Introduces the CopyFileRange syscall which first appears in Linux 4.5.
Allows copying file content between file descriptors within the kernel
without transferring data to user space. This syscall also allows the
kernel to take advantage of reflinking or other fast copy methods on
supported file systems.
http://man7.org/linux/man-pages/man2/copy_file_range.2.html
Change-Id: Id365f1e5d4d5ddf7159478e3a13084c9576ebd5c
Reviewed-on: https://go-review.googlesource.com/39992
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Right now the process for adding in new constants, errors, or syscalls
for Linux is a pain and unreliable. The scripts are designed to be run
on the target architecture and use the header files installed on the
user's system. This makes it hard to generate files for all the
architectures or to have consistency between users. See golang/go#15282.
This CL fixes this issue by making all of the files for the 11 supported
architectures directly from source checkouts of Linux, glibc, and bluez.
This is done using Docker, the gcc cross-compilers, and qemu emulation.
Previously discussed here:
https://go-review.googlesource.com/c/37589/
A README.md file is also added to explain how all the parts of the build
system work.
In order to get the build working for all the architectures, I made
some changes to the other scripts called from mkall_linux.go:
- Files only used for generating linux code, moved to linux/
- linux/mksysnum.pl supports a specified CC compiler.
- The generated C code in mkerrors.sh changed to avoid a warning
- mkerrors.sh headers changed to fix powerpc64 bug in sys/ioctl.h
- linux/types.go no longer needs to export Ptrace structs in lowercase
Build instructions:
- Host system needs to be x86-64 Linux
- Install Docker (https://docs.docker.com/engine/installation/)
- ./mkall.sh (That's it!!!)
Change-Id: I87067c14442ba12f8d51991349a43a9d73f38ae0
Reviewed-on: https://go-review.googlesource.com/37943
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
This CL introduces the Getrandom syscall. Allowing users of
sufficiently new kernels (3.17) to detect when the operating system
lacks enough entropy to generate cryptographic random data. See:
http://man7.org/linux/man-pages/man2/getrandom.2.html
As we have updated the syscall numbers, the Getrandom function can
just be generated normally. The only other change is to mkerrors.sh
to include the GRND_* flags for getrandom from <linux/random.h>.
It is important to note that currently the mkerrors.sh script cannot
actually be run as it has fallen behind the 4.10 kernel. The GRND_*
flags were added manually.
This was originally part of a larger commit that was split up.
See CL: https://go-review.googlesource.com/37570
Change-Id: I464088a9c9f7224abdcdae476b30a4f2f412d9f5
Reviewed-on: https://go-review.googlesource.com/37589
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Matt Layher <mdlayher@gmail.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Pause is a legacy syscall not available on linux-arm64. Use ppoll with
all args as 0 to emulate - this is the way musl libc does Pause when the
pause syscall isn't available.
With the changes in syscall_linux* and regenerating zsyscall_linux*,
this calling Pause on linux-arm64 works and returns EINTR as expected.
Change-Id: I88236290313f18c742d826e759e86ff260a8b383
Reviewed-on: https://go-review.googlesource.com/22014
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Since Linux 3.11, O_TMPFILE flag can be used in open syscall to create
an unnamed file in a directory. The file occupies space in the
filesystem, and can be given a name using linkat syscall. If the file is
closed without being given a name, its contents are deleted.
See the manpage open(2) in Linux for details.
Exports O_TMPFILE for Linux in 386 and amd64 (other architectures
already had it). Exports Linkat syscall and AT_SYMLINK_FOLLOW (used for
giving a name to the file) for all Linux in all architectures.
Fixesgolang/go#7830.
Change-Id: Ib82e44f405b227e227b9cbf317c2657b32e046f5
Reviewed-on: https://go-review.googlesource.com/21003
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
arm64 doesn't have a Dup2 syscall, instead Dup3 is supposed
to be used. Since Dup3 is linux-specific, provide a wrapper
to make writing portable code easier.
Updates golang/go#10235
To verify it, added a testcase for Dup and Dup2.
Change-Id: I066bb60d62b2bd64d7ba0fdfbb334ce2213c78e9
Reviewed-on: https://go-review.googlesource.com/20178
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Riku Voipio <riku.voipio@linaro.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This system call is used to reassociate the current thread with a Linux
namespace (e.g. a network namespace or a mount namespace). This system
call is key to interacting with the primitives enabling Linux containers.
The users of this system call will most likely want to wrap their calls
with a pair of LockOSThread / UnlockOSThread calls. Here is an example
that is a reasonably close approximation of the `ns_exec' program given
as an example in `man 2 setns':
package main
import (
"log"
"os"
"os/exec"
"runtime"
"golang.org/x/sys/unix"
)
func main() {
if len(os.Args) < 3 {
log.Fatalf("%s /proc/PID/ns/FILE cmd args...", os.Args[0])
}
fd, err := unix.Open(os.Args[1], unix.O_RDONLY, 0)
if err != nil {
log.Fatalf("open: %s", err)
}
runtime.LockOSThread()
defer runtime.UnlockOSThread()
if err = unix.Setns(fd, 0); err != nil {
log.Fatalf("setns: %s", err)
}
cmd := exec.Command(os.Args[2], os.Args[3:]...)
cmd.Stdin = os.Stdin
cmd.Stdout = os.Stdout
cmd.Stderr = os.Stderr
err = cmd.Run()
if err != nil {
log.Fatalf("exec: %s", err)
}
}
Fixesgolang/go#5968.
Change-Id: I78dc54667cfaef4f9e99a08d48f6e423686f1b22
Reviewed-on: https://go-review.googlesource.com/20054
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
