In issue 50113, we see that a thread blocked in a system call can result
in a hang of AllThreadsSyscall. To resolve this, we must send a signal
to these threads to knock them out of the system call long enough to run
the per-thread syscall.
Stepping back, if we need to send signals anyway, it should be possible
to implement this entire mechanism on top of signals. This CL does so,
vastly simplifying the mechanism, both as a direct result of
newly-unnecessary code as well as some ancillary simplifications to make
things simpler to follow.
Major changes:
* The rest of the mechanism is moved to os_linux.go, with fields in mOS
instead of m itself.
* 'Fixup' fields and functions are renamed to 'perThreadSyscall' so they
are more precise about their purpose.
* Rather than getting passed a closure, doAllThreadsSyscall takes the
syscall number and arguments. This avoids a lot of hairy behavior:
* The closure may potentially only be live in fields in the M,
hidden from the GC. Not necessary with no closure.
* The need to loan out the race context. A direct RawSyscall6 call
does not require any race context.
* The closure previously conditionally panicked in strange
locations, like a signal handler. Now we simply throw.
* All manual fixup synchronization with mPark, sysmon, templateThread,
sigqueue, etc is gone. The core approach is much simpler:
doAllThreadsSyscall sends a signal to every thread in allm, which
executes the system call from the signal handler. We use (SIGRTMIN +
1), aka SIGSETXID, the same signal used by glibc for this purpose. As
such, we are careful to only handle this signal on non-cgo binaries.
Synchronization with thread creation is a key part of this CL. The
comment near the top of doAllThreadsSyscall describes the required
synchronization semantics and how they are achieved.
Note that current use of allocmLock protects the state mutations of allm
that are also protected by sched.lock. allocmLock is used instead of
sched.lock simply to avoid holding sched.lock for so long.
Fixes#50113
Change-Id: Ic7ea856dc66cf711731540a54996e08fc986ce84
Reviewed-on: https://go-review.googlesource.com/c/go/+/383434
Reviewed-by: Austin Clements <austin@google.com>
Trust: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Generally speaking Go functions make no guarantees
about what has happened to result parameters on error,
and Pipe is no exception: callers should avoid looking at
p if Pipe returns an error.
However, we had a bug in which ForkExec was using the
content of p after a failed Pipe, and others may too.
As a robustness fix, make Pipe avoid writing to p on failure.
Updates #50057
Change-Id: Ie8955025dbd20702fabadc9bbe1d1a5ac0f36305
Reviewed-on: https://team-review.git.corp.google.com/c/golang/go-private/+/1291271
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-on: https://go-review.googlesource.com/c/go/+/370577
Run-TryBot: Filippo Valsorda <filippo@golang.org>
Trust: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Alex Rakoczy <alex@golang.org>
These were identical. This is a preliminary step
towards remove allocs per UDP receive.
Change-Id: I83106cd3f1fe4bc5bae2d1b0ebd23eedd820abed
Reviewed-on: https://go-review.googlesource.com/c/go/+/361258
Trust: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
These were identical. This is a preliminary step
towards remove allocs per UDP send.
Change-Id: I21e1264c7d4747baa626ddb93afff4c1cf225d13
Reviewed-on: https://go-review.googlesource.com/c/go/+/361256
Trust: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Most architectures currently already implement Pipe using the pipe2
syscall. Only 386, amd64 and mips{,le} still use the pipe syscall.
However, some systems (e.g. Android seccomp policies) block that
syscall, see #40828 for an example.
The pipe2 syscall was added in Linux kernel version 2.6.27. The minimum
required Linux kernel version for Go 1.18 will be changed to 2.6.32
per #45964 so it is possible to unify the implementation of Pipe using
the pipe2 syscall.
For #45964
Change-Id: I8ed6a391300c95f3107b4ec6b27d320e42fb535b
Reviewed-on: https://go-review.googlesource.com/c/go/+/350530
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 minimum required Linux kernel version for Go 1.18 will be changed to
2.6.32, see #45964. The current minimum required version is 2.6.23 and
accept4 was added in 2.6.28, so the fallback to accept in Accept on
Linux can be removed.
For #45964
Change-Id: I78fc4e5b58417bbc540912c9dbf1b1b3db888fea
Reviewed-on: https://go-review.googlesource.com/c/go/+/346849
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 minimum required Linux kernel version for Go 1.18 will be changed to
2.6.32, see #45964. The current minimum required version is 2.6.23 and
utimensat was added in 2.6.22, so the fallback isn't even necessary for
the current minimum supported version. Remove the fallback to utimes.
For #45964
Change-Id: I5536f6ea7a34944dd9165f1533c10692171fb0c5
Reviewed-on: https://go-review.googlesource.com/c/go/+/346790
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>
Android seems to block the accept syscall in newer versions. Use accept4
instead on kernel versions that support it (Linux 2.6.28 and newer) and
fall back to accept on ENOSYS.
Fixes#45767
Change-Id: If190ace0e0213207fdaf6eeb79a5543ef18456de
Reviewed-on: https://go-review.googlesource.com/c/go/+/313769
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
This replaces five implementations scattered across low level packages.
(And I plan to use it in a sixth soon.)
Three of the five were byte-for-byte identical.
Change-Id: I3bbbeeac63723a487986c912b604e10ad1e042f4
Reviewed-on: https://go-review.googlesource.com/c/go/+/301549
Trust: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
This change fixes two failng tests on linux-ppc64x:
- TestAllThreadsSyscall() exposed a real bug in the ppc64x support:
- It turns out that the r2 syscall return value is not defined
on all architectures. Notably linux-ppc64x so address that by
introducing a private architectural constant in the syscall
package, archHonorsR2: true if r2 has a determanistic value.
- TestSetuidEtc() was sensitive to /proc/<PID>/status content:
- The amount of padding space has changed with kernel vintage.
- Stress testing revealed a race with /proc files disappearing.
Fixes#42178
Change-Id: Ie6fc0b8f2f94a409ac0e5756e73bfce113274709
Reviewed-on: https://go-review.googlesource.com/c/go/+/266202
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
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>
Follow glibc's implementation and check secondary group memberships
using Getgroups.
No test since we cannot easily change file permissions when not running
as root and the test is meaningless if running as root.
Same as CL 238722 did for x/sys/unix
Updates #39660
Change-Id: I6af50e27b255e33405558947a0ab3dfbc33b2d50
Reviewed-on: https://go-review.googlesource.com/c/go/+/238937
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reimplement syscall wrappers for linux/arm64 in terms of supported
syscalls (or in case of Ustat make it return ENOSYS) and remove the
manually added SYS_* consts for the deprecated syscalls. Adapted from
golang.org/x/sys/unix where this is already done since CL 119655.
Change-Id: I94ab48a4645924df3822497d0575f1a1573d509f
Reviewed-on: https://go-review.googlesource.com/c/go/+/199140
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
On linux/riscv64, the renameat syscall no longer exists and has been
superseded by renameat2. Thus we'll have to use Renameat2 to implement
Renameat on linux/riscv64 for #27532. Prepare for this by moving the
Renameat definition to the GOARCH specific files.
Follow CL 157899 which did the same for golang.org/x/sys/unix
Change-Id: I9670213cc3987df48fee962ddee36915a7785560
Reviewed-on: https://go-review.googlesource.com/c/go/+/192077
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The Linux kernel faccessat system call does not take a flags parameter.
The flag parameter to the C library faccessat function is implemented in C.
The syscall.Faccessat function takes a flags parameter. In older releases
we have passed the flags parameter to the kernel, which ignored it.
In CL 120015 we started returning an error if any flags were set.
That seems clearly better than ignoring them, but it turns out that some
code was using the flags. The code was previously subtly broken.
Now it is obviously broken. That is better, but we can do better still:
we can implement the flags as the C library does. That is what this CL does.
Change-Id: I259bd6f240c3951e939b81c3032dead3d9c567b4
Reviewed-on: https://go-review.googlesource.com/126415
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
As mentioned in #25845, port CL 46474 from golang.org/x/sys/unix to the
syscall package.
Currently Linux' fchmodat(2) syscall implementation doesn't support the
flags parameter (though it might in future versions [1]). Fchmodat in
the syscall package takes the parameter and (wrongly) passes it on to the
syscall which will ignore it.
According to the POSIX.1-2008 manual page [2], AT_SYMLINK_NOFOLLOW is
the only valid value for the flags parameter and EOPNOTSUPP should be
returned in case changing the mode of a symbolic link is not supported
by the underlying system. EINVAL should be returned for any other value
of the flags parameter.
[1] https://patchwork.kernel.org/patch/9596301/
[2] http://pubs.opengroup.org/onlinepubs/9699919799/functions/chmod.html
Updates #20130
Updates #25845
Change-Id: I1021dd0e6a4f4cb3557cb1c1b34dd618c378cda6
Reviewed-on: https://go-review.googlesource.com/118658
Reviewed-by: Ian Lance Taylor <iant@golang.org>
By simply rearranging the logic, we avoid the overhead of a superfluous
call to getsockopt. For, if p is already non empty, there's no point
in having to check if we need to attach dummy payload. This has
performance benefits when using send/recvmsg for high speed
communications.
Change-Id: Id85cff17328ecbf6d09dd52fbeeaa691dbe69b75
Reviewed-on: https://go-review.googlesource.com/108338
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The syscall package is frozen and we don't want to encourage anyone to
implement these syscalls.
Change-Id: I6b6e33e32a4b097da6012226aa15300735e50e9f
Reviewed-on: https://go-review.googlesource.com/96315
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Add the rawSyscallNoError wrapper function which is used for Linux
syscalls that don't return an error and convert all applicable
occurences of RawSyscall to use it instead.
Fixes#22924
Change-Id: Iff1eddb54573d459faa01471f10398b3d38528dd
Reviewed-on: https://go-review.googlesource.com/84485
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
The getdents64 syscall is only available for mips64/mips64le starting
with Linux kernel 3.10. Since mips64le requires at least 4.8 according
to [1] (regarding #16848) using it should be fine.
[1] https://golang.org/wiki/MinimumRequirements
This CL changes the binary layout of type Dirent for mips64/mips64le,
but not the public API. But since the currently used layout doesn't
match the struct linux_dirent returned by the getdents syscall this
should be fine as well.
Fixes#23624
Change-Id: Iaa7306fa6e4442ad2fed41c60b37627a7314f117
Reviewed-on: https://go-review.googlesource.com/91055
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Directly return error instead of assigning to err and then returning.
Change-Id: Ie5c466cac70cc6d52ee72ebba3e497e0da8a5797
Reviewed-on: https://go-review.googlesource.com/73531
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
syscall.Exit and runtime.exit do the same thing.
Why duplicate code?
CL 45115 fixed bug where windows runtime.exit was correct,
but syscall.Exit was broken. So CL 45115 fixed windows
syscall.Exit by calling runtime.exit.
Austin suggested that all OSes should do the same, and
this CL implements his idea.
While making changes, I discovered that nacl syscall.Exit
returned error
func Exit(code int) (err error)
and I changed it into
func Exit(code int)
like all other OSes. I assumed it was a mistake and it
is OK to do because cmd/api does not complain about it.
Also I changed plan9 runtime.exit to accept int32 just
like all other OSes do.
Change-Id: I12f6022ad81406566cf9befcc6edc382eebd413b
Reviewed-on: https://go-review.googlesource.com/66170
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: David du Colombier <0intro@gmail.com>
The previous implementation forced all Unix socket to have a name
strictly shorter than len(sa.raw.Path) to allow a terminating NULL
byte to be added. This requirement does not apply to abstract socket
names under Linux, so for this case we allow the full length.
Fixes#21965
Change-Id: I1d1f58b6b6172d589428c7230cfeae984de78b4b
Reviewed-on: https://go-review.googlesource.com/66190
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
This commit relaxes SendmsgN behavior of introducing a dummy 1-byte
payload when sending ancillary-only messages.
The fake payload is not needed for SOCK_DGRAM type sockets, and actually
breaks interoperability with other fd-passing software (journald is one
known example). This introduces an additional check to avoid injecting
dummy payload in such case.
Full reference at https:/golang.org/issue/6476#issue-51285243
Fixes#6476
Change-Id: I19a974b4e7920e002bd0556259ab766572358520
Reviewed-on: https://go-review.googlesource.com/45872
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Fixes#21437
Change-Id: I55fbf5114ae1bb7f4aa1a20450e8d5309756cd5b
Reviewed-on: https://go-review.googlesource.com/55430
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Delete use stub from asm.s, leaving only a dummy file.
Deleting the file causes Windows build to fail.
Fixes#16607
Change-Id: Ic5a55e042e588f1e1bc6605a3d309d1eabdeb288
Reviewed-on: https://go-review.googlesource.com/36716
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Don't panic, crash, or return references to uninitialized memory when
ParseDirent is passed invalid input.
Move common dirent parsing to syscall.go with minimal platform-specific
functions in syscall_$GOOS.go.
Fixes#15653
Change-Id: I5602475e02321fe381064488401c14b33bec6886
Reviewed-on: https://go-review.googlesource.com/23780
Run-TryBot: Damien Neil <dneil@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The tree's pretty inconsistent about single space vs double space
after a period in documentation. Make it consistently a single space,
per earlier decisions. This means contributors won't be confused by
misleading precedence.
This CL doesn't use go/doc to parse. It only addresses // comments.
It was generated with:
$ perl -i -npe 's,^(\s*// .+[a-z]\.) +([A-Z]),$1 $2,' $(git grep -l -E '^\s*//(.+\.) +([A-Z])')
$ go test go/doc -update
Change-Id: Iccdb99c37c797ef1f804a94b22ba5ee4b500c4f7
Reviewed-on: https://go-review.googlesource.com/20022
Reviewed-by: Rob Pike <r@golang.org>
Reviewed-by: Dave Day <djd@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Syscall getdents64 is relatively new in linux/mips64, only since kernel
version 3.10. To support older kernel, syscall getdents is used for
mips64.
Change-Id: I892b05dff7d93e7ddb0d700abd6a5e6d4084ab4c
Reviewed-on: https://go-review.googlesource.com/14449
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Minux Ma <minux@golang.org>
Change parameter name from uid to gid, to fix an obvious copy-paste
error.
Change-Id: Iba13a45c87fde9625b82976a7d7901af4b705230
Reviewed-on: https://go-review.googlesource.com/16474
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
There is no SYS_INOTIFY_INIT on linux/arm64, only SYS_INOTIFY_INIT1.
Change-Id: I97f430f2c2b910fb19dce495ff1adf591b8634fc
Reviewed-on: https://go-review.googlesource.com/9870
Run-TryBot: Minux Ma <minux@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Dave Cheney <dave@cheney.net>
This avoids hanging when a Go program uses a FUSE filesystem and the
dup system call has to close a file descriptor. When dup uses
RawSyscall then the goroutine calling dup will occupy a scheduler slot
(a p structure) during the call, and may block waiting for some other
goroutine to respond to the close call on the FUSE filesystem.
Changing to Syscall avoids the problem. This makes Dup a tiny bit
slower but is quite unlikely to make a difference for any real
programs.
Fixes#10202.
Change-Id: If6490a8f9b3c9cfed6acbfb4bfd1eaeac62ced17
Reviewed-on: https://go-review.googlesource.com/8095
Reviewed-by: Rob Pike <r@golang.org>
Fixes newly introduced test on linux/arm64 because linux/arm64 doesn't
have the getpgrp syscall. Getpgid(0) is documented to be equivalent to
Getpgrp.
Change-Id: I8f30f4f8de8c32fe04a29c9c4a9330d4e4e6b46d
Signed-off-by: Shenghou Ma <minux@golang.org>
Reviewed-on: https://go-review.googlesource.com/8022
Reviewed-by: Aram Hăvărneanu <aram@mgk.ro>
Reviewed-by: Rob Pike <r@golang.org>
Updates #9974
This proposal tackles the body of syscalls which have been replaced,
and are now deprecated in linux. This is needed for the arm64 port as
arm64 is the first linux architecture to remove the "legacy" forms of
these syscalls.
The *AT variants were added in kernel 2.6.16, so well before our 2.6.23
cutoff (hey, it'll even work on RHEL5).
Discussion: https://groups.google.com/forum/#!topic/golang-dev/zpeFtN2z5Fc
Change-Id: I473a7c9a295d6f776fcdc75dcce06cbe9e3564ee
Reviewed-on: https://go-review.googlesource.com/5837
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Updates #9974
This proposal moves the definition of Dup2 from the generic syscall_linux.go
to the GOOS specific variants. This is in preparation for the arm64 port.
For all existing platforms Dup2 is not affected. When arm64 is added we'll use
either a forwarding method to Dup3 or
//sysnb Dup2(oldfd int, newfd int) (err error) = SYS_DUP3
Because mksycall.pl does not sort symbols before generating the output file
the diff includes some unavoidable code moves as Dup2 is processed latter in
the run.
Discussion: https://groups.google.com/forum/#!topic/golang-dev/zpeFtN2z5Fc
Change-Id: Icdedf55bb29e749c4230e1ee371bf9d0bd0cfb38
Reviewed-on: https://go-review.googlesource.com/5835
Reviewed-by: Minux Ma <minux@golang.org>
Run-TryBot: Dave Cheney <dave@cheney.net>
Updates #9974
This proposal moves the definition of Pipe an Pipe2 from the generic
syscall_linux.go to the GOOS specific variants. This is in preparation
for the arm64 port.
For platforms where pipe2(2) is not supported in the minimum 2.6.23 kernel,
amd64 and 386, we retain pipe(2). For all other platforms pipe(2) is removed
and Pipe forwards to pipe2(2).
Because mksycall.pl does not sort symbols before generating the output file
the diff includes some unavoidable code moves as Pipe and Pipe2 are processed
latter in the run.
Discussion: https://groups.google.com/forum/#!topic/golang-dev/zpeFtN2z5Fc
Change-Id: Ie26d6761eeb9760dbaff974ee8bc0d57a9ceaee4
Reviewed-on: https://go-review.googlesource.com/5833
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Given:
p := alloc()
fn_taking_ptr(p)
p is NOT recorded as live at the call to fn_taking_ptr:
it's not needed by the code following the call.
p was passed to fn_taking_ptr, and fn_taking_ptr must keep
it alive as long as it needs it.
In practice, fn_taking_ptr will keep its own arguments live
for as long as the function is executing.
But if instead you have:
p := alloc()
i := uintptr(unsafe.Pointer(p))
fn_taking_int(i)
p is STILL NOT recorded as live at the call to fn_taking_int:
it's not needed by the code following the call.
fn_taking_int is responsible for keeping its own arguments
live, but fn_taking_int is written to take an integer, so even
though fn_taking_int does keep its argument live, that argument
does not keep the allocated memory live, because the garbage
collector does not dereference integers.
The shorter form:
p := alloc()
fn_taking_int(uintptr(unsafe.Pointer(p)))
and the even shorter form:
fn_taking_int(uintptr(unsafe.Pointer(alloc())))
are both the same as the 3-line form above.
syscall.Syscall is like fn_taking_int: it is written to take a list
of integers, and yet those integers are sometimes pointers.
If there is no other copy of those pointers being kept live,
the memory they point at may be garbage collected during
the call to syscall.Syscall.
This is happening on Solaris: for whatever reason, the timing
is such that the garbage collector manages to free the string
argument to the open(2) system call before the system call
has been invoked.
Change the system call wrappers to insert explicit references
that will keep the allocations alive in the original frame
(and therefore preserve the memory) until after syscall.Syscall
has returned.
Should fix Solaris flakiness.
This is not a problem for cgo, because cgo wrappers have
correctly typed arguments.
LGTM=iant, khr, aram, rlh
R=iant, khr, bradfitz, aram, rlh
CC=dvyukov, golang-codereviews, r
https://golang.org/cl/139360044
