/c/go/+/250944
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
The post-index offset of VLD[1-4]R instructions is decided by the
"size" field not "Q" field, the current assembler uses "Q" fileld
to check the correctness of post-index offset which is not correct.
This patch fixes it.
Fixes#40725
Change-Id: If1cde7f21c6b3ee0e491649eb567700bd1475c84
Reviewed-on: https://go-review.googlesource.com/c/go/+/249757
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The Value field of ast.BasicLit is a string field holding the literal
string. For CHARs and STRINGs, the BasicLit.Value literal includes
quotes, so to use the value in practise one will often need to Unquote
it.
Since this is a common gotcha (I've been bitten by this a few times),
document it, and suggest the use of the strconv.Unquote functions.
Fixes#39590
Change-Id: Ie3e13f5a2a71bb1b59e03bc5b3a16d8e2e7c01d4
Reviewed-on: https://go-review.googlesource.com/c/go/+/244960
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
On Server.Shutdown, all idle connections are closed.
A caveat for new connections is that they are marked idle
after 5 seconds.
Previously new HTTP/2 connections were marked New, and after 5 seconds,
they would then become idle. With this change, we now mark HTTP/2
connections as Active to allow the proper shutdown sequence to occur.
Fixes#36946Fixes#39776
Change-Id: I31efbf64b9a2850ca544da797f86d7e1b3378e8b
Reviewed-on: https://go-review.googlesource.com/c/go/+/240278
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Updates x/net/http2 to git rev 62affa334b73ec65ed44a326519ac12c421905e3
x/net/http2: reject HTTP/2 Content-Length headers containing a sign
https://go-review.googlesource.com/c/net/+/236098/ (fixes#39017)
also updates the vendored version of golang.org/x/net by running
go get golang.org/x/net@62affa334b73ec65ed44a326519ac12c421905e3
go mod tidy
go mod vendor
go generate -run bundle net/http
Change-Id: I7ecfdb7644574c44c3616e3b47664eefd4c926f3
Reviewed-on: https://go-review.googlesource.com/c/go/+/253238
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Ensures that calling TempDir() in either of Cleanup or Benchmark
doesn't cause test failures which were previously caused by the
created directory having been deleted after the first run, yet
we prevented the recreation of the directory due to our selection
of concurrency primitive sync.Once. This change recreates the
temporary directory if it doesn't exist, regardless of how
many times Cleanup and Benchmark are invoked.
Fixes#41062
Change-Id: I925d9f7207d7c369a193d1e17da7a59a586244a7
Reviewed-on: https://go-review.googlesource.com/c/go/+/251297
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The linker assumed macOS is AMD64 (and 386 in the past). It
passes darwin/amd64-specific flags to the external linker when
building for macOS. They don't work for ARM64-based macOS. So
only pass them on AMD64.
Disable DWARF combining for macOS ARM64 for now. The generated
binary doesn't run. (TODO: fix.)
For macOS ARM64 port. External linking now works.
Change-Id: Iab53bc48f4fadd9b91de8898b4b450ea442667a2
Reviewed-on: https://go-review.googlesource.com/c/go/+/253019
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
The current implementation stores the comparison pseudo-ops of arm64
conditional instructions (CSEL/CSEL0) in Aux, this patch modifies it
and stores it in AuxInt, which can avoid the allocation.
Change-Id: I0b69e51f63acd84c6878c6a59ccf6417501a8cfc
Reviewed-on: https://go-review.googlesource.com/c/go/+/252517
Run-TryBot: fannie zhang <Fannie.Zhang@arm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This defines a macro for the regabi GOEXPERIMENT when assembling
runtime assembly code.
In general, assembly code will be shielded from the calling convention
change, but there is a small amount of runtime assembly that is going
to have to change. By defining a macro, we can easily make the small
necessary changes. The other option is to use build tags, but that
would require duplicating nontrivial amounts of unaffected code,
leading to potential divergence issues. (And unlike Go code, assembly
code can't depend on the compiler optimizing away branches on a
feature constant.) We consider the macro preferable, especially since
this is expected to be temporary as we transition to the new calling
convention.
Updates #40724.
Change-Id: I73984065123968337ec10b47bb12c4a1cbc07dc5
Reviewed-on: https://go-review.googlesource.com/c/go/+/252258
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This is the "feature flag" for the register calling convention work
(though since this work is expected to extend over a few releases,
it's not version-prefixed). This will let us develop the register
calling convention on the main branch while maintaining an easy toggle
between the old and new ABIs.
Updates #40724.
Change-Id: I129c8d87d34e6fa0910b6fa43efb35b706021637
Reviewed-on: https://go-review.googlesource.com/c/go/+/252257
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The primary responsibility of declare() to associate a symbol (Sym) with
a declaration (Node), so "oldname" will work. Function literals are
anonymous, so their symbols does not need to be declared.
Passes toolstash-check.
Change-Id: I739b1054e3953e85fbd74a99148b9cfd7e5a57eb
Reviewed-on: https://go-review.googlesource.com/c/go/+/249078
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
They are skipped while processing Func.Dcl anyway.
This CL does not pass toolstash-check, because it reduces the length
of Func.Dcl length, while that length is used to generate autotmp
variables name.
Change-Id: I408183e62ce6c34e5f04c89814ebb9570957e37b
Reviewed-on: https://go-review.googlesource.com/c/go/+/252418
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Fixes the file server to reject requests of the form:
"Range": "bytes=--N"
where "-N" is a negative suffix-length as designated by the
grammar in RFC 7233 Section 2.1, "Byte-Ranges", which specifies
that suffix-length MUST be of the form 1*DIGIT aka a non-negative digit.
Thus requests such as:
"Range": "bytes=--2"
will be rejected with a "416 Range Not Satisfiable" response.
Fixes#40940
Change-Id: I3e89f8326c14af30d8bdb126998a50e02ba002d9
Reviewed-on: https://go-review.googlesource.com/c/go/+/252497
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
While debugging #40771, I realized that the chatty printer should only
ever print to a single io.Writer (normally os.Stdout). The other
Writer implementations in the chain write to local buffers, but if we
wrote a test's output to a local buffer, then we did *not* write it to
stdout and we should not store it as the most recently logged test.
Because the chatty printer should only ever print to one place, it
shouldn't receive an io.Writer as an argument — rather, it shouldn't
be used at all for destinations other than the main output stream.
On the other hand, when we flush the output buffer to stdout in the
top-level flushToParent call, it is important that we not allow some
other test's output to intrude between the test summary header and the
remainder of the test's output. cmd/test2json doesn't know how to
parse such an intrusion, and it's confusing to humans too.
No test because I couldn't reproduce the user-reported error without
modifying the testing package. (This behavior seems to be very
sensitive to output size and/or goroutine scheduling.)
Fixes#40771
Updates #38458
Change-Id: Ic19bf1d535672b096ba1c8583a3b74aab6d6d766
Reviewed-on: https://go-review.googlesource.com/c/go/+/249026
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Process may crash becaues acquireLockRank and releaseLockRank may
be called in nosplit context. With optimizations and inlining
disabled, these functions won't get inlined or have their morestack
calls eliminated.
Nosplit is not strictly required for lockWithRank, unlockWithRank
and lockWithRankMayAcquire, just keep consistency with lockrank_on.go
here.
Fixes#40843
Change-Id: I5824119f98a1da66d767cdb9a60dffe768f13c81
GitHub-Last-Rev: 38fd3ccf6e
GitHub-Pull-Request: golang/go#40844
Reviewed-on: https://go-review.googlesource.com/c/go/+/248878
Reviewed-by: Dan Scales <danscales@google.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
cgo effectively prepends -I${SRCDIR} to the header include path of all
preambles it processes, so when an #include <> matches a header file
both in the source directory and also another include directory, the
local copy will be used in preference.
This behaviour is surprising but unfortunately also longstanding and
relied upon by packages in the wild, so the best we can do is to
document it.
Fixes#41059
Change-Id: If6d2818294b2bd94ea0fe5fd6ce77e54b3e167a6
Reviewed-on: https://go-review.googlesource.com/c/go/+/251758
Reviewed-by: Ian Lance Taylor <iant@golang.org>
I decided to add package and module diagrams to the test cases to make
them easier to follow.
While adding those diagrams, I noticed some strong similarities among
a couple of the graphs, so I consolidated those cases (and deleted the
redundant tests).
For #36460
Change-Id: Id6cd04fc871379b83851c2d1af89ea9296a0f3e5
Reviewed-on: https://go-review.googlesource.com/c/go/+/251997
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
That test tested that import paths with non-ASCII unicode paths
were allowed by the Go command. Remove this test case because
golang.org/cl/251878 removes that support.
Also rewrite a test case in TestRepoRootForImportPath in the test
for cmd/go/internal/get to reflect that unicode directory names are now
disallowed.
Updates #29101
Change-Id: I669e220facd04fc82ccd05dd08e8f1ff4d48b1fd
Reviewed-on: https://go-review.googlesource.com/c/go/+/252297
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Previously, if an example test invoked runtime.Goexit, it would
pass yet hang until a timeout, while regular tests that invoke
runtime.Goexit do fail. This change removes that inconsistent
behavior and makes such example tests fail, and panic with an
indication of having invoked runtime.Goexit.
Fixes#41084
Change-Id: I0ffa152204f2b1580f4d5d6961ba1ce6b13fc022
Reviewed-on: https://go-review.googlesource.com/c/go/+/251857
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The copy of CheckImportPath in path.go and the regular expression for github
repos in vcsPaths together allow import paths with unicode letters with import
paths. These all come from github repos with non-ASCII unicode letters
with paths in directories. This mainly shows up in GOPATH mode, but could
also show up in Module mode when getting a module in GOPROXY=direct mode.
We expect there to not be any significant affected users of this change--
an investingation of github repos that would produce import paths that
would comply with the copy CheckImportPaths that's being removed, but not
modload.CheckImportPaths only surfaced a handful of cases, all of which
seemed to be small test or demonstation repos. But this CL is being
submitted early in the cycle so that it can be backed out if need be.
Updates #29101
Change-Id: I719df4af5b318e1330e90d8a0bffe5bb8d816f4f
Reviewed-on: https://go-review.googlesource.com/c/go/+/251878
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This pulls in golang.org/cl/250920 which rejects Windows shortnames as
path components in module.CheckImportPath (as is already done in
cmd/go/internal/get's copy of CheckImportPath). This will allow us to replace
the copy of CheckImportPath with the original.
This also pulls in golang.org/cl/250919 which rejects + in CheckPath and
CheckImportPath, and golang.org/cl/235597, which adds methods to the zip
package for gorelease, but shouldn't affect cmd.
This change also updates the cmd/go test case TestScript/mod_bad_filenames
to reflect that golang.org/x/mod/zip error messages now include filenames
for bad file names that can't be included in zip archives.
Updates #29101
Change-Id: I7f654325dc33b19bc9c6f77a56546747add5a47f
Reviewed-on: https://go-review.googlesource.com/c/go/+/251877
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Rather than probe and guess if sendfile will work inside ResponseWriter.ReadFrom(src),
this change fixes the underlying issue of starting to respond before src is readable
We'll no longer send a status OK if a header has not yet been written and reading
from src is destined to fail. This small change implicitly takes care of the need for
the server to sniff the response body to determine the Content-Type.
This allows splice to work on Linux when src is a socket or any non-regular file that's spliceable.
The extra read of 512 bytes may raise an objection, and that's fair, but
we're already swapping some syscall prep work for another and a read of
512 probably will not impact the overall performance. For shorter
bodies, there's likely less setup time. A little initial slop is not too
unusual in zero copy network code, and sometimes actually helps.
Fixes#40888
Change-Id: I4a8e2ad0ace1318bae66dae5671d06ea6d4838ed
GitHub-Last-Rev: 097364ea86
GitHub-Pull-Request: golang/go#40903
Reviewed-on: https://go-review.googlesource.com/c/go/+/249238
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
commit 72ec930fa7 added basic support for
relocations, but assumed that the symbol value would be 0, likely because
.debug_info always has address == 0 in the ELF section headers.
CL 195679 added further support for relocations, but explicitly encoded
the original assumption that section addresses would be 0.
This change removes that assumption: all relocations will now be
properly computed based on the target symbol value even when that symbol
is a section with a non-zero address.
Typically, sections that are part of a LOAD program segment have
non-zero addresses. For example, .debug_ranges relocations could be
relative to .text, which usually has an address > 0.
Fixes#40879
Change-Id: Ib0a616bb8b05d6c96d179b03ca33a10946fc5d59
GitHub-Last-Rev: 4200de7326
GitHub-Pull-Request: golang/go#41038
Reviewed-on: https://go-review.googlesource.com/c/go/+/250559
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Remove an extra int32-representable check when deciding to use an int32
constant as an immediate value.
Comment out a broken optimization that relies on MaxUint32 being
representable by a signed int32. It never triggers and when fixed, the
signedness of the auxint prevents other optimization passes from
handling it properly, thus causing segfaults in the runtime.
Remove a couple offset representable in 32-bits checks on 32-bit aux
vals.
toolstash-check clean
Change-Id: I148b53403fde523c90d692cb90e412460664b439
Reviewed-on: https://go-review.googlesource.com/c/go/+/230458
Reviewed-by: Keith Randall <khr@golang.org>
We currently use two fields to store the targets of branches.
Some phases use p.To.Val, some use p.Pcond. Rewrite so that
every branch instruction uses p.To.Val.
p.From.Val is also used in rare instances.
Introduce a Pool link for use by arm/arm64, instead of
repurposing Pcond.
This is a cleanup CL in preparation for some stack frame CLs.
Change-Id: If8239177e4b1ea2bccd0608eb39553d23210d405
Reviewed-on: https://go-review.googlesource.com/c/go/+/251437
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Also don't restart DWARF reading from beginning when we are testing
multiple entries.
Also reformat relocationTests slice to use indexed literals.
Change-Id: Ia5f17214483394d0ef033be516df61f0bdc521b6
Reviewed-on: https://go-review.googlesource.com/c/go/+/251637
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
When zero or less, maxIdleTime and maxLifetime means unlimited.
Helper function shortestIdleTimeLocked must not return the
minimum of the two until both are verified to be greater
then zero.
Fixes#40841
Change-Id: I1130332baf4ad259cd90c10f4221f5def8510655
Reviewed-on: https://go-review.googlesource.com/c/go/+/248817
Reviewed-by: Daniel Theophanes <kardianos@gmail.com>
The order array was zero initialized by the compiler, but ends up being
overwritten by the runtime anyway.
So let the runtime takes full responsibility for initializing, save us
one instruction per select.
Fixes#40399
Change-Id: Iec1eca27ad7180d4fcb3cc9ef97348206b7fe6b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/251517
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
The current command will run this entire set of tests, which takes a
noticeable amount of time. Contributors may wish to run only a subset of
these tests to save time/compute (e.g. when iterating on a CL that
failed tests in that subset). Listing file(s) as operands to the command
will run only those tests.
Change-Id: I1874c43681a594190bc40b61cee0b8d321be73f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/242997
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
This was fixed by CL 242084. Retroactively add some tests that would
have failed before the fix.
Also, remove some existing duplicate tests.
Change-Id: I95f7a215d4a9651ded6d739f89c574f33f573c60
Reviewed-on: https://go-review.googlesource.com/c/go/+/251397
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
There was some duplication of logic interpreting the implicit type of
an operand in assignableTo and convertUntyped. Factor out this logic to
a new 'implicitType' function, which returns the implicit type of an
untyped operand when used in a context where a target type is expected.
I believe this resolves some comments about code duplication. There is
other similar code in assignable, assignableTo, and convertUntypes, but
I found it to to be sufficiently semantically distinct to not warrant
factoring out.
Change-Id: I199298a2e58fcf05344318fca0226b460c57867d
Reviewed-on: https://go-review.googlesource.com/c/go/+/242084
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
cmd/go.TestScript/test_main_twice demonstrates a program that invokes
(*M).Run twice in a row. If we only restore os.Exit(0) in m.afterOnce,
we will fail to restore it after the second run and fail the test
process despite both runs passing.
Updates #29062
Updates #23129
Change-Id: Id22ec68f1708e4583c8dda14a8ba0efae7178b85
Reviewed-on: https://go-review.googlesource.com/c/go/+/251262
Run-TryBot: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
We're using sort.SliceStable, so no need to keep track of indexes as well.
Use a more robust test for whether a node is a call.
Add a test that we're actually reordering comparisons. This test fails
without the alg.go changes in this CL because eqstring uses OCALLFUNC
instead of OCALL for its data comparisons.
Update #8606
Change-Id: Ieeec33434c72e3aa328deb11cc415cfda05632e2
Reviewed-on: https://go-review.googlesource.com/c/go/+/237921
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
We currently use two fields to store the targets of branches.
Some phases use p.To.Val, some use p.Pcond. Rewrite so that
every branch instruction uses p.To.Val.
p.From.Val is also used in rare instances.
Introduce a Pool link for use by arm/arm64, instead of
repurposing Pcond.
This is a cleanup CL in preparation for some stack frame CLs.
Change-Id: I9055bf0a1d986aff421e47951a1dedc301c846f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/243318
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
cgo parses GCC’s error messages to classify C identifiers referenced
from Go programs (are they integer constants? type names?). If GCC tries
to colorize its errors, cgo can’t figure out what GCC is saying. GCC
avoids escape sequences in this scenario by default, but the default
behavior can be overridden in at least two places:
- The user can set `CGO_COPTS=-fdiagnostics-color`.
- Whoever compiled GCC can configure GCC itself to always colorize
output.
The most reliable way to ensure that GCC doesn’t colorize output is to
append `-fdiagnostics-color=never` to the GCC command line; do so.
Fixes#40415
Change-Id: Id4bdf8d92fac8b038340b4264f726e8fe38875b4
Reviewed-on: https://go-review.googlesource.com/c/go/+/248398
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
This merges an lis + subf into subfic, and for 32b constants
lwa + subf into oris + ori + subf.
The carry bit is no longer used in code generation, therefore
I think we can clobber it as needed. Note, lowered borrow/carry
arithmetic is self-contained and thus is not affected.
A few extra rules are added to ensure early transformations to
SUBFCconst don't trip up earlier rules, fold constant operations,
or otherwise simplify lowering. Likewise, tests are added to
ensure all rules are hit. Generic constant folding catches
trivial cases, however some lowering rules insert arithmetic
which can introduce new opportunities (e.g BitLen or Slicemask).
I couldn't find a specific benchmark to demonstrate noteworthy
improvements, but this is generating subfic in many of the default
bent test binaries, so we are at least saving a little code space.
Change-Id: Iad7c6e5767eaa9dc24dc1c989bd1c8cfe1982012
Reviewed-on: https://go-review.googlesource.com/c/go/+/249461
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Carlos Eduardo Seo <cseo@linux.vnet.ibm.com>
This is a port of CL 244628, updated to move some existing regression
tests into the fixedbugs directory, and to use subtests. Also,
'TestFixed' is renamed to 'TestFixedBugs'.
Change-Id: I43aac3f75f2bd850567d08e8b008d91aeb717064
Reviewed-on: https://go-review.googlesource.com/c/go/+/247904
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Better still would be to avoid passing around module.Version instances
with invalid Version strings in the first place, so that any time we
see a module.Version we know that it is actually a version of a module
(and not a structurally-similar datum with something else tacked on to
one of the fields). But that's a bigger cleanup for which I don't
currently have enough bandwidth.
Fixes#41060
Change-Id: I32fba5619105cbf67dd03691064c82b8ebb3ce18
Reviewed-on: https://go-review.googlesource.com/c/go/+/250951
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
TestBenchmark is broken due to lack of a Config.Importer, but
unfortunately fails silently due to an unchecked error.
Fix the importer and check the error. Also improve the output to include
allocation stats.
Finally, don't run TestBenchmark on go/types by default. If the
benchmark is being used during a refactoring of go/types itself, results
for go/types will not be comparable.
Change-Id: Ib6bdb6807403b3ec99762f535e2496c94bd9b6e0
Reviewed-on: https://go-review.googlesource.com/c/go/+/249517
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Add s390x support to the addressing modes pass. This significantly
reduces the number of rules we need to have to handle indexed
addressing modes on s390x.
There are some changes introduced with the new approach. Notably
pointer calculations of the form '(ADD x (ADDconst y [c]))' won't
get fully merged into address fields right now, the constant offset
will remain separate. That is a relatively minor issue though.
file before after Δ %
addr2line 4120904 4120960 +56 +0.001%
api 4944005 4948765 +4760 +0.096%
asm 4977431 4984335 +6904 +0.139%
buildid 2683760 2683504 -256 -0.010%
cgo 4557976 4558408 +432 +0.009%
compile 19103577 18916634 -186943 -0.979%
cover 4883694 4885054 +1360 +0.028%
dist 3545177 3553689 +8512 +0.240%
doc 3921766 3921518 -248 -0.006%
fix 3295254 3302182 +6928 +0.210%
link 6539222 6540286 +1064 +0.016%
nm 4105085 4107757 +2672 +0.065%
objdump 4546015 4545439 -576 -0.013%
pack 2416661 2415485 -1176 -0.049%
pprof 13267433 13265489 -1944 -0.015%
test2json 2762180 2761996 -184 -0.007%
trace 10145090 10135626 -9464 -0.093%
vet 6772946 6771738 -1208 -0.018%
total 106588176 106418865 -169311 -0.159%
Fixes#37891.
Change-Id: If60d51f31eb2806b011432a6519951b8668cb42f
Reviewed-on: https://go-review.googlesource.com/c/go/+/250958
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The AssignableTo API is specifically for non-constant values, but is
currently called by gopls for constant completions. Add a test to ensure
that we handle this edge case correctly.
Change-Id: I83115cbca2443a783df1c3090b5741260dffb78e
Reviewed-on: https://go-review.googlesource.com/c/go/+/250258
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
The error returned by convertUntyped is 'cannot convert _ to _', which
can be misleading in contexts where an explicit conversion would be
allowed.
Arguably the error message from convertUntyped should just be 'cannot
use _ as _', as 'convert' has an explicit meaning within the spec.
Making that change caused a large number of test failures, so for now we
just fix this for assignments by interpreting the error.
For #22070
Change-Id: I4eed6f39d1a991e8df7e035ec301d28a05150eb5
Reviewed-on: https://go-review.googlesource.com/c/go/+/242083
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
The intermediate SSA opcodes* are no longer generated during the
lowering pass. The shifting rules have been improved using ISEL.
Therefore, we can remove them and the rules which expand them.
* The removed opcodes are:
LoweredAdd64Carry
ADDconstForCarry
MaskIfNotCarry
FlagCarryClear
FlagCarrySet
Change-Id: I1ebe2726ed988f29ed4800c8f57b428f7a214cd0
Reviewed-on: https://go-review.googlesource.com/c/go/+/249462
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
This change rejects module paths that don't conform to
the new checkModulePathLax function, when loading a go.mod
file. The change uses the checkModulePathLax function instead of
CheckPath because there are still many users who are using
unpublished modules with unpublishable paths, and we don't
want to break them all.
Next, before this change, when go mod init is run in GOPATH,
it would try to use the location of the directory within GOPATH
to infer the module path. After this change, it will only use
that inferred module path if it conforms to module.CheckPath.
Change-Id: Idb36d1655cc76aae82671e87ba634609503ad1a2
Reviewed-on: https://go-review.googlesource.com/c/go/+/211597
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
'go get' will now warn about retracted versions in the build list,
after updating go.mod. The warning instructs users to run
'go get module@latest' to upgrade or downgrade away from the retracted
version.
'go get' now allows users to explicitly request a specific retracted
version.
For #24031
Change-Id: I15fda918dc84258fb35b615dcd33b0f499481bd7
Reviewed-on: https://go-review.googlesource.com/c/go/+/228383
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The -retracted flag causes 'go list' to load information about
retracted module module versions.
When -retracted is used with -f or -json, the Retracted field is set
to a string containing the reason for the retraction on retracted
module versions. The string is based on comments on the retract
directive. This field is also populated when the -u flag is used.
When -retracted is used with -versions, retracted versions are shown.
Normally, they are omitted.
For #24031
Change-Id: Ic13d516eddffb1b8404e21034f78cecc9896d1b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/228382
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The go command now recognizes 'retract' directives in go.mod. A
retract directive may be used by a module author to indicate a
version should not be used. The go command will not automatically
upgrade to a retracted version. Retracted versions will not be
considered when resolving version queries like "latest" that don't
refer to a specific version.
Internally, when the go command resolves a version query, it will find
the highest release version (or pre-release if no release is
available), then it will load retractions from the go.mod file for
that version. Comments on retractions are treated as a rationale and
may appear in error messages. Retractions are only loaded when a query
is resolved, so this should have no impact on performance for most
builds, except when go.mod is incomplete.
For #24031
Change-Id: I17d643b9e03a3445676dbf1a5a351090c6ff6914
Reviewed-on: https://go-review.googlesource.com/c/go/+/228380
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Query and other functions now accept an "allowed" function that
returns an error (previously, the function returned a bool). If the
error is equivalent to ErrDisallowed, it indicates the version is
excluded (or, in a future CL, retracted). This provides predicates a
chance to explain why a version is not allowed.
When a query refers to a specific revision (by version, branch, tag,
or commit name), most callers will not use the Allowed predicate. This
allows commands like 'go list -m' and 'go mod download' to handle
disallowed versions when explicitly requested. 'go get' will reject
excluded versions though.
When a query does not refer to a specific revision (for example,
"latest"), disallowed versions will not be considered.
When an "allowed" predicate returns an error not equivalent to
ErrDisallowed, it may be ignored or returned, depending on the
case. This never happens for excluded versions, but it may happen for
retractions (in a future CL). This indicates a list of retractions
could not be loaded. This frequently happens when offline, and it
shouldn't cause a fatal or warning in most cases.
For #24031
Change-Id: I4df6fb6bd60e3e0259e5b3b4bf71a307b4b32298
Reviewed-on: https://go-review.googlesource.com/c/go/+/228379
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
cmd/internal/objabi/doc.go has comments decribing the (old)
object file format. But cmd/internal/objabi has nothing to do
with object files, and never did. Delete.
Move some comment to cmd/internal/goobj, where the (new) object
file format is actually defined, and update to reflect the
current status.
Change-Id: Ied96089df4be35e5d259a572ed60ee00f2cd0d1d
Reviewed-on: https://go-review.googlesource.com/c/go/+/249958
Reviewed-by: Than McIntosh <thanm@google.com>
This CL applies strong aux typing to the remaining s390x rewrite
rules in preparation for strong aux typing becoming the default.
Passes toolstash-check on s390x.
Change-Id: Id585b0db492780737818024e1b22b4837435b525
Reviewed-on: https://go-review.googlesource.com/c/go/+/250558
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
JavaScript environments are quite unpredictable because bundlers add
mocks for compatibility and libraries can polute the global namespace.
Detect more of such situations:
- Add check that require("fs") returns an object.
- Fix check that require("fs") returns an non-empty object.
- Add check that "module" is defined.
Fixes#40730
Change-Id: I2ce65fc7db64bbbb0b60eec79a4cfe5c3fec99c0
Reviewed-on: https://go-review.googlesource.com/c/go/+/248758
Run-TryBot: Richard Musiol <neelance@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
CL 220905 added code to identify alternate transports that always error
by using http2erringRoundTripper. This does not work when the transport
is from another package, e.g., http2.erringRoundTripper.
Expose a new method that allow detection of such a RoundTripper.
Switch to an interface that is both a RoundTripper and can return the
underlying error.
Fixes#40213
Change-Id: I170739857ab9e99dffb5fa55c99b24b23c2f9c54
Reviewed-on: https://go-review.googlesource.com/c/go/+/243258
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
In order to generate more accurate or informative error messages from
the type checker, it can be helpful to interpret error messages in
context. This is currently achieved in a number of ways:
+ Return a boolean value, and then reverse-engineer the error at the
callsite (as in representable->representableConst).
+ Return a value causing the error (as in Checker.missingMethod), and
add the error at the callsite.
+ Pass a "reason" string pointer to capture the error (as in
Checker.assignableTo), and add the error at the callsite.
+ Pass a "context" string pointer, and use this when writing errors in
the delegated method.
In all cases, it is the responsibility of whatever code calls
Checker.error* to set the operand mode to invalid.
These methods are used as appropriate, depending on whether multiple
errors are generated, whether additional context is needed, and whether
the mere presence of an error needs to be interpreted at the callsite.
However, this practice has some downsides: the plurality of error
handling techniques can be a barrier to readability and composability.
In this CL, we introduce Yet Another Pattern, with the hope that it can
replace some or all of the existing techniques: factor out side-effect
free functions that evaluate a single error, and add helpers for
recording this error in the Checker.
As a proof of concept this is done for Checker.representable and
Checker.convertUntyped. If the general pattern does not seem appropriate
for replacing some or all of the error-handling techniques listed above,
we should revert to an established technique.
Some internal error APIs are refactored to operate on an error, rather
than a types.Error, with internal error metadata extracted using
errors.As. This seemed to have negligible impact on performance, but we
should be careful about actually wrapping errors: I expect that many
users will expect err to be a types.Error.
Change-Id: Ic5c6edcdc02768cd84e04638fad648934bcf3c17
Reviewed-on: https://go-review.googlesource.com/c/go/+/242082
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Fix a data race for clients that mutate requests after receiving a
response error which is caused by the writeLoop goroutine left
running, this can be seen on canceled requests.
Fixes#37669
Change-Id: I0e0e4fd63266326b32587d8596456760bf848b13
Reviewed-on: https://go-review.googlesource.com/c/go/+/232799
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Old behavior is still enabled because it doesn't hurt to leave
it in and existing users of this feature (there are dozens of
us!) will not be surprised. Adding this finer control allows
users to avoid writing ssa.html where they can't, shouldn't, or
just don't want to.
Example, both ways:
$ GOSSAFUNC="(*Reader).Reset" go test -c -o ./a compress/gzip
dumped SSA to bytes/ssa.html
dumped SSA to strings/ssa.html
dumped SSA to bufio/ssa.html
dumped SSA to compress/gzip/ssa.html
$ GOSSAFUNC="compress/gzip.(*Reader).Reset" go test -c -o ./a compress/gzip
dumped SSA to compress/gzip/ssa.html
Updates #40919.
Change-Id: I06b77c3c1d326372a32651570b5dd6e56dfb1d7f
Reviewed-on: https://go-review.googlesource.com/c/go/+/250340
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The StripPrefix wrapper strips a prefix string from the request's
URL.Path field, but doesn't touch the RawPath field. This leads to the
confusing situation when StripPrefix handles a request with URL.RawPath
populated (due to some escaped characters in the request path) and the
wrapped request's RawPath contains the prefix but Path does not.
This change modifies StripPrefix to strip the prefix from both Path and
RawPath. If there are escaped characters in the prefix part of the
request URL the stripped handler serves a 404 instead of invoking the
underlying handler with a mismatched Path/RawPath pair.
This is a backward incompatible change for a very small minority of
requests; I would be surprised if anyone is depending on this behavior,
but it is possible. If that's the case, we could make a more
conservative change where the RawPath is trimmed if possible, but when
the prefix contains escaped characters then we don't 404 but rather send
through the invalid Path/RawPath pair as before.
Fixes#24366
Change-Id: I7030b8c183a3dfce307bc0272bba9a18df4cfe08
Reviewed-on: https://go-review.googlesource.com/c/go/+/233637
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Previously, Readdirnames returned a *PathError on Windows and Plan 9,
but a *SyscallError on POSIX systems.
In contrast, similar methods (such as Stat) return a *PathError on all platforms.
Fixes#38923
Change-Id: I26395905b1e723933f07b792c7aeee7c335949cd
Reviewed-on: https://go-review.googlesource.com/c/go/+/233917
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Right now we just prevent such types from being on the heap. This CL
makes it so they cannot appear on the stack either. The distinction
between heap and stack is pretty vague at the language level (e.g. it
is affected by -N), and we don't need the flexibility anyway.
Once go:notinheap types cannot be in either place, we don't need to
consider pointers to such types to be pointers, at least according to
the garbage collector and stack copying. (This is the big win of this
CL, in my opinion.)
The distinction between HasPointers and HasHeapPointer no longer
exists. There is only HasPointers.
This CL is cleanup before possible use of go:notinheap to fix#40954.
Update #13386
Change-Id: Ibd895aadf001c0385078a6d4809c3f374991231a
Reviewed-on: https://go-review.googlesource.com/c/go/+/249917
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
par.Work performs two different tasks: deduplicating work (a task
which overlaps with par.Cache), and executing limited active work in
parallel. It also requires the caller to re-invoke Do whenever the
workqueue transititions from empty to non-empty.
The new par.Queue only performs the second of those two tasks, and
presents a simpler API: it starts and stops its own goroutines as
needed (indicating its idle state via a channel), rather than
expecting the caller to drive the transitions explicitly.
For #36460
Change-Id: I5c38657dda63ab55718497467d05d41744ff59f2
Reviewed-on: https://go-review.googlesource.com/c/go/+/247766
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Previously they were cached per mvsReqs instance. However, the
contents of the go.mod file of a given dependency version can only
vary if the 'replace' directives that apply to that version have
changed, and the only time we change 'replace' directives is in 'go
mod edit' (which does not care about the build list or MVS).
This not only simplifies the mvsReqs implementation, but also makes
more of the underlying logic independent of mvsReqs.
For #36460
Change-Id: Ieac20c2fcd56f64d847ac8a1b40f9361ece78663
Reviewed-on: https://go-review.googlesource.com/c/go/+/244774
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Previously, this cache was a member of the (ephemeral) modload.loader
struct. However, the Go language version for a given module version
does not vary based on the build list, the set of loaded packages, the
build tags in use, the meaning of the "all" pattern, or anything else
that can be configured for an instance of the package loader. The map
containing that information is therefore not appropriate as a field of
the (configurable, package-list-dependent) loader struct.
The Go language version mapping could, in theory, be read from the
go.mod file in the module cache (or replacement directory) every time
it is needed: this map is just a cache, and as such it belongs
alongside the other caches and indexes in the modload package, which
are currently found in modfile.go.
We may want to do the same sort of global caching for the mapping from
each module.Version to its list of direct requirements (which are
similarly idempotent), but for now that is left for a future change.
For #36460
For #36876
Change-Id: I90ac176ffea97f30c47d6540c3dfb874dc9cfa4f
Reviewed-on: https://go-review.googlesource.com/c/go/+/244078
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
Previously, we suppressed the module version annotation if the last
error in the stack was a *module.ModuleError, regardless of its path.
However, if the error is for a replacement module, that produces a
confusing error message: the error is attributed to the last module in
the error path, but actually originates in the replacement (which is
not otherwise indicated).
Now, we print both the original and the replacement modules when they
differ, which may add some unfortunate redundancy in the output but at
least doesn't drop the very relevant information about replacements.
Fixes#35039
Change-Id: I631a7398033602b1bd5656150a4fad4945a87ade
Reviewed-on: https://go-review.googlesource.com/c/go/+/247765
Reviewed-by: Jay Conrod <jayconrod@google.com>
When we print the stack from a BuildListError, we print the main
module first and the error last. That was the opposite of the order in
which in was stored in memory, leading to (arguably) more complex code
and (definitely) my own inability to reason about the contents of the
slice.
For now, it's still more convenient to construct the stack reversed,
so we do that and then reverse it before packing it into the error.
For #36460
Change-Id: I6312fb67b2ad9bf9b64071fe829854833208bad7
Reviewed-on: https://go-review.googlesource.com/c/go/+/244759
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Previously, when we encountered an excluded version in any module's
requirements, we would resolve it to the next higher version.
Unfortunately, the meaning of “the next higher version” can change
over time.
Moreover, users who use 'exclude' directives normally either already
require some higher version (using the 'exclude' directive to prune
out invalid requirements from some intermediate version), or already
require some lower version (using the 'exclude' directive to prevent
'go get -u' from upgrading to a known-bad version). In both of these
cases, resolving an upgrade for the excluded version is needless work
even in the best case: it adds work for the 'go' command when there is
already a perfectly usable selected version of the module in the
requirement graph.
Instead, we now interpret the 'exclude' directive as dropping all
references to the excluded version.
This implements the approach described in
https://golang.org/issue/36465#issuecomment-572694990.
Fixes#36465
Updates #36460
Change-Id: Ibf0187daced417b4cc23b97125826778658e4b0f
Reviewed-on: https://go-review.googlesource.com/c/go/+/244773
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This allows semver-based comparisons of the version without additional allocations.
Also comment on the reason for the loops that iterate over modFile instead.
(I was reading the vendor code in order to add the lazy-loading version check,
and this section was a bit unclear to me.)
For #36460
Change-Id: I11559d81ffb4eba0e4e10e6fa3c01990b11f9180
Reviewed-on: https://go-review.googlesource.com/c/go/+/240622
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
Due to a typo, this test case was not actually exercising the bug
described in golang/go#33656. Update it to do so. Interestingly, the
comparison is now valid (as it should be) -- I suspect #33656 is
actually fixed.
Fixes#33656
Change-Id: If50a917f6477d8eb4f82f5a2a96bf5d9123ff0d4
Reviewed-on: https://go-review.googlesource.com/c/go/+/241263
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
These exported functions are mostly trivial wrappers, but do make
certain assumptions about how the underlying Checker APIs can be called.
Add some simple tests.
Change-Id: I68e9ae875353c12d118ec961a6f3834385fbbb97
Reviewed-on: https://go-review.googlesource.com/c/go/+/241262
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
With this patch, opt pass can expose more obvious constant-folding
opportunites.
Example:
func test(i int) int {return (i+8)-(i+4)}
The previous version:
MOVD "".i(FP), R0
ADD $8, R0, R1
ADD $4, R0, R0
SUB R0, R1, R0
MOVD R0, "".~r1+8(FP)
RET (R30)
The optimized version:
MOVD $4, R0
MOVD R0, "".~r1+8(FP)
RET (R30)
This patch removes some existing reassociation rules, such as "x+(z-C)",
because the current generic rewrite rules will canonicalize "x-const"
to "x+(-const)", making "x+(z-C)" equal to "x+(z+(-C))".
This patch also adds test cases.
Change-Id: I857108ba0b5fcc18a879eeab38e2551bc4277797
Reviewed-on: https://go-review.googlesource.com/c/go/+/237137
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This patch adds the ARM6464Bitfield auxInt to auxIntType() and
returns its Go type as "arm64Bitfield" type, which is defined
as int16 type.
And the Go type of SymOff auxInt is int32, but some functions
(such as min(), areAdjacentOffsets() and read16/32/64(),etc.)
use SymOff as an input parameter and treat its type as int64,
this patch adds the type conversion for these rules.
Passes toolstash-check -all.
Change-Id: Ib234b48d0a97ef244dd37878e06b5825316dd782
Reviewed-on: https://go-review.googlesource.com/c/go/+/234378
Reviewed-by: Keith Randall <khr@golang.org>
Add a new conversion function to convert aux type to Op type.
Passes toolstash-check -all.
Change-Id: I25d649a5296f6f178d64320dfc5d291e0a597e24
Reviewed-on: https://go-review.googlesource.com/c/go/+/233739
Reviewed-by: Keith Randall <khr@golang.org>
Add a check code for checking whether the "c" value can be
represented as a signed 32 bit integer in some rules.
Passes toolstash-check -all.
Change-Id: I84e4431dc75945985695516d34565b841c8b53e0
Reviewed-on: https://go-review.googlesource.com/c/go/+/233738
Reviewed-by: Keith Randall <khr@golang.org>
CL 249962 added wasm StorepNoWB implementation in assembly, it's now
like all other architectures. This CL adds a general test that the
second param of StorepNoWB must be force to escape.
Fixes#40975
Change-Id: I1eccc7e50a3ec742a1912d65f25b15f9f5ad9241
Reviewed-on: https://go-review.googlesource.com/c/go/+/249761
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Turns out if your failure is in a function with a name like "Reset()"
there will be a lot of hits on the same hashcode. Adding package sensitivity
solves this problem.
In additionm, it turned out that in the case that a logfile was specified
for the GOSSAHASH logging, that it was opened in create mode, which meant
that multiple compiler invocations would reset the file to zero length.
Opening in append mode works better; the automated harness
(github.com/dr2chase/gossahash) takes care of truncating the file before use.
Change-Id: I5601bc280faa94cbd507d302448831849db6c842
Reviewed-on: https://go-review.googlesource.com/c/go/+/246937
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The second argument of StorepNoWB must be forced to escape.
The current Go code does not explicitly enforce that property.
By implementing in assembly, and not using go:noescape, we
force the issue.
Test is in CL 249761. Issue #40975.
This CL is needed for CL 249917, which changes how go:notinheap
works and breaks the previous StorepNoWB wasm code.
I checked for other possible errors like this. This is the only
go:notinheap that isn't in the runtime itself.
Change-Id: I43400a806662655727c4a3baa8902b63bdc9fa57
Reviewed-on: https://go-review.googlesource.com/c/go/+/249962
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Current peFile.goarch looks for symbols like "_rt0_386_windows" to
determine GOARCH. But "_rt0_386_windows" is not present in executables
built with cgo.
Use pe.FileHeader.Machine instead. This should work with any Windows
executable, not just with Go built executable.
Fixes#39682
Change-Id: Ie0ffce664f4b8b8fed69b2ecc482425b042a38d5
Reviewed-on: https://go-review.googlesource.com/c/go/+/240957
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
In
type T1 struct { t2 }
type t2 int
func (t2) M()
T1 has method M because it embeds t2, which has M. Classify
the example
func ExampleT1_M
with T1 instead of ignoring it, as is done currently. There is no
other way to provide an example for such a method, since its original
type is unexported.
Continue to ignore examples on methods from embedded types that are
exported, unless in AllMethods mode. Examples for those methods could
be written on the original type.
The change involves removing a check in classifyExamples. The check
isn't necessary to get the above behavior because
reader.collectEmbeddedMethods and sortedFuncs already generate the
appropriate list of methods.
For #40172.
Change-Id: Ibe7d965ecba6426466184e6e6655fc05989e9caf
Reviewed-on: https://go-review.googlesource.com/c/go/+/249557
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
checkptr has code to recognize &^ expressions, but it didn't take into
account that "p &^ x" gets rewritten to "p & ^x" during walk, which
resulted in false positive diagnostics.
This CL changes walkexpr to mark OANDNOT expressions with Implicit
when they're rewritten to OAND, so that walkCheckPtrArithmetic can
still recognize them later.
It would be slightly more idiomatic to instead mark the OBITNOT
expression as Implicit (as it's a compiler-generated Node), but the
OBITNOT expression might get constant folded. It's not worth the extra
complexity/subtlety of relying on n.Right.Orig, so we set Implicit on
the OAND node instead.
To atone for this transgression, I add documentation for nodeImplicit.
Fixes#40917.
Change-Id: I386304171ad299c530e151e5924f179e9a5fd5b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/249477
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
The optimization that replaces inline markers with pre-existing
instructions assumes that 'Prog' values produced by the compiler are
still reachable after the assembler has run. This was not true on
s390x where the assembler was removing NOP instructions from the
linked list of 'Prog' values. This led to broken inlining data
which in turn caused an infinite loop in the runtime traceback code.
Fix this by stopping the s390x assembler backend removing NOP
values. It does not make any difference to the output of the
assembler because NOP instructions are 0 bytes long anyway.
Fixes#40473.
Change-Id: I9b97c494afaae2d5ed6bca4cd428b4132b5f8133
Reviewed-on: https://go-review.googlesource.com/c/go/+/249448
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This change does context propagation (and only context propagation)
necessary to add context to modfetch.Download and pkg.LoadImport.
This was done by adding context to their callers, and then
adding context to all call-sites, and then repeating adding
context to callers of those enclosing functions and their
callers until none were left. In some cases the call graph expansion
was pruned by using context.TODOs.
The next CL will add a span to Download. I kept it out of this
change to avoid making it any larger (and harder to review)
than it needs to be.
Updates #38714
Change-Id: I7a03416e04a14ca71636d96f2c1bda2c4c30d348
Reviewed-on: https://go-review.googlesource.com/c/go/+/249021
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Opening the html writer can fail, and the failure printer wants
to use the entry block's line number. So make sure we set up
the entry block first.
Fixes#40919
Change-Id: I4ffa2839b45a721bbaf04ff84418e8108fa1cc37
Reviewed-on: https://go-review.googlesource.com/c/go/+/249497
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
par.Work is used in a number of places as a parallel
work queue. This change replaces it with goroutines
and channels in a number of simpler places where it's
used.
This is the same CL as golang.org/cl/240062 and golang.org/cl/248326
except for the following changes in convert.go (all line numbers
from this CL), as well as fixing up imports in download.go:
- On line 44, the "*" before modules.Versions is removed (we were
trying to assign to a nil value on lines 72 and 73).
- Line 64 is new, and ensures that we receive on the semaphore
channel once the goroutine function exits. (The previous versions
of this CL only received at the end of the function, ignoring
the return point in the branch in the middle of the function.)
- The semaphore channel receive right before line 74 is gone,
replaced with the deferred receive above.
- The if block at line 83 is new, accounting for cases where
modfetch.ImportRepoRev returned an error in the goroutine,
so that versions[i] is ignored.
Change-Id: I0e33670bb2eb0a1e4d7a5fa693a471e61ffbc8b7
Reviewed-on: https://go-review.googlesource.com/c/go/+/249020
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Previously, on Unix systems, when the profiler was enabled or disabled,
we called setitimer once per thread. With this change we instead call
it once per process.
Change-Id: I90f0189b562e11232816390dc7d55ed154bd836d
Reviewed-on: https://go-review.googlesource.com/c/go/+/240003
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
People sometimes expect Join to trim .. components from its arguments
before joining, and are surprised that it doesn't. This is bad if they
were relying on that assumed behaviour to prevent directory traversal
attacks.
While a careful reading of the documentation for Join and Clean
might dispel this notion, it is not obvious at first glance.
Add a case to the examples to nudge people in the right direction.
Updates #40373
Change-Id: Ib5792c12ba1000811a0c0eb77048196d0b26da60
Reviewed-on: https://go-review.googlesource.com/c/go/+/249177
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Add a new lowering rule to match and replace such instances
with the MADDLD instruction available on power9 where
possible.
Likewise, this plumbs in a new ppc64 ssa opcode to house
the newly generated MADDLD instructions.
When testing ed25519, this reduced binary size by 936B.
Similarly, MADDLD combination occcurs in a few other less
obvious cases such as division by constant.
Testing of golang.org/x/crypto/ed25519 shows non-trivial
speedup during keygeneration:
name old time/op new time/op delta
KeyGeneration 65.2µs ± 0% 63.1µs ± 0% -3.19%
Signing 64.3µs ± 0% 64.4µs ± 0% +0.16%
Verification 147µs ± 0% 147µs ± 0% +0.11%
Similarly, this test binary has shrunk by 66488B.
Change-Id: I077aeda7943119b41f07e4e62e44a648f16e4ad0
Reviewed-on: https://go-review.googlesource.com/c/go/+/248723
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Remove some unnecessary code. Most significantly, we can skip testing
"if ch == nil { block() }", because this is already the semantics
implied by normal send/receive operations.
Updates #40410.
Change-Id: I4acd33383cc876719fc3b998d85244d4ac1ff9d9
Reviewed-on: https://go-review.googlesource.com/c/go/+/245126
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Currently, we include a "kind" field on scase to distinguish the three
kinds of cases in a select statement: sends, receives, and defaults.
This commit removes by kind field by instead arranging for the
compiler to always place sends before receives, and to provide their
counts separately. It also passes an explicit "block bool" parameter
to avoid needing to include a default case in the array.
It's safe to shuffle cases like this because the runtime will
randomize the order they're polled in anyway.
Fixes#40410.
Change-Id: Iaeaed4cf7bddd576d78f2c863bd91a03a5c82df2
Reviewed-on: https://go-review.googlesource.com/c/go/+/245125
Reviewed-by: Keith Randall <khr@golang.org>
Per-case PCs are only needed for race detector builds, so this allows
skipping allocating stack space for them for non-race builds.
It's possible to arrange the PCs and order arrays consecutively in
memory so that we could just reuse the order0 pointer to identify
both. However, there's more risk of that silently going wrong, so this
commit passes them as separate arguments for now. We can revisit this
in the future.
Updates #40410.
Change-Id: I8468bc25749e559891cb0cb007d1cc4a40fdd0f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/245124
Reviewed-by: Keith Randall <khr@golang.org>
Currently, selectgo does an initial pass over the cases array to look
for entries with nil channels, so they can be easily recognized and
skipped later on. But this still involves actually visiting the cases.
This commit changes selectgo to omit cases with nil channels when
constructing pollorder, so that they'll be skipped over entirely later
on. It also checks for caseDefault up front, which will facilitate
changing it to use a "block bool" parameter instead.
Updates #40410.
Change-Id: Icaebcb8f08df03cc33b6d8087616fb5585f7fedd
Reviewed-on: https://go-review.googlesource.com/c/go/+/245123
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
selectgo will report at most one block event, so there's no need to
keep a releasetime for every select case. It suffices to simply track
the releasetime of the case responsible for the wakeup.
Updates #40410.
Change-Id: I72679cd43dde80d7e6dbab21a78952a4372d1e79
Reviewed-on: https://go-review.googlesource.com/c/go/+/245122
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The current wakeup protocol for channel communications is that the
second goroutine sets gp.param to the sudog when a value is
successfully communicated over the channel, and to nil when the wakeup
is due to closing the channel.
Setting nil to indicate channel closure works okay for chansend and
chanrecv, because they're only communicating with one channel, so they
know it must be the channel that was closed. However, it means
selectgo has to re-poll all of the channels to figure out which one
was closed.
This commit adds a "success" field to sudog, and changes the wakeup
protocol to always set gp.param to sg, and to use sg.success to
indicate successful communication vs channel closure.
While here, this also reorganizes the chansend code slightly so that
the sudog is still released to the pool if the send blocks and then is
awoken because the channel closed.
Updates #40410.
Change-Id: I6cd9a20ebf9febe370a15af1b8afe24c5539efc6
Reviewed-on: https://go-review.googlesource.com/c/go/+/245019
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Keith Randall <khr@golang.org>
Replace strings.Split by strings.IndexByte and explicit
slicing to avoid the allocation of the return slice
of strings.Split.
name old time/op new time/op delta
Marshal 43.3µs ± 1% 36.7µs ± 1% -15.23% (p=0.000 n=9+9)
name old alloc/op new alloc/op delta
Marshal 10.7kB ± 0% 9.2kB ± 0% -13.96% (p=0.000 n=10+10)
name old allocs/op new allocs/op delta
Marshal 444 ± 0% 366 ± 0% -17.57% (p=0.000 n=10+10)
Change-Id: I9e727defa23f7e5fc684f246de0136fe28cf8d25
Reviewed-on: https://go-review.googlesource.com/c/go/+/231738
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
We could instead fix cmd/test2json to treat PAUSE lines as *not*
changing the active test name, but that seems like it would be more
confusing to humans, and also wouldn't fix tools that parse output
using existing builds of cmd/test2json.
Fixes#40657
Change-Id: I937611778f5b1e7dd1d6e9f44424d7e725a589ed
Reviewed-on: https://go-review.googlesource.com/c/go/+/248727
Run-TryBot: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Jean de Klerk <deklerk@google.com>
There were two places where the -short flag was added in order to
speed up tests when run in short mode, in CL 178399 and CL 177417.
It appears viable to re-use the GO_TEST_SHORT value so that -short
flag is not used when the tests are executed on a longtest builder,
where it is not a goal to skip slow tests for improved performance.
Do so, in order to make the testing configurations simpler and more
predictable.
Factor out the flag name out of the string returned by short, so that
it can be used in context of 'go test' which can accept a -short flag,
and a test binary which requires the use of a longer -test.short flag.
For #39054.
For #29252.
Change-Id: I52dfbef73cc8307735c52e2ebaa609305fb05933
Reviewed-on: https://go-review.googlesource.com/c/go/+/233898
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
When external linking, for large binaries, the external linker
may insert a trampoline for the write barrier call, which looks
0000000005a98cc8 <__long_branch_runtime.gcWriteBarrier>:
5a98cc8: 86 01 82 3d addis r12,r2,390
5a98ccc: d8 bd 8c e9 ld r12,-16936(r12)
5a98cd0: a6 03 89 7d mtctr r12
5a98cd4: 20 04 80 4e bctr
It clobbers R12 (and CTR, which is never live across a call).
As at compile time we don't know whether the binary is big and
what link mode will be used, I think we need to mark R12 as
clobbered for write barrier call. For extra safety (future-proof)
we mark caller-saved register that cannot be used for function
arguments, which includes R11, as potentially clobbered as well.
Fixes#40851.
Change-Id: Iedd901c5072f1127cc59b0a48cfeb4aaec81b519
Reviewed-on: https://go-review.googlesource.com/c/go/+/248917
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
This reverts the following changes:
• cmd/go: add tracing for querying and downloading from the proxy
CL 242786, commit 1a35583418
• cmd/go: do context propagation for tracing downloads
CL 248327, commit c0cf190d22
• cmd/go/internal: remove some users of par.Work
CL 248326, commit f30044a03b
Reason for revert: broke linux 386 and amd64 longtest builders.
The problem started with CL 248326, but CL 248327 and CL 242786
are reverted as well due to conflicts.
Updates #38714.
Fixes#40861.
Change-Id: I68496b4e5a27e47a42183553c3a645b288edac83
Reviewed-on: https://go-review.googlesource.com/c/go/+/249017
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Carlos Amedee <carlos@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Some of the existing optimizations aren't triggered because they
are handled by the generic rules so this CL removes them. Also
some constraints were copied without much thought from the amd64
rules and they don't make sense on s390x, so we remove those
constraints.
Finally, add a 'multiply by the sum of two powers of two'
optimization. This makes sense on s390x as shifts are low latency
and can also sometimes be optimized further (especially if we add
support for RISBG instructions).
name old time/op new time/op delta
IntMulByConst/3-8 1.70ns ±11% 1.10ns ± 5% -35.26% (p=0.000 n=10+10)
IntMulByConst/5-8 1.64ns ± 7% 1.10ns ± 4% -32.94% (p=0.000 n=10+9)
IntMulByConst/12-8 1.65ns ± 6% 1.20ns ± 4% -27.16% (p=0.000 n=10+9)
IntMulByConst/120-8 1.66ns ± 4% 1.22ns ±13% -26.43% (p=0.000 n=10+10)
IntMulByConst/-120-8 1.65ns ± 7% 1.19ns ± 4% -28.06% (p=0.000 n=9+10)
IntMulByConst/65537-8 0.86ns ± 9% 1.12ns ±12% +30.41% (p=0.000 n=10+10)
IntMulByConst/65538-8 1.65ns ± 5% 1.23ns ± 5% -25.11% (p=0.000 n=10+10)
Change-Id: Ib196e6bff1e97febfd266134d0a2b2a62897989f
Reviewed-on: https://go-review.googlesource.com/c/go/+/248937
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Clean up some pclntab state, specifically:
1) Remove the oldPclnState type.
2) Move a structure out of pclnState, that was holding some memory.
3) Stop passing container around everywhere and calling emitPcln. Use a
slice of function symbols instead.
Change-Id: I74e916564cd769a706750d024e55ee0d811a79da
Reviewed-on: https://go-review.googlesource.com/c/go/+/248379
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Move the pctables out of pclntab_old. Creates a new generator symbol,
runtime.pctab, which holds all the deduplicated pctables. Also, tightens
up some of the types in runtime.
Darwin, cmd/compile statistics:
alloc/op
Pclntab_GC 26.4MB ± 0% 13.8MB ± 0%
allocs/op
Pclntab_GC 89.9k ± 0% 86.4k ± 0%
liveB
Pclntab_GC 25.5M ± 0% 24.2M ± 0%
No significant change in binary size.
Change-Id: I1560fd4421f8a210f8d4b508fbc54e1780e338f9
Reviewed-on: https://go-review.googlesource.com/c/go/+/248332
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
For an unsigned integer, it's useful to convert its order test with 0/1
to its equality test with 0. We can save a comparison instruction that
followed by a conditional branch on arm64 since it supports
compare-with-zero-and-branch instructions. For example,
if x > 0 { ... } else { ... }
the original version:
CMP $0, R0
BLS 9
the optimized version:
CBZ R0, 8
Updates #21439
Change-Id: Id1de6f865f6aa72c5d45b29f7894818857288425
Reviewed-on: https://go-review.googlesource.com/c/go/+/246857
Reviewed-by: Keith Randall <khr@golang.org>
Corrects ErrReader's signature to what was accepted in the approved
proposal, and also removes an exported ErrIO which wasn't part of
the proposal and is unnecessary.
The new signature allows users to customize their own errors.
While here, started examples, with ErrReader leading the way.
Updates #38781
Change-Id: Ia7f84721f11061343cfef8b1adc2b7b69bc3f43c
Reviewed-on: https://go-review.googlesource.com/c/go/+/248898
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
If the AND has other uses, we end up saving an argument to the AND
in another register, so we can use it for the TEST. No point in doing that.
Change-Id: I73444a6aeddd6f55e2328ce04d77c3e6cf4a83e0
Reviewed-on: https://go-review.googlesource.com/c/go/+/241280
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
I noticed that there is a Todo comment here. This variable is only used for filename when dump a function's ssa passes result in details. It is no problem to print a function alone, but may be edited by not only one goroutine if dump multiple functions at the same time. Although it looks only dump one function's ssa passes now. As far as I am concerned this variable can be a member variable of the struct Func. I'm not sure if this change is necessary. Looking forward to your advices, thank you very much.
Change-Id: I35dd7247889e0cc7f19c0b400b597206592dee75
Reviewed-on: https://go-review.googlesource.com/c/go/+/244918
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
There are some architecture-independent rules in #21439, since an
unsigned integer >= 0 is always true and < 0 is always false. This CL
adds these optimizations to generic rules.
Updates #21439
Change-Id: Iec7e3040b761ecb1e60908f764815fdd9bc62495
Reviewed-on: https://go-review.googlesource.com/c/go/+/246617
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This change does context propagation (and only context propagation)
necessary to add context to modfetch.Download and pkg.LoadImport.
This was done by adding context to their callers, and then
adding context to all call-sites, and then repeating adding
context to callers of those enclosing functions and their
callers until none were left. In some cases the call graph expansion
was pruned by using context.TODOs.
The next CL will add a span to Download. I kept it out of this
change to avoid making it any larger (and harder to review)
than it needs to be.
Updates #38714
Change-Id: I5bf2d599aafef67334c384dfccd5e255198c85b4
Reviewed-on: https://go-review.googlesource.com/c/go/+/248327
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
If we are parsing a test output, and the test does not end in the
usual PASS or FAIL line (say, because it panicked), then we need the
exit status of the test binary in order to determine whether the test
passed or failed. If we don't have that status available, we shouldn't
guess arbitrarily — instead, we should omit the final "pass" or "fail"
action entirely.
(In practice, we nearly always DO have the final status, such as when
running 'go test' or 'go tool test2json some.exe'.)
Fixes#40132
Change-Id: Iae482577361a6033395fe4a05d746b980e18c3de
Reviewed-on: https://go-review.googlesource.com/c/go/+/248624
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
par.Work is used in a number of places as a parallel
work queue. This change replaces it with goroutines
and channels in a number of simpler places where it's
used.
Change-Id: I0620eda46ec7b2c0599a8b9361639af7bb73a05a
Reviewed-on: https://go-review.googlesource.com/c/go/+/248326
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This change adds context, and a span to modload.LoadBuildList and
propagates context into modload.BuildList. It's the start
of a run of CLs to add trace spans for module operations.
Updates #38714
Change-Id: I0d58dd394051526338092dc9a5ec29a9e087e4e4
Reviewed-on: https://go-review.googlesource.com/c/go/+/248325
Run-TryBot: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
These commands are build-like commands that do their own flag
processing, so the value of debug-trace isn't available until
the command starts running. Start tracing in the cmd's run
function.
Updates #38714
Change-Id: I4d633e6ee907bf09feac52c2aff3daceb9b20e12
Reviewed-on: https://go-review.googlesource.com/c/go/+/248324
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This could help make it easier to identify blocking
dependencies when examining traces. Flows can be turned
off when viewing traces to remove potential distractions.
Updates #38714
Change-Id: Ibfd3f1a1861e3cac31addb053a2fca7ee796c4d7
Reviewed-on: https://go-review.googlesource.com/c/go/+/248322
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Action.Func is now a func(*Builder, context.Context, *Action), so that
contexts can be propagated into the action funcs. While context
is traditionally the first parameter of a function, it's the second
parameter of Action.Func's type to continue to allow for methods
on Builder to be used as functions taking a *Builder as the first
parameter. context.Context is instead the first parameter on
those functions.
Change-Id: I5f058d6a99a1e96fe2025f2e8ce30a033d12e935
Reviewed-on: https://go-review.googlesource.com/c/go/+/248321
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This change adds a trace event for each action and also
annotates each of the action execution goroutines with trace.Goroutine
so that the actions eaxecuted by each goroutine appear on different threads in
the chrome trace viewer.
Updates #38714
Change-Id: I2e58dc5606b2e3f7f87076a61e1cc6a2014255c5
Reviewed-on: https://go-review.googlesource.com/c/go/+/248320
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
trace.StartGoroutine will associate the trace information on the context
with a new chrome profiler thread id. The chrome profiler doesn't
expect multiple trace events to have the same thread id, so this
will allow us to display concurrent events on the trace.
Updates #38714
Change-Id: I888b0cce15a5a01db66366716fdd85bf86c832cd
Reviewed-on: https://go-review.googlesource.com/c/go/+/248319
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
On ARM, for a JMP/CALL relocation, the instruction bytes is
encoded in Reloc.Add (issue #19811). I really hate it, but before
it is fixed we have to follow the rule and emit the right bits
from r.Add.
Fixes#40769.
Change-Id: I862e105408d344c5cc58ca9140d2e552e4364453
Reviewed-on: https://go-review.googlesource.com/c/go/+/248399
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Joel Sing <joel@sing.id.au>
Reviewed-by: Than McIntosh <thanm@google.com>
cgo_import_dynamic pragma indicates a symbol is imported from a
dynamic library. Currently, the linker does not actually link
against the dynamic library, so we have to "force" it by using
//go:cgo_import_dynamic _ _ "dylib"
syntax, which links in the library unconditionally.
This CL changes it to link in the library automatically when a
symbol is imported from the library, without using the "force"
syntax. (The "force" syntax is still supported.)
Remove the unconditional imports in the runtime. Now,
Security.framework and CoreFoundation.framework are only linked
when the x509 package is imported (or otherwise specified).
Fixes#40727.
Change-Id: Ied36b1f621cdcc5dc4a8f497cdf1c554a182d0e0
Reviewed-on: https://go-review.googlesource.com/c/go/+/248333
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Currently, the GC stores the object marks for checkmarks mode in the
heap bitmap using a rather complex encoding: for one word objects, the
checkmark is stored in the pointer/scalar bit since one word objects
must be pointers; for larger objects, the checkmark is stored in what
would be the scan/dead bit for the second word of the object. This
encoding made more sense when the runtime used the first scan/dead bit
as the regular mark bit, but we moved away from that long ago.
This encoding and overloading of the heap bitmap bits causes a great
deal of complexity in many parts of the allocator and garbage
collector and leads to some subtle bugs like #15903.
This CL moves the checkmarks mark bits into their own per-arena bitmap
and reclaims the second scan/dead bit as a regular scan/dead bit.
I tested this by enabling doubleCheck mode in heapBitsSetType and
running in both regular and GODEBUG=gccheckmark=1 mode.
Fixes#15903.
No performance degradation. (Very slight improvement on a few
benchmarks, but it's probably just noise.)
name old time/op new time/op delta
BiogoIgor 16.6s ± 1% 16.4s ± 1% -0.94% (p=0.000 n=25+24)
BiogoKrishna 19.2s ± 3% 19.2s ± 3% ~ (p=0.638 n=23+25)
BleveIndexBatch100 6.12s ± 5% 6.17s ± 4% ~ (p=0.170 n=25+25)
CompileTemplate 206ms ± 1% 205ms ± 1% -0.43% (p=0.005 n=24+24)
CompileUnicode 82.2ms ± 2% 81.5ms ± 2% -0.95% (p=0.001 n=22+22)
CompileGoTypes 755ms ± 3% 754ms ± 4% ~ (p=0.715 n=25+25)
CompileCompiler 3.73s ± 1% 3.73s ± 1% ~ (p=0.445 n=25+24)
CompileSSA 8.67s ± 1% 8.66s ± 1% ~ (p=0.836 n=24+22)
CompileFlate 134ms ± 2% 133ms ± 1% -0.66% (p=0.001 n=24+23)
CompileGoParser 164ms ± 1% 163ms ± 1% -0.85% (p=0.000 n=24+24)
CompileReflect 466ms ± 5% 466ms ± 3% ~ (p=0.863 n=25+25)
CompileTar 182ms ± 1% 182ms ± 1% -0.31% (p=0.048 n=24+24)
CompileXML 249ms ± 1% 248ms ± 1% -0.32% (p=0.031 n=21+25)
CompileStdCmd 10.3s ± 1% 10.3s ± 1% ~ (p=0.459 n=23+23)
FoglemanFauxGLRenderRotateBoat 8.66s ± 1% 8.62s ± 1% -0.47% (p=0.000 n=23+24)
FoglemanPathTraceRenderGopherIter1 20.3s ± 3% 20.2s ± 2% ~ (p=0.893 n=25+25)
GopherLuaKNucleotide 29.7s ± 1% 29.8s ± 2% ~ (p=0.421 n=24+25)
MarkdownRenderXHTML 246ms ± 1% 247ms ± 1% ~ (p=0.558 n=25+24)
Tile38WithinCircle100kmRequest 779µs ± 4% 779µs ± 3% ~ (p=0.954 n=25+25)
Tile38IntersectsCircle100kmRequest 1.02ms ± 3% 1.01ms ± 4% ~ (p=0.658 n=25+25)
Tile38KNearestLimit100Request 984µs ± 4% 986µs ± 4% ~ (p=0.627 n=24+25)
[Geo mean] 552ms 551ms -0.19%
https://perf.golang.org/search?q=upload:20200723.6
Change-Id: Ic703f26a83fb034941dc6f4788fc997d56890dec
Reviewed-on: https://go-review.googlesource.com/c/go/+/244539
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Martin Möhrmann <moehrmann@google.com>
The heapBitsSetType function has a slow doubleCheck debugging mode
that checks the bitmap written out by the rest of the function using
far more obvious logic. But even this has some surprisingly complex
logic in it. Simplify it a bit. This also happens to fix the logic on
32-bit.
Fixes#40335.
Change-Id: I5cee482ad8adbd01cf5b98e35a270fe941ba4940
Reviewed-on: https://go-review.googlesource.com/c/go/+/244538
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
The runtime has its own implementation of string indexing. To reduce
code duplication and cognitive load, replace this with calls to the
internal/bytealg package. We can't do this on Plan 9 because it needs
string indexing in a note handler (which isn't allowed to use the
optimized bytealg version because it uses SSE), so we can't just
eliminate the index function, but this CL does down-scope it so make
it clear it's only for note handlers on Plan 9.
Change-Id: Ie1a142678262048515c481e8c26313b80c5875df
Reviewed-on: https://go-review.googlesource.com/c/go/+/244537
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Since NetBSD 7, hw.ncpuonline reports the number of CPUs online, while
hw.cpu reports the number of CPUs configured. Try hw.cpuonline first and
fall back to hw.ncpu in case it fails (which is the case on NetBSD
before 7.0).
This follows the behavior on OpenBSD (see CL 161757). Also, Go
in pkgsrc is patched to use hw.cpuonline, so this CL would allow said
patch to be dropped.
Updates #30824
Change-Id: Id1c19dff2c1e4401e6074179fae7c708ba0e3098
Reviewed-on: https://go-review.googlesource.com/c/go/+/231957
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Benny Siegert <bsiegert@gmail.com>
Instead of returning a bad verb error format for runes above
utf8.Maxrune return a quoted utf8.RuneError rune (\ufffd).
This makes the behaviour consistent with the "c" verb and
aligns behaviour to not return bad verb error format when
a verb is applied to the correct argument type.
Fixes#14569
Change-Id: I679485f6bb90ebe408423ab68af16cce38816cd0
Reviewed-on: https://go-review.googlesource.com/c/go/+/248759
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
The two existing benchmarks encode randomized pixels, which isn't very
representative. The two new benchmarks encode a PNG photo as a GIF.
Also rename the benchmarks for consistency.
Also fix the bytes-per-op measure for paletted images, which are 1 (not
4) bytes per pixel.
Also simplify BenchmarkEncodeRandomPaletted (formerly just called
BenchmarkEncode). It doesn't need to generate a random palette (and the
GIF encoder largely doesn't care about the palette's RGBA values).
Use palette.Plan9 instead, a pre-existing 256-element color palette.
Change-Id: I10a6ea4e9590bb0d9f76e8cc0f4a88d43b1d650d
Reviewed-on: https://go-review.googlesource.com/c/go/+/248218
Reviewed-by: David Symonds <dsymonds@golang.org>
There is an optimization rule that removes calls to racefuncenter and
racefuncexit, if there are no other race calls in the function. The
rule removes the call to racefuncenter, but it does *not* remove the
store of its argument to the outargs section of the frame. If the
outargs section is now size 0 (because the calls to racefuncenter/exit
were the only calls), then that argument store clobbers the frame
pointer instead.
The fix is to remove the argument store when removing the call to
racefuncenter. (Racefuncexit doesn't have an argument.)
Change-Id: I183ec4d92bbb4920200e1be27b7b8f66b89a2a0a
Reviewed-on: https://go-review.googlesource.com/c/go/+/248262
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
These operations are async unsafe on architectures that use
frame pointers.
The reason is they rely on data being safe when stored below the stack
pointer. They do:
45da69: 48 89 6c 24 f0 mov %rbp,-0x10(%rsp)
45da6e: 48 8d 6c 24 f0 lea -0x10(%rsp),%rbp
45da73: e8 7d d0 ff ff callq 45aaf5 <runtime.duffzero+0x115>
45da78: 48 8b 6d 00 mov 0x0(%rbp),%rbp
This dance ensures that inside duffzero, it looks like there is a
proper frame pointer set up, so that stack walkbacks work correctly if
the kernel samples during duffzero.
However, this instruction sequence depends on data not being clobbered
even though it is below the stack pointer.
If there is an async interrupt at any of those last 3 instructions,
and the interrupt decides to insert a call to asyncPreempt, then the
saved frame pointer on the stack gets clobbered. The last instruction
above then restores junk to the frame pointer.
To prevent this, mark these instructions as async unsafe.
(The body of duffzero is already async unsafe, as it is in package runtime.)
Change-Id: I5562e82f9f5bd2fb543dcf2b6b9133d87ff83032
Reviewed-on: https://go-review.googlesource.com/c/go/+/248261
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Martin Möhrmann <moehrmann@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The default ctrl+c handler should process exits in situations where it
makes sense, like console apps, but not in situations where it doesn't,
like libraries or services. Therefore, we should remove the exit(2) so
that the default handler is used for this. This also uses the more
proper windows exit code of STATUS_CONTROL_C_EXIT, with the base case
handler installed by KernelBase.dll. In particular, this helps in the
case of services, which previously would terminate when receiving
shutdown signals, instead of passing them onward to the service program.
In this CL, contrary to CL 244959, we do not need to special case
services with expensive detection algorithms, or rely on hard-coded
library/archive flags.
Fixes#40167.
Fixes#40074.
Change-Id: I9bf6ed6f65cefeff754d270aa33fa4df8d0b451f
Reviewed-on: https://go-review.googlesource.com/c/go/+/243597
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
Currently, generated struct wrapper for closure is not handled in
mustHeapAlloc. That causes compiler crashes when the wrapper struct
is too large for stack, and must be heap allocated instead.
Fixes#39292
Change-Id: I14c1e591681d9d92317bb2396d6cf5207aa93e08
Reviewed-on: https://go-review.googlesource.com/c/go/+/244917
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
CL 230121 fixed the bug that struct literal blank fields type array/struct
can not be initialized. But it still misses some cases when an expression
causes "candiscard(value)" return false. When these happen, we recursively
call fixedlit with "var_" set to "_", and hit the bug again.
To fix it, just making splitnode return "nblank" whenever "var_" is "nblank".
Fixes#38905
Change-Id: I281941b388acbd551a4d8ca1a235477f8d26fb6e
Reviewed-on: https://go-review.googlesource.com/c/go/+/232617
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Simplifying some code without compromising performance.
My CPU is Intel Xeon Gold 6161, 2.20GHz, 64-bit operating system.
The memory is 8GB. This is my test environment, I hope to help you judge.
Benchmark:
name old time/op new time/op delta
Log1p-4 21.8ns ± 5% 21.8ns ± 4% ~ (p=0.973 n=20+20)
Change-Id: Icd8f96f1325b00007602d114300b92d4c57de409
Reviewed-on: https://go-review.googlesource.com/c/go/+/233940
Reviewed-by: Robert Griesemer <gri@golang.org>
If the timeout triggers before writing to the done channel, the
goroutine will be blocked waiting for a corresponding read that’s
no longer existent, thus a goroutine leak. This change fixes that by
using a buffered channel instead.
Change-Id: I9cf4067a58bc5a729ab31e4426edd78bd359e8e0
GitHub-Last-Rev: a7d811a7be
GitHub-Pull-Request: golang/go#40236
Reviewed-on: https://go-review.googlesource.com/c/go/+/242902
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Adds parentheses so as to properly bind <- to the right most
channel.
This meant that previously given:
ChanOf(<-chan T)
it would mistakenly try to look up the type as
chan <-chan T
instead of
chan (<-chan T)
Fixes#39897
Change-Id: I8564916055f5fadde3382e41fe8820a1071e5f13
GitHub-Last-Rev: f8f2abe8d4
GitHub-Pull-Request: golang/go#39898
Reviewed-on: https://go-review.googlesource.com/c/go/+/240280
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Taking the live variable set from the last return point is problematic.
See #40629 for details, but there may not be a return point, or it may
be before the final defer.
Additionally, keeping track of the last call as a *Value doesn't quite
work. If it is dead-code eliminated, the storage for the Value is reused
for some other random instruction. Its live variable information,
if it is available at all, is wrong.
Instead, just mark all the open-defer argument slots as live
throughout the function. (They are already zero-initialized.)
Fixes#40629
Change-Id: Ie456c7db3082d0de57eaa5234a0f32525a1cce13
Reviewed-on: https://go-review.googlesource.com/c/go/+/247522
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
modload.Init initialized the build cache with the intent of providing
a better error message in Go 1.12, when the build cache became
mandatory (in module mode, packages aren't installed outside the build
cache). Unfortunately, this didn't provide a more descriptive error
(the cache calls base.Fatalf with its own message), and it caused
errors for commands that don't use the cache (like 'go mod edit').
This CL removes the cache initialization from modload.Init. The
builder will initialize it when it's needed.
For #39882
Change-Id: Ibc01ae4e59358dcd08a07ffc97bf556514d0366f
Reviewed-on: https://go-review.googlesource.com/c/go/+/240548
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
In CL 239797, str.GlobsMatchPath was copied to golang.org/x/mod/module
as MatchPrefixPatterns. This CL updates x/mod, switches calls to use
the new function, and deletes the old function.
For #38725
Change-Id: I7241032228b574aa539426a92d2f5aad9ee001e2
Reviewed-on: https://go-review.googlesource.com/c/go/+/240061
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
modfetch.WriteGoSum now accepts a map[module.Version]bool parameter.
This is used to prevent some new sums from being saved to go.sum when
they would be removed by the next 'go mod tidy'. Previusly, sums were
saved for modules looked up during import resolution.
A new function, modload.TrimGoSum, is also introduced, which marks
sums for deletion. 'go mod tidy' now uses this. The new logic
distinguishes between go.mod sums and content sums, which lets 'go mod
tidy' delete sums for modules in the build graph but not the build
list.
Fixes#31580Fixes#36260Fixes#33008
Change-Id: I06c4125704a8bbc9969de05265967ec1d2e6d3e8
Reviewed-on: https://go-review.googlesource.com/c/go/+/237017
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The Builder isn't needed by MkEnv, and Builder.Init doesn't have side
effects that change the environment. Builder.Init does currently call
CheckGOOSARCHPair, but that's being moved out in CL 234658.
Builder.Init creates the temporary work directory used by the
builder. For the builder created in MkEnv, this directory is never
used. Creating this directory can cause unnecessary errors for
commands that don't use a builder like 'go clean' and 'go list'.
Fixes#38395
Updates #24398
Change-Id: Ib93ae55afdf958000470657f4c4ff5bd92700e46
Reviewed-on: https://go-review.googlesource.com/c/go/+/236563
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The prior implementation used the format verb %v which unfortunately
improperly wrapped any underlying scanner errors, and we couldn't use
errors.Is nor errors.As. This change fixes that by using the %w verb.
Added a unit to ensure that both error sub string matching works, but
also that errors.Is works as expected.
Fixes#38099
Change-Id: Iea667041dd8081d961246f77f2542330417292dc
Reviewed-on: https://go-review.googlesource.com/c/go/+/248337
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Daniel Theophanes <kardianos@gmail.com>
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The optimization that replaces inline markers with pre-existing
instructions assumes that 'Prog' values produced by the compiler are
still reachable after the assembler has run. This was not true on
s390x where the assembler was removing NOP instructions from the
linked list of 'Prog' values. This led to broken inlining data
which in turn caused an infinite loop in the runtime traceback code.
Fix this by stopping the s390x assembler backend removing NOP
values. It does not make any difference to the output of the
assembler because NOP instructions are 0 bytes long anyway.
Fixes#40473.
Change-Id: Ib4fabadd1de8adb80421f75950ee9aad2111147a
Reviewed-on: https://go-review.googlesource.com/c/go/+/247697
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Switch pcdata over to content addressable symbols. This is the last
step before removing these from pclntab_old.
No meaningful benchmarks changes come from this work.
Change-Id: I3f74f3d6026a278babe437c8010e22992c92bd89
Reviewed-on: https://go-review.googlesource.com/c/go/+/247399
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
CL 246497 introduced bugs in gosym that the long tests caught. These two
bugs were:
1) In 1.16, 0 is now a valid file number from pcfile tables.
2) Also, in 1.16, when we scan all functions looking for a pc/file pair,
the values returned from pcfile are no longer the direct offset into
the file table. Rather, the values from pcfile are the offset into
the cu->file look-up table.
This CL fixes those two issues.
Change-Id: I0cd280bdcaeda89faaf9fac41809abdb87734499
Reviewed-on: https://go-review.googlesource.com/c/go/+/248317
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Previously, the assembler removed NOPs from the Prog list in
obj9.go. NOPs shouldn't be removed if they were added as
an inline mark, as described in the issue below.
Fixes#40689
Once the NOPs were left in the Prog list, some instructions
were flagged as invalid because they had an operand which was
not represented in optab. In order to preserve the previous
assembler behavior, entries were added to optab for those
operand cases. They were not flagged as errors before because
the NOP instructions were removed before the code to check the
valid opcode/operand combinations.
Change-Id: Iae5145f94459027cf458e914d7c5d6089807ccf8
Reviewed-on: https://go-review.googlesource.com/c/go/+/247842
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Paul Murphy <murp@ibm.com>
Reviewed-by: Michael Munday <mike.munday@ibm.com>
Reviewed-by: Keith Randall <khr@golang.org>
Minor cleanup: remove the symbol attribute AttrSeenGlobal, since it is
redundant with the existing attribute AttrOnList (no need to have what
amounts to a separate flag for checking the same property).
Change-Id: Ia269b64de37c2bb4a2314bbecf3d2091c6d57424
Reviewed-on: https://go-review.googlesource.com/c/go/+/239477
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Add a draft version of a blurb on improvements to the linker. This
will need to be finalized later in the release since there are still
some additional changes to be made to the linker in 1.16.
Updates #40703.
Change-Id: Id85c7e129071cc2faacb09c53a2968bd52b0a7b4
Reviewed-on: https://go-review.googlesource.com/c/go/+/248238
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Go 1.14 included a (rather awful) workaround for a Linux kernel bug
that corrupted vector registers on x86 CPUs during signal delivery
(https://bugzilla.kernel.org/show_bug.cgi?id=205663). This bug was
introduced in Linux 5.2 and fixed in 5.3.15, 5.4.2 and all 5.5 and
later kernels. The fix was also back-ported by major distros. This
workaround was necessary, but had unfortunate downsides, including
causing Go programs to exceed the mlock ulimit in many configurations
(#37436).
We're reasonably confident that by the Go 1.16 release, the number of
systems running affected kernels will be vanishingly small. Hence,
this CL removes this workaround.
This effectively reverts CLs 209597 (version parser), 209899 (mlock
top of signal stack), 210299 (better failure message), 223121 (soft
mlock failure handling), and 244059 (special-case patched Ubuntu
kernels). The one thing we keep is the osArchInit function. It's empty
everywhere now, but is a reasonable hook to have.
Updates #35326, #35777 (the original register corruption bugs).
Updates #40184 (request to revert in 1.15).
Fixes#35979.
Change-Id: Ie213270837095576f1f3ef46bf3de187dc486c50
Reviewed-on: https://go-review.googlesource.com/c/go/+/246200
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
In Go 1.15 we switched the default linking mode for PIE on
Linux/AMD64 and Linux/ARM64 to internal linking. Clarify that
the previous behavior (external linking) can still be used with
a flag.
Fixes#40719.
Change-Id: Ib7042622bc91e1b1aa31f520990d03b5eb6c56bb
Reviewed-on: https://go-review.googlesource.com/c/go/+/248199
Reviewed-by: Ian Lance Taylor <iant@golang.org>
In the dev.link branch we have continued developing the new object
file format support and the linker improvements described in
https://golang.org/s/better-linker . Since the last merge (May 1st
2020), more progress has been made to improve the new linker, with
improvements on both linker speed and memory usage.
Fixes#40703.
Change-Id: I9924ea88d981845c3a40ec8c25820120fc21c003
This change propagates context into PackagesForErrors and Do for
the purpose of tracing, and calls trace.StartSpan on PackagesForErrors
and Do, so that the trace now shows the broad outline of where
the "Loading" and "Execution" phases are in the build.
Updates #38714
Change-Id: Ib9a7cf7030210f68f76663d1c8a7461e0a226611
Reviewed-on: https://go-review.googlesource.com/c/go/+/238541
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The ViewerEvent, ViewerData and ViewerFrame structs are moved into
cmd/internal/traceviewer, and renamed Event, Data, and Frame.
The structs are the same, except for the following: A definition
for the JSON "bp" field that's defined in the trace format, but
missing in the structs has been added. Also, the Tid and Pid fields
on Event have been renamed TID and PID to better match Go style.
Finally, the footer field on ViewerData, which hasn't been used
for a while, has been removed.
This CL is in preparation for the usage of these structs by cmd/go's
tracing functionality.
Updates #38714
Change-Id: I345f23617b96d4629b876ae717f89d56a67e05a3
Reviewed-on: https://go-review.googlesource.com/c/go/+/239098
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Jay Conrod <jayconrod@google.com>
While working on deduplicating pcdata, I found that the following hashed
symbols would result in the same:
[] == [0,0,0,0....]
This makes using content addressable symbols untenable for pcdata.
Adding the length to the hash keeps the dream alive.
No difference in binary size (darwin, cmd/compile), spurious
improvements in DWARF phase memory.
Change-Id: I21101f7754a3d870922b0dea39c947cc8509432f
Reviewed-on: https://go-review.googlesource.com/c/go/+/247903
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Creates two new symbols: runtime.cutab, and runtime.filetab, and strips
the filenames out of runtime.pclntab_old.
All stats are for cmd/compile.
Time:
Pclntab_GC 48.2ms ± 3% 45.5ms ± 9% -5.47% (p=0.004 n=9+9)
Alloc/op:
Pclntab_GC 30.0MB ± 0% 29.5MB ± 0% -1.88% (p=0.000 n=10+10)
Allocs/op:
Pclntab_GC 90.4k ± 0% 73.1k ± 0% -19.11% (p=0.000 n=10+10)
live-B:
Pclntab_GC 29.1M ± 0% 29.2M ± 0% +0.10% (p=0.000 n=10+10)
binary sizes:
NEW: 18565600
OLD: 18532768
The size differences in the binary are caused by the increased size of
the Func objects, and (less likely) some extra alignment padding needed
as a result. This is probably the maximum increase in size we'll size
from the pclntab reworking.
Change-Id: Idd95a9b159fea46f7701cfe6506813b88257fbea
Reviewed-on: https://go-review.googlesource.com/c/go/+/246497
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Austin Clements <austin@google.com>
This template is based on CL 220278 and previous ones like it.
Include Compiler and Linker sections proactively, they can be
removed if they don't end up being needed for Go 1.16.
Use two spaces of indentation for TODOs to set a better precedent
for the final text that will take its place.
'relnote -html' does not report any changes at this time.
For #40700.
Change-Id: I096b0ce0d33aaaa6fae9c91c0d2dfb89b9c5e94c
Reviewed-on: https://go-review.googlesource.com/c/go/+/248198
Reviewed-by: Carlos Amedee <carlos@golang.org>
This is the start of the Go 1.16 development cycle, so update the
Version value accordingly. It represents the Go 1.x version that
will soon open up for development (and eventually become released).
Historically, we used to bump this at an arbitrary time throughout
the development cycle, but it's better to be more predictable about
updating it. The start of a development cycle should be the most
appropriate time: it clearly marks the boundary between 1.15 and
1.16 development, and doing it early can help catch issues in other
tooling. See issue #38704 for more background.
There is no longer a need to update the list of Go versions in
src/go/build/doc.go because it does not exist as of CL 232981.
For #40705.
Updates #38704.
Updates #37018.
Change-Id: Id8ee733b5e79c53b6cd03509c6560614d8743833
Reviewed-on: https://go-review.googlesource.com/c/go/+/248038
Reviewed-by: Carlos Amedee <carlos@golang.org>
CL 245485 introduced a map for used files in a function. When
numbering symbols, make sure we traverse the files in
deterministic order.
Should fix longtest builders.
Change-Id: I1006bc5425116ab40e33a61e8f5acd1bdb4abad9
Reviewed-on: https://go-review.googlesource.com/c/go/+/247997
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Read Go object files using cmd/internal/goobj2 package directly,
instead of using cmd/internal/goobj as an intermediate layer.
Now cmd/internal/archive is only about reading archives.
Change-Id: Ifecb217fb26c16c26fc1bbc3fba0ed44710020ed
Reviewed-on: https://go-review.googlesource.com/c/go/+/246443
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Rename cmd/internal/goobj package to cmd/internal/archive. This
is in preparation of a refactoring of object and archive file
reading packages.
With this CL, the cmd/internal/archive contains logic about
reading Go object files. This will be moved to other places in
later CLs.
Change-Id: Ided7287492a4766183d6e49be840a7f361504d1d
Reviewed-on: https://go-review.googlesource.com/c/go/+/246442
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
This update was created using the updatecontrib command:
go get golang.org/x/build/cmd/updatecontrib
cd gotip
updatecontrib
With manual changes based on publicly available information
to canonicalize letter case and formatting for a few names.
For #12042.
Change-Id: I66dc5ee28d9a64bc9d150e72d136d8f71e50373b
Reviewed-on: https://go-review.googlesource.com/c/go/+/247767
Reviewed-by: Carlos Amedee <carlos@golang.org>
Reviewed-by: Andrew Bonventre <andybons@golang.org>
Currently, nanotime1 (and walltime1) is not reentrant, in that it
sets m.vdsoSP at entry and clears it at exit. If a signal lands
in between, and nanotime1 is called from the signal handler, it
will clear m.vdsoSP while we are still in nanotime1. If (in the
unlikely event) it is signaled again, m.vdsoSP will be wrong,
which may cause the stack unwinding code to crash.
This CL makes it reentrant, by saving/restoring the previous
vdsoPC and vdsoSP, instead of setting it to 0 at exit.
TODO: have some way to test?
Change-Id: I9ee53b251f1d8a5a489c71d4b4c0df1dee70c3e5
Reviewed-on: https://go-review.googlesource.com/c/go/+/246763
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
In order to prevent renumbering of filenames in pclntab generation, use
the per-package file list (previously only used for DWARF generation) as
file-indices. This is the largest step to eliminate renumbering of
filenames in pclntab.
Note, this is probably not the final state of the file table within the
object file. In this form, the linker loads all filenames for all
objects. I'll move to storing the filenames as regular string
symbols,and defaulting all string symbols to using the larger hash value
to make generation of pcln simplest, and most memory friendly.
Change-Id: I23daafa3f4b4535076e23100200ae0e7163aafe0
Reviewed-on: https://go-review.googlesource.com/c/go/+/245485
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Add a new PPC64-only linker test that does a build with the
-debugppc64textsize debugging option (selecting a lower the threshold
for text section splitting) to verify that no bugs have been
introduced in the linker code that manages this process.
Change-Id: Iea3f16a04c894d528eab2cb52f1ec1d75a2770cc
Reviewed-on: https://go-review.googlesource.com/c/go/+/241499
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Add a new debugging command line option (-debugppc64textsize=N) that
forces the start of a new text section after ".text" hits N bytes as
opposed to the architected limit of 2^26. This is intended to enable
testing of the linker code paths that handle multiple .text sections
on PPC64 without resorting to building giant applications.
Updates #20492.
Change-Id: I74ab7fd1e412e9124de5bd0d8d248c5e73225ae3
Reviewed-on: https://go-review.googlesource.com/c/go/+/241073
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The code in the the linker's genelfsym() routine was not properly
including runtime.text.%d marker symbols that are emitted on PPC64
when a very large text section is split into chunks. This bug was
introduced in CL 233338 when portions of asmb2() were converted
from sym.Symbol to loader.Sym usage.
Change-Id: Idfed944c41e1805f78f35be67bcdd18bdefd7819
Reviewed-on: https://go-review.googlesource.com/c/go/+/241498
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
On AIX, container symbols are handled in a weird way (unlike
other platforms): the outer symbol needs to have size (but still
no data), and the inner symbols must not be in the symbol table
(otherwise it overlaps with the outer symbol, which the system
linker doesn't like).
As of CL 241598, pclntab becomes a container symbol. We need to
follow the rule above for AIX.
Fix AIX build.
Change-Id: Ie2515a4cabbd8cf3f6d3868643a28f64ca3365a2
Reviewed-on: https://go-review.googlesource.com/c/go/+/246479
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Currently, at compile time, for each itab symbol, we create an
"itablink" symbol which holds solely the address of the itab
symbol. At link time, all the itablink symbols are grouped
together to form the itablinks slice.
This CL removes the itablink symbols, and directly generate the
itablinks slice in the linker. This removes a number of symbols,
which are dupOK and generally have long names. And also removes
a special handling of itablink symbols in the deadcode pass which
iterates through all symbols.
Change-Id: I475c3c8899e9fbeec9abc7647b1e4a69aa5c3c5a
Reviewed-on: https://go-review.googlesource.com/c/go/+/245901
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The hash maps are used to deduplicate hashed symbols. Once we
loaded all the symbols, we no longer need the hash maps. Drop
them.
Linking cmd/compile,
name old live-B new live-B delta
Loadlib_GC 13.1M ± 0% 11.3M ± 0% -13.62% (p=0.008 n=5+5)
Change-Id: I4bb1f84e1111a56d9e777cd6a68f7d974b60e321
Reviewed-on: https://go-review.googlesource.com/c/go/+/245721
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Extend the content-addressable symbol mechanism to itab symbols.
Itab symbols require global uniqueness (as at run time we compare
pointers), so it needs to be reliably deduplicated. Currently the
content hash depends on symbol name expansion, so we can only do
this when all Go packages are built with know package paths. Fall
back to checking names if any Go package is built with unknown
package path.
Change-Id: Icf5e8873755050c20e5fc6549f6de1c883254c89
Reviewed-on: https://go-review.googlesource.com/c/go/+/245719
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The type descriptor symbol of a defined (named) type (and pointer
to it) is defined only in the package that defines the type. It
is not dupOK, unlike other type descriptors. So it can be
referenced by index. Currently it is referenced by name for
cross-package references, because the index is not exported and
so not known to the referencing package.
This CL passes the index through the export data, so the symbol
can be referenced by index, and does not need to be looked up by
name. This also makes such symbol references consistent: it is
referenced by index within the defining package and also cross-
package, which makes it easier for content hashing (in later CLs).
One complication is that we need to set flags on referenced
symbols (specifically, the UsedInIface flag). Before, they are
non-package refs, which naturally carry flags in the object file.
For indexed refs, we currently don't put their flags in the
object file. Introduce a new block for this.
Change-Id: I8126f8e318ac4e6609eb2ac136201fd6c264c256
Reviewed-on: https://go-review.googlesource.com/c/go/+/245718
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Leaving creation of the funcID till the linker requires the linker to
load the function and file names into memory. Moving these into the
compiler/assembler prevents this.
This work is a step towards moving all func metadata into the compiler.
Change-Id: Iebffdc5a909adbd03ac263fde3f4c3d492fb1eac
Reviewed-on: https://go-review.googlesource.com/c/go/+/244024
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Austin Clements <austin@google.com>
For dupOK symbols, their attributes should be OR'd. Most of the
attributes are expected to be set consistently across multiple
definitions, but UsedInIface must be OR'd, and for alignment we
need to pick the largest one. Currently the attributes are not
always OR'd, depending on addSym returning true or false. This
doesn't cause any real problem, but it would be a problem if we
make type descriptor symbols content-addressable.
This CL removes the second result of addSym, and lets preloadSyms
always set the attributes. Also removes the alignment handling on
addSym, handles it in preloadSyms only.
Change-Id: I06b3f0adb733f6681956ea9ef54736baa86ae7bc
Reviewed-on: https://go-review.googlesource.com/c/go/+/245720
Reviewed-by: Jeremy Faller <jeremy@golang.org>
We're reworking pclntab generation in the linker, and with that we're
moving FuncID generation in to the compiler. Determining the FuncID is
done by a lookup on the package.function name; therefore, we need the
package whenever we make the TEXT symbols.
Change-Id: I805445ffbf2f895f06ce3a91fb09126d012bf86e
Reviewed-on: https://go-review.googlesource.com/c/go/+/245318
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Move the function names out of runtime.pclntab_old, creating
runtime.funcnametab. There is an unfortunate artifact in this change in
that calculating the funcID still requires loading the name. Future work
will likely pull this out and put it into the object file Funcs.
ls -l cmd/compile (darwin):
before: 18524016
after: 18519952
The difference in size can be attributed to alignment in pclntab_old.
Change-Id: Ibcbb230d4632178f8fcd0667165f5335786381f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/243223
Reviewed-by: Austin Clements <austin@google.com>
Non functional change.
As runtime.pclntab breaks up, it'll be easier if we can just pass around
the pclntab state. Also, eliminate the globals in pclntab.
Change-Id: I2a5849e8f5f422a336a881e53a261e3997d11c44
Reviewed-on: https://go-review.googlesource.com/c/go/+/242599
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
As of July 2020, a fair amount of the new linker's live memory, and
runtime is spent generating pclntab. In an effort to streamline that
code, this change starts breaking up the generation of runtime.pclntab
into smaller chunks that can run later in a link. These changes are
described in an (as yet not widely distributed) document that lays out
an improved format. Largely the work consists of breaking up
runtime.pclntab into smaller pieces, stopping much of the data
rewriting, and getting runtime.pclntab into a form where we can reason
about its size and look to shrink it. This change is the first part of
that work -- just pulling out the header, and demonstrating where a
majority of that work will be.
Change-Id: I65618d0d0c780f7e5977c9df4abdbd1696fedfcb
Reviewed-on: https://go-review.googlesource.com/c/go/+/241598
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Austin Clements <austin@google.com>
The symbol's data in the object file (sym.P) may already not
contain trailing zeros (e,g, for [10]int{1}), but sometimes it
does (e.g. for [10]int{1,0}). The linker can already handle this
case. We just always trim the trailing zeros for content hashing,
so it can deduplicate [10]int{1} and [10]int{1,0}.
Note: in theory we could just trim the zeros in the symbol data
as well. But currently the linker depends on reading symbol data
for certain symbols (e.g. type symbol decoding), and trimming
will complicates things in the linker.
Change-Id: I9e90e41e6ac808b36855b0713a85e61c33bf093a
Reviewed-on: https://go-review.googlesource.com/c/go/+/245717
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
We have Reloc and Reloc2. Reloc2 is the better approach and most
code uses Reloc2. There are still uses of Reloc. This CL migrates
them to Reloc2, and removes Reloc.
Change-Id: Id5f6a6019e1e044add682d05e70ebb1548ec58d9
Reviewed-on: https://go-review.googlesource.com/c/go/+/245577
Reviewed-by: Jeremy Faller <jeremy@golang.org>
We used to generate all external relocations in memory, then emit
the relocation records at a later pass. The data structures were
chosen so that it takes as little memory as possible. Now we just
stream out external relocations, and ExtReloc is just a local
variable. Change the data structure to avoid repeated read of
some fields. Also get rid of ExtRelocView, as it is no longer
necessary.
Change-Id: I40209bbe4387af231b29788125c3b4ebb0ff4a33
Reviewed-on: https://go-review.googlesource.com/c/go/+/245479
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
We accidentally passed the address of a local to a function
pointer, where we should pass the address of a global.
Linking cmd/compile with external linking:
Asmb2_GC 32.5ms ± 5% 21.6ms ± 3% -33.57% (p=0.016 n=5+4)
Asmb2_GC 29.2MB ± 0% 6.4MB ± 0% -78.20% (p=0.008 n=5+5)
Asmb2_GC 1.43M ± 0% 0.00M ± 4% -99.98% (p=0.008 n=5+5)
Change-Id: I4754189bcc20f824627d95858ba35285d53c614d
Reviewed-on: https://go-review.googlesource.com/c/go/+/245337
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Do them in the same CL so ARM's archreloc doesn't need to support
both streaming and non-streaming.
TODO: we haven't switched to using mmap to emit external
relocations on Windows.
Change-Id: Ica2ee89c03fc74839efd6b9e26c80585fcdce45c
Reviewed-on: https://go-review.googlesource.com/c/go/+/244357
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
OutData was used for a symbol to point to its data in the output
buffer, in order to apply relocations. Now we fold relocation
application to Asmb next to symbol data writing. We can just pass
the output data as a local variable.
Linking cmd/compile,
name old time/op new time/op delta
Asmb_GC 19.0ms ±10% 16.6ms ± 9% -12.50% (p=0.032 n=5+5)
name old alloc/op new alloc/op delta
Asmb_GC 3.78MB ± 0% 0.14MB ± 1% -96.41% (p=0.008 n=5+5)
name old live-B new live-B delta
Asmb_GC 27.5M ± 0% 23.9M ± 0% -13.24% (p=0.008 n=5+5)
Change-Id: Id870a10dce2a0a7447a05029c6d0ab39b47d0a12
Reviewed-on: https://go-review.googlesource.com/c/go/+/244017
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Support streaming external relocations on ARM64. Support
architecture-specific relocations.
Also support streaming external relocations on Darwin. Do it in
the same CL so ARM64's archreloc doesn't need to support both
streaming and non-streaming.
Change-Id: Ia7fee9957892f98c065022c69a51f47402f4d6e2
Reviewed-on: https://go-review.googlesource.com/c/go/+/243644
Reviewed-by: Jeremy Faller <jeremy@golang.org>
For content-addressable symbols, we build its content hash based
on the symbol data and relocations. When the compiler builds the
symbol data, it may not always include the trailing zeros, e.g.
the data of [10]int64{1,2,3} is only the first 24 bytes.
Therefore, we may end up with symbols with the same contents
(thus same hash) but different sizes. This is not actually a hash
collision. In this case, we can deduplicate them and keep the one
with the larger size.
Change-Id: If6834542d7914cc00f917d7db151955e5aee6f30
Reviewed-on: https://go-review.googlesource.com/c/go/+/243718
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
For content-addressable symbols with relocations, we build a
content hash based on its content and relocations. Depending on
the category of the referenced symbol, we choose different hash
algorithms such that the hash is globally consistent.
For now, we only support content-addressable symbols with
relocations when the current package's import path is known, so
that the symbol names are fully expanded. Otherwise, if the
referenced symbol is a named symbol whose name is not fully
expanded, the hash won't be globally consistent, and can cause
erroneous collisions. This is fine for now, as the deduplication
is just an optimization, not a requirement for correctness (until
we get to type descriptors).
Change-Id: I639e4e03dd749b5d71f0a55c2525926575b1ac30
Reviewed-on: https://go-review.googlesource.com/c/go/+/243142
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
For now, we only do this for symbols without relocations.
Mark static temps "local", as they are not referenced across DSO
boundaries. And deduplicating a local symbol and a non-local
symbol can be problematic.
Change-Id: I0a3dc4138aaeea7fd4f326998f32ab6305da8e4b
Reviewed-on: https://go-review.googlesource.com/c/go/+/243141
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Currently, when external linking, in relocsym (in asmb pass), we
convert Go relocations to an in-memory representation of external
relocations, and then in asmb2 pass we write them out to the
output file. This is not memory efficient.
This CL makes it not do the conversion but directly stream out
the external relocations based on Go relocations. Currently only
do this on AMD64 ELF systems.
This reduces memory usage, but makes the asmb2 pass a little
slower.
Linking cmd/compile with external linking:
name old time/op new time/op delta
Asmb_GC 83.8ms ± 7% 70.4ms ± 4% -16.03% (p=0.008 n=5+5)
Asmb2_GC 95.6ms ± 4% 118.2ms ± 5% +23.65% (p=0.008 n=5+5)
TotalTime_GC 1.59s ± 2% 1.62s ± 1% ~ (p=0.151 n=5+5)
name old alloc/op new alloc/op delta
Asmb_GC 26.0MB ± 0% 4.1MB ± 0% -84.15% (p=0.008 n=5+5)
Asmb2_GC 8.19MB ± 0% 8.18MB ± 0% ~ (p=0.222 n=5+5)
name old live-B new live-B delta
Asmb_GC 49.2M ± 0% 27.4M ± 0% -44.38% (p=0.008 n=5+5)
Asmb2_GC 51.5M ± 0% 29.7M ± 0% -42.33% (p=0.008 n=5+5)
TODO: figure out what is slow. Possible improvements:
- Remove redundant work in relocsym.
- Maybe there is a better representation for external relocations
now.
- Fine-grained parallelism in emitting external relocations.
- The old elfrelocsect only iterates over external relocations,
now we iterate over all relocations. Is it too many?
Change-Id: Ib0a8ee8c88d65864c62b89a8d634614f7f2c813e
Reviewed-on: https://go-review.googlesource.com/c/go/+/242603
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
For symbols of size 8 bytes or below, we can map them to 64-bit
hash values using the identity function. There is no need to use
longer and more expensive hash functions.
For them, we introduce another pseudo-package, PkgIdxHashed64. It
is like PkgIdxHashed except that the hash function is different.
Note that the hash value is not affected with trailing zeros,
e.g. "A" and "A\0\0\0" have the same hash value. This allows
deduplicating a few more symbols. When deduplicating them, we
need to keep the longer one.
Change-Id: Iad0c2e9e569b6a59ca6a121fb8c8f0c018c6da03
Reviewed-on: https://go-review.googlesource.com/c/go/+/242362
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Fill in the data at compile time, and get rid of the preprocess
function in the linker.
We need to be careful with symbol alignment: data symbols are
generally naturally aligned, except for string symbols which are
not aligned. When deduplicating two symbols with same content but
different alignments, we need to keep the biggest alignment.
Change-Id: I4bd96adfdc5f704b5bf3a0e723457c9bfe16a684
Reviewed-on: https://go-review.googlesource.com/c/go/+/242081
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
This CL introduces content-addressable symbols (a.k.a. hashed
symbols) to object files. Content-addressable symbols are
identified and referenced by their content hashes, instead of by
names.
In the object file, a new pseudo-package index PkgIdxHashed is
introduced, for content-addressable symbols, and a new block is
added to store their hashes. The hashes are used by the linker to
identify and deduplicate the symbols.
For now, we only support content-addressable symbols that are
always locally defined (i.e. no cross-package references).
As a proof of concept, make string constant symbols content-
addressable.
Change-Id: Iaf53efd74c0ffb54fa95f784628cc84e95844536
Reviewed-on: https://go-review.googlesource.com/c/go/+/242079
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The special case is no longer needed, didn't actually work, and
we no longer even save this map anywhere (see CL 240621 for more
information).
Change-Id: I19bcf32cace22decf50fd6414d4519cc51cbb0be
Reviewed-on: https://go-review.googlesource.com/c/go/+/241982
Reviewed-by: Austin Clements <austin@google.com>
Set the AttrStatic flag on compiler-emitted TOC symbols for ppc64; these
symbols don't need to go into the final symbol table in Go binaries.
This fixes a buglet introduced by CL 240539 that was causing failures
on the aix builder.
Change-Id: If8b63bcf6d2791f1ec5a0c371d2d11e806202fd2
Reviewed-on: https://go-review.googlesource.com/c/go/+/241637
Run-TryBot: Than McIntosh <thanm@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Don't emit symbol table entries for compiler-generated file-local
symbols (this category includes .stmp_* temporaries and *.stkobj
symbols). Note that user-written static symbols within assembler
sources will still be added to the symbol table. Apply the same test
when emitting DWARF for global variables.
Change-Id: I4db77a2750a0b575e051dfea895c4742cf6709a6
Reviewed-on: https://go-review.googlesource.com/c/go/+/240539
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Mark compiler-generated ".stmp_%d" and "<fn>.stkobj" symbols as
AttrStatic, so as to tell the linker that they do not need to be
inserted into its name lookup tables.
Change-Id: I59ffd11659b2c54c2d0ad41275d05c3f919e3b88
Reviewed-on: https://go-review.googlesource.com/c/go/+/240497
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Following CL 240399 and CL 240400, do the same for Mach-O.
Linking cmd/compile with external linking,
name old time/op new time/op delta
Asmb2_GC 32.7ms ± 2% 13.5ms ± 6% -58.56% (p=0.008 n=5+5)
name old alloc/op new alloc/op delta
Asmb2_GC 16.5MB ± 0% 6.4MB ± 0% -61.15% (p=0.008 n=5+5)
Change-Id: I0fd7019d8713d1940e5fbbce4ee8eebd926451a1
Reviewed-on: https://go-review.googlesource.com/c/go/+/241178
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
The loader writeBlock() function has code that tries to skip the
initial portion of the input symbols list depending on the address of
the section being written-- this code is dead (skipping is never
triggered) due to similar skipping in the callers; remove this
preamble.
Change-Id: I9769694a3194faf73ebebbbc10ceba4928c3087c
Reviewed-on: https://go-review.googlesource.com/c/go/+/241067
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Introduce a new loader method "SetCarrierSym", to be used when
establishing container/containee symbol relationships for symbol
bucketing in the symtab phase.
This new method is intended to be employed in situations where you
have a series of related symbols will be represented by a single
carrier symbol as a combined entity. The pattern here is that the
sub-symbols contain content but will be anonymous from a symbol table
perspective; the carrier symbol has no content itself but will appear
in the symbol table. Examples of carrier symbols that follow this
model are "runtime.itablink" and "runtime.typelink".
Change-Id: I1a3391a71062c7c740cb108b3fa210b7f69b81ed
Reviewed-on: https://go-review.googlesource.com/c/go/+/240509
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Introduce a new loader method "AddInteriorSym" to be used when
establishing container/containee symbol relationships for host object
sub-symbols and GOT/dynamic sub-symbols.
Interior symbols are employed in situations where you have a
"container" or "payload" symbol that has content, and then a series of
"interior" sub-symbols that point into a portion of the container
symbol's content. Each interior symbol will typically have a useful
name / size / value, but no content of its own. From a symbol table
perspective the container symbol is anonymous, but the interior
symbols are added to the output symbol table.
Change-Id: I919ed5dbbfe2ef2c9a76214f7ea9b384a1be6297
Reviewed-on: https://go-review.googlesource.com/c/go/+/240508
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
In CL 240399 we changed to precompute the size for ELF relocation
records and use mmap to write them, but we left architectures
where elfreloc1 write non-fixed number of bytes. This CL handles
those architectures. When a Go relocation will turn into multiple
ELF relocations, in relocsym we account this difference and add
it to the size calculation. So when emitting ELF relocations, we
know the number of ELF relocations to be emitted.
Change-Id: I6732ab674b442f4618405e5412a77f6e4a3315d7
Reviewed-on: https://go-review.googlesource.com/c/go/+/241079
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Now that we write ELF relocation records in mapped memory with
known sizes and offsets, we can write them in parallel.
Further speed up Asmb2 pass. Linking cmd/compile with external
linking,
Asmb2 141ms ± 4% 97ms ± 5% -30.98% (p=0.000 n=10+9)
Change-Id: I52c2b9230e90ed4421c21d7ef13a4f1e996f6054
Reviewed-on: https://go-review.googlesource.com/c/go/+/240400
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Currently, ELF relocations are generated sequentially in the heap
and flushed to output file periodically. In fact, in some cases,
the output size of the relocation records can be easily computed,
as a relocation entry has fixed size. We only need to count the
number of relocation records to compute the size.
Once the size is computed, we can mmap the output with the proper
size, and directly write relocation records in the mapped memory.
It also opens the possibility of writing relocations in parallel
(not done in this CL).
Note: on some architectures, a Go relocation may turn into
multiple ELF relocations, which makes size calculation harder.
This CL does not handle those cases, and it still writes
sequentially in the heap there.
Linking cmd/compile with external linking,
name old time/op new time/op delta
Asmb2 190ms ± 2% 141ms ± 4% -25.74% (p=0.000 n=10+10)
name old alloc/op new alloc/op delta
Asmb2_GC 66.8MB ± 0% 8.2MB ± 0% -87.79% (p=0.008 n=5+5)
name old live-B new live-B delta
Asmb2_GC 66.9M ± 0% 55.2M ± 0% -17.58% (p=0.008 n=5+5)
Change-Id: If7056bbe909dc90033eef6b9c4891fcca310602c
Reviewed-on: https://go-review.googlesource.com/c/go/+/240399
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Currently, on most platforms, the start/end symbols runtime.text
and runtime.etext are defined in symtab pass and assigned values
in address pass. In some cases (darwin+dynlink or AIX+external),
however, they are defined and assigned values in textaddress pass
(because they need non-zero sizes). Then their values get
overwritten in address pass. This is bad. The linker expects
their values to be consistent. In particular, in CL 239281,
findfunctab is split to two parts. The two parts need to have a
consistent view of the start/end symbols. If its value changes in
between, bad things can happen.
This CL fixes it by always defining runtime.text/etext symbols in
the textaddress pass.
Fix darwin and AIX builds.
Change-Id: Ifdc1bcb69d99be1b7e5b4fd31d473650c03e3b9d
Reviewed-on: https://go-review.googlesource.com/c/go/+/240065
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Create a new class of symbols internal to the linker. These symbols live
in the Loader, and are real smybols, but have no data, only size. After
symbols are allocated in the binary in asmb() a function is called that
is responsible for filling in the data.
This allows the linker to create large symbols, but not pay the price on
the heap memory.
Change-Id: Ib4291fc6e578478057ed2ec163d7b27426f1d5ff
Reviewed-on: https://go-review.googlesource.com/c/go/+/239280
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The Go compiler includes special treatment for a small set of very
commonly used type symbols (26 to be exact); for these types it
doesn't bother to emit type descriptors for "normal" compilations, and
instead only generates them for the runtime package, so as to reduce
object file bloat.
This patch moves the set of type symbols in question from the
PkgIdxNone index space (in the object file) to the PkgIdxBuiltin
space, which saves some work in the compiler and loader (reduces each
package's index space slightly).
Change-Id: I039c805e05c1aef26f035e52760fd0a0af40f7a5
Reviewed-on: https://go-review.googlesource.com/c/go/+/239658
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Change the object file writer to avoid adding entries to the object
file string table for builtin functions. This helps save some very
small amount of space in the object file.
Change-Id: Ic3b94a154e00eb4c7378b57613580c7073b841bc
Reviewed-on: https://go-review.googlesource.com/c/go/+/239657
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
On AIX, in relocsym we call Xcoffadddynrel, which adds a
relocation record to a global array. relocsym already runs in
parallel. In the past we only parallelize over segments, and
we call Xcoffadddynrel only for symbols in data segment, so it is
effectively called sequentially. In CL 239197 we started to do
more fine-grained parallelism, so we need to make sure it is safe
to call Xcoffadddynrel in parallel.
Fix AIX build.
Change-Id: I3128193995a5a99d9fa04c8e728e590f17298da3
Reviewed-on: https://go-review.googlesource.com/c/go/+/239561
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The code in the compiler's DWARF line table generation emits line
table ops at the end of each function fragment to reset the state
machine registers back to an initial state, so that when the line
table fragments for each function are stitched together into a
compilation unit, each fragment will have a clean starting point. The
set-file/set-line ops emitted in this code were being applied to the
last row of the line table, however, meaning that they were
overwriting the existing values.
To avoid this problem, add code to advance the PC past the end of the
last instruction in the function, and switch to just using an
end-of-sequence operator at the end of each function instead of
explicit set-file/set-line ops.
Updates #39757.
Change-Id: Ieb30f83444fa86fb1f2cd53862d8cc8972bb8763
Reviewed-on: https://go-review.googlesource.com/c/go/+/239286
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: David Chase <drchase@google.com>
We can apply relocations of a symbol right after the symbol data
is copied to output buffer. This should help locality and
parallelism (parallelizing over blocks, instead of over segments).
Linking cmd/compile,
Asmb+Reloc 23.9ms ±18% 16.5ms ±11% -30.73% (p=0.008 n=5+5)
Linking cmd/compile with external linking,
Asmb+Reloc 74.0ms ± 3% 33.8ms ± 8% -54.32% (p=0.008 n=5+5)
In external linking mode, allocation goes up slightly, as we do
smaller batching now. It doesn't seem too bad.
Asmb+Reloc 15.0MB ± 0% 16.7MB ± 0% +11.22% (p=0.008 n=5+5)
Change-Id: Ide33d9ff86c39124c8f5cfc050d7badc753a1ced
Reviewed-on: https://go-review.googlesource.com/c/go/+/239197
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
This patch introduces parallelization of DWARF generation on a per
compilation unit basis. Each compilation unit now operates on a
separate set of symbols, so it's safe to send each compilation unit to
a goroutine to be processed in parallel.
Doing this requires some restructing to ensure that any new symbols
needed are created up front, since we can't create any new syms during
the parallel portion. Similarly, the parallel portion can't set any
symbol attributes, so the check that verifies we haven't doubly listed
any DIE syms had to be reworked, and setting of reachability has to be
delayed until after the parallel phase is complete.
Change-Id: I3042b76e9b597bb1a6a44dce19efba2d02bed76b
Reviewed-on: https://go-review.googlesource.com/c/go/+/237679
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
The loader's SymbolBuilder Add*/Set* methods include a call to mark
the underlying symbol as reachable (as a convenience, so that callers
would not have to set it explicitly). This code was carried over from
the corresponding sym.Symbol methods; back in the sym.Symbol world
unreachable symbols were never removed from the AllSyms slice, hence
setting and checking reachability was a good deal more important.
With the advent of the loader and the new deadcode implementation,
there is less of a need for this sort of fallback, and in addition the
implicit attr setting introduces data races in the the loader if there
are SymbolBuilder Add*/Set* method calls in parallel threads, as well
as adding overhead to the methods.
This patch gets rid of the implicit reachability setting, and instead
marks reachability in CreateSymForUpdate, as well as adding a few
explicit SetAttrReachable calls where needed.
Change-Id: I029a0c5a4a24237826a7831f9cbe5180d44cbc40
Reviewed-on: https://go-review.googlesource.com/c/go/+/237678
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Refactor some of the linker's DWARF generation methods so as to have
helper routines that do all the work for a given comp unit for a given
section (range, loc, etc). No change in functionality, this is just a
reorg in preparation for a later patch in this sequence.
Change-Id: I86fc789220326a4e522904a5924c8971d6757189
Reviewed-on: https://go-review.googlesource.com/c/go/+/237677
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Rework the way symbols are handled in DWARF line table generation to
eliminate copying the data payload for all SWDWARFLINES syms (emitted
by the compiler) into the payload of the ".debug_line" section symbol
(generated by the linker). Instead, chain together the SWDWARFLINES
symbols into a list, then append that list to the section sym list in
dwarfp (this moves us from a single monolithic .debug_line to a
.debug_line section sym followed by a list symbols (one per function
and an epilog symbol per compilation unit). To enable this work, move
the emission of the DW_LNE_set_address op (at the start of each
function) from the linker to the compiler.
Change-Id: Iec61b44a451f7a386c82a89bf944de482b018789
Reviewed-on: https://go-review.googlesource.com/c/go/+/237427
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Don't emit the .debug_pubnames/.debug_pubtypes sections. These sections
are not used by either GDB or Delve, and C++ compilers [notably GCC
and Clang] no longer emit the sections by default.
Change-Id: Ic3309755e88c8e1aa28a29366bc7f0df1748fe64
Reviewed-on: https://go-review.googlesource.com/c/go/+/237426
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The linker's asmb phase for ELF has a chunk of code that decides where
to place the .symtab and .strtab sections, which appear after after
the DWARF data; this code currently tries to align the start of the
.symtab section using the value of -R (stored in *FlagRound). This
patch gets rid of this additional alignment and instead just aligns
.symtab by pointer size. The -R value is needed for loadable
segments/sections (such as text or data), not for non-loadable
sections (e.g. symtab). On most architectures the *FlagRound value is
4k, however on ARM64 it is 64k, meaning that aligning symtab on this
boundary can waste a good chunk of space.
Change-Id: Ib51f3ad5611f5614768355eb8533084ba117a8e9
Reviewed-on: https://go-review.googlesource.com/c/go/+/238019
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Currently, a method of a reachable type is live if it matches a
method of a reachable interface. In fact, we only need to retain
the method if the type is actually converted to an interface. If
the type is never converted to an interface, there is no way to
call the method through an interface method call (but the type
descriptor could still be used, e.g. in calling
runtime.newobject).
A type can be used in an interface in two ways:
- directly converted to interface. (Any interface counts, as it
is possible to convert one interface to another.)
- obtained by reflection from a related type (e.g. obtaining an
interface of T from []T).
For the former, we let the compiler emit a marker on the type
descriptor symbol when it is converted to an interface. In the
linker, we only need to check methods of marked types.
For the latter, when the linker visits a marked type, it needs to
visit all its "child" types as marked (i.e. potentially could be
converted to interface).
This reduces binary size:
cmd/compile 18792016 18706096 (-0.5%)
cmd/go 14120572 13398948 (-5.1%)
Change-Id: I4465c7eeabf575f4dc84017214c610fa05ae31fd
Reviewed-on: https://go-review.googlesource.com/c/go/+/237298
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Merge conflicts are mostly recently changed nm/objdump output
format and its tests. Resolved easily (mostly just using the
format on master branch).
Change-Id: I99d8410a9a02947ecf027d9cae5762861562baf5
Instead of using container/heap package, implement a simple
specialized heap algorithm for the work queue in the deadcode
pass, to avoid allocations and function pointer calls.
Linking cmd/compile,
name old time/op new time/op delta
Deadcode_GC 59.8ms ± 4% 42.2ms ± 4% -29.45% (p=0.008 n=5+5)
name old alloc/op new alloc/op delta
Deadcode_GC 3.53MB ± 0% 2.10MB ± 0% -40.57% (p=0.008 n=5+5)
name old allocs/op new allocs/op delta
Deadcode_GC 187k ± 0% 8k ± 0% -95.48% (p=0.008 n=5+5)
Change-Id: Ibb21801d5b8e4a7eaf429856702e02720cd1772f
Reviewed-on: https://go-review.googlesource.com/c/go/+/236565
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Currently, in the deadcode pass, when checking whether a defined
method satisfies an interface, it compares the string
representation of the defined method and the interface method.
In fact, it can simply compare the method name and the type
descriptor (as we do in runtime). Make it so.
Change-Id: Ideb2b2410e5eedcd20ac31e3af41f5499fc90225
Reviewed-on: https://go-review.googlesource.com/c/go/+/236564
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Now the compiler-generated fingerprint is a hash of the export
data. We don't need to hash it ourselves in the linker. And the
linker doesn't need to read export data at all.
Fixes#33820.
Change-Id: I54bf3ebfd0f0c72aa43a352d7b2e0575dd62970d
Reviewed-on: https://go-review.googlesource.com/c/go/+/236119
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Currently, the compiler generates a fingerprint for each package,
which is used by the linker for index consistency check.
When building plugin or shared object, currently the linker also
generates a hash, by hashing the export data. At run time, when
a package is referenced by multiple DSOs, this hash is compared
to ensure consistency.
It would be good if we can unify this two hashes. This way, the
linker doesn't need to read the export data (which is intended
for the compiler only, and is not always available for the
linker). The export data hash is sufficient for both purposes.
It is consistent with the current hash geneated by the linker.
And the export data includes indices for exported symbols, so its
hash can be used to catch index mismatches.
Updates #33820.
Change-Id: I2bc0d74930746f54c683a10dfd695d50ea3f5a38
Reviewed-on: https://go-review.googlesource.com/c/go/+/236118
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Background: when compiling a function, it's possible that a local
variable will be optimized away, which could potentially degrade the
debugging experience if the compiler fails to emit DWARF information
for the variable's type. To mitigate this situation, the compiler
emits R_USETYPE relocations for the function's auto/param variables as
a signal to the linker to generate DWARF for the types in question,
even if the type is not specifically attached to a DWARF param or var.
This patch change the logic in the compiler to avoid emitting a
R_USETYPE relocation if the type in question is already referenced by
a concrete DWARF param or auto record. This cuts down on the amount of
work the linker has to do, also makes object files a bit smaller on
average (about 1% for the runtime package).
Change-Id: I4d24da458d0658edf90c5dca0bf21d5ddc3961d8
Reviewed-on: https://go-review.googlesource.com/c/go/+/234837
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Rework the code in the linker that visits DWARF subprorgam DIEs to
reduce number of symbol name instantiations and name lookups, by
making better use of relocation target symbol types.
Change-Id: Ifb2a4e24874b8c891d7fdf17dd749c3f9139157a
Reviewed-on: https://go-review.googlesource.com/c/go/+/234685
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This change splits the SDWARFINFO symbol type (a generic container of
DWARF content) into separate sub-classes. The new symbol types are
SDWARFCUINFO comp unit DIE, also CU info and CU packagename syms
SDWARFCONST constant DIE
SDWARFFCN subprogram DIE (default and concrete)
SDWARFABSFCN abstract function DIE
SDWARFTYPE type DIE
SDWARFVAR global variable DIE
Advantage of doing this: in the linker there are several places where
we have to iterate over a symbol's relocations to pick out references
to specific classes of DWARF sub-symbols (for example, looking for all
abstract function DIEs referenced by a subprogram DIE, or looking at
all the type DIEs used in a subprogram DIE). By splitting SDWARFINFO
into parts clients can now look only at the relocation target's sym
type as opposed to having to materialize the target sym name, or do a
lookup.
Change-Id: I4e0ee3216d3c8f1a78bec3d296c01e95b3d025b5
Reviewed-on: https://go-review.googlesource.com/c/go/+/234684
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Most Go objects are compiled with known package path, so the
symbol name is already fully expanded. Nevertheless, currently
in the linker strings.Replace is called unconditionally, and most
of the time it doesn't do anything.
This CL records a per-object flag in the object file, and do the
name expansion only when the name is not expanded at compile time.
This gives small speedups for the linker. Linking cmd/compile:
name old time/op new time/op delta
Loadlib 35.1ms ± 2% 32.8ms ± 4% -6.43% (p=0.008 n=5+5)
Symtab 15.8ms ± 2% 14.0ms ± 8% -11.45% (p=0.008 n=5+5)
TotalTime 399ms ± 1% 385ms ± 2% -3.63% (p=0.008 n=5+5)
Change-Id: I735084971a051cd9be4284ad294c284cd5b545f4
Reviewed-on: https://go-review.googlesource.com/c/go/+/234490
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
This deletes all sym.Symbol and sym.Reloc references. This is
certainly not complete, and there are more cleanups to do. But I
feel this makes a good first round.
Change-Id: I7621d016957f7ef114be5f0606fcb3ad6aee71c8
Reviewed-on: https://go-review.googlesource.com/c/go/+/234097
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Currently, for the special field tracking symbol go.track.XXX,
when they are reachable, we set its type to SCONST. There is no
need to do that. Just leave it unset (as Sxxx). The symbol is
done after this point.
Change-Id: I966d80775008f7fb5d30fbc6b9e4a30ae8316b6c
Reviewed-on: https://go-review.googlesource.com/c/go/+/233998
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Use uint32 consistently for local index (this is what the object
file uses).
Use a index, instead of a pointer, to refer to the object file.
This reduces memory usage and GC work.
This reduces some allocations. Linking cmd/compile,
name old alloc/op new alloc/op delta
Loadlib_GC 19.9MB ± 0% 16.9MB ± 0% -15.33% (p=0.008 n=5+5)
name old live-B new live-B delta
Loadlib_GC 12.6M ± 0% 11.3M ± 0% -9.97% (p=0.008 n=5+5)
Change-Id: I20ce60bbb6d31abd2e9e932bdf959e2ae840ab98
Reviewed-on: https://go-review.googlesource.com/c/go/+/233779
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Remove the loader's PropagateSymbolChangesBackToLoader and
PropagateLoaderChangesToSymbols shim functions. These were used at one
point to enable conversion of phases in the linker that were
"downstream" of loadlibfull -- given the current wavefront position
there's not much point keeping them around.
Change-Id: I3f01f25b70b1b80240369c8f3a10dca89931610f
Reviewed-on: https://go-review.googlesource.com/c/go/+/233817
Run-TryBot: Than McIntosh <thanm@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Minor renaming cleanup to get rid of a couple of old sym.Symbol
adddynrel helpers and rename the current crop of adddynrel2
methods/functions back to adddynrel.
Change-Id: I67e76decff84d603ef765f3b6a0cd78c7f3743ec
Reviewed-on: https://go-review.googlesource.com/c/go/+/233523
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Adds in support for remaining architectures to the linker's ELF asmb2
path, along with deleting most of the older sym.Symbol based code.
Change-Id: I67c96525db72b7d6dd3187cf2b9f6faddc296291
Reviewed-on: https://go-review.googlesource.com/c/go/+/233362
Reviewed-by: Cherry Zhang <cherryyz@google.com>
This patch converts the linker's Asmb2 phase to use loader APIs
for AMD64 (other architectures to be converted in a subsequent
patch).
Change-Id: I5a9aa9b03769cabc1a22b982f48fd113213d7bcf
Reviewed-on: https://go-review.googlesource.com/c/go/+/233338
Run-TryBot: Than McIntosh <thanm@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
The undef pass basically double-checks the relocation targets are
defined. We already do that in the reloc pass, and for external
relocations we check that when we emit relocations. The undef pass
doesn't seem necessary.
Change-Id: Iecfa654dc014fdc6e59c624cbf5948ad65fd367a
Reviewed-on: https://go-review.googlesource.com/c/go/+/232577
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
The function Abbrevs() was returning an array of structures by value,
which is not very efficient (this was showing up in a kubernetes
kubelet linker profile). Switch the function to return a slice
instead.
Improves linker DwarfGenerateDebugSyms running time when
linking the compiler in compilebench:
DwarfGenerateDebugSyms 29.2ms ±144% 23.9ms ±125% -17.89% (p=0.000 n=99+99)
Change-Id: I1132816563f208c63eb82a7932d9f2bcb2455324
Reviewed-on: https://go-review.googlesource.com/c/go/+/231558
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Add code to the loader to enforce the invariant that there is only a
single level of 'outer' symbol nesting. That is, if outer(X) = Y, then
outer(Y) is always zero.
Revise foldSubSymbolOffset based on the new invariant, allowing it to
be inlined, and then fix the various "for s.Outer != nil" loops in the
linker to just use an "if" instead of a loop.
Change-Id: Ib895702bc6de52718248f09a5368b84cb2e0a3fa
Reviewed-on: https://go-review.googlesource.com/c/go/+/231137
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
2020-05-01 16:49:36 +00:00
2011 changed files with 106441 additions and 107736 deletions
@@ -660,16 +666,13 @@ For example, you should not set <code>$GOHOSTARCH</code> to
<code>arm</code> on an x86 system.
</p>
<li><code>$GO386</code> (for <code>386</code> only, default is auto-detected
if built on either <code>386</code> or <code>amd64</code>, <code>387</code> otherwise)
<li><code>$GO386</code> (for <code>386</code> only, defaults to <code>sse2</code>)
<p>
This controls the code generated by gc to use either the 387 floating-point unit
(set to <code>387</code>) or SSE2 instructions (set to <code>sse2</code>) for
floating point computations.
This variable controls how gc implements floatingpoint computations.
</p>
<ul>
<li><code>GO386=387</code>: use x87 for floating point operations; should support all x86 chips (Pentium MMX or later).</li>
<li><code>GO386=sse2</code>: use SSE2 for floating point operations; has better performance than 387, but only available on Pentium 4/Opteron/Athlon 64 or later.</li>
<li><code>GO386=softfloat</code>: use software floating point operations; should support all x86 chips (Pentium MMX or later).</li>
<li><code>GO386=sse2</code>: use SSE2 for floating point operations; has better performance but only available on Pentium 4/Opteron/Athlon 64 or later.</li>
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
