Age | Commit message (Collapse) | Author |
|
1401: cleanup: remove redundant unwrap in Dir::from_fd r=asomers a=scottlamb
Co-authored-by: Scott Lamb <slamb@slamb.org>
|
|
|
|
|
|
Co-authored-by: Dominik Hassler <hadfl@omnios.org>
Co-authored-by: Joshua M. Clulow <josh@sysmgr.org>
|
|
Apparently the Linux kernel can return smaller sizes when the value in
the last element of sockaddr_ll (`sll_addr`) is smaller than the
declared size of that field.
|
|
libc removed it in version 0.2.87.
https://github.com/rust-lang/libc/pull/2079
|
|
After the child returns from a fork() of a multi-threaded process, it is
undefined behaviour to call non-async-signal-safe functions according to
POSIX. Since forkpty() is implemented in terms of fork(), those
restrictions should apply to it too.
|
|
|
|
|
|
|
|
|
|
|
|
Test fix obtained from
https://github.com/nix-rust/nix/issues/1384#issuecomment-774708486
|
|
This is useful to allow returning an iterator based on a directory iterator
without needing a self-referential struct.
|
|
1387: fix(unsafe): remove unnecessary unsafe r=asomers a=matu3ba
libc 0.2.82 exposes status signals with macros generating safe functions
Co-authored-by: Jan Philipp Hafer <jan.hafer@rwth-aachen.de>
|
|
closes #1380
libc 0.2.82 exposes status signals with macros generating safe functions
|
|
|
|
1382: Don't implement Clone on Dir, SignalFd, and PtyMaster r=asomers a=asomers
Since they close their file descriptors on Drop, it's almost impossible
to use Clone without creating a double-close situation.
Also, check for EBADF in SignalFd::drop and Dir::drop.
Co-authored-by: Alan Somers <asomers@gmail.com>
|
|
Since they close their file descriptors on Drop, it's almost impossible
to use Clone without creating a double-close situation.
Also, check for EBADF in SignalFd::drop and Dir::drop.
|
|
|
|
Since rust-lang/libc is now using the correct data type on s390x (c_uint) for
statfs constants, the s390x exemptions in the statfs MAGIC definitions like
EXT4_SUPER_MAGIC can be removed.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
|
|
|
|
Need to use the right cfg option for the conditional compilation.
target_os is the right option to use when targeting FreeBSD. target_env
was used before which seems to be a typo.
|
|
Closes #1339.
|
|
Allow nix to compile on Fuchsia by conditionally avoiding libc
functionality that does not exist for Fuchsia.
|
|
Adds a high level `Persona` bitflags enum, as well as `personality::get()`
and `personality::set()` for interacting with `libc::personality()`.
|
|
It's been unused since PR #133
|
|
On Cirrus CI, this test would sometimes fail, probably due to pause
being awoken by SIGRT_1. sigwait is a superior alternative to pause.
Fixes #1354
|
|
Travis has been super-slow lately (> 6 hours per build). Cirrus is much
faster: about 20 minutes. Cirrus also has slightly better test
coverage, mainly because it doesn't use SECCOMP.
Also,
* Fix the Redox CI build. The old Travis configuration didn't actually
build for Redox, so we never noticed that Redox can't be built with a
stable compiler. Thanks to @coolreader18 for finding this.
* Disable the udp_offload tests on cross-tested platforms. These tests
are failing with ENOPROTOOPT in Cirrus-CI. I suspect it's due to a
lack of support in QEMU. These tests were skipped on Travis because
its kernel was too old.
* Fix require_kernel_version on Cirrus-CI. Cirrus reports the Linux
kernel version as 4.19.112+, which the semver crate can't handle.
* Fix test_setfsuid on Cirrus. When run on Cirrus, it seems like the
file in /tmp gets deleted as soon as it's closed. Probably an
overzealous temporary file cleaner. Use /var/tmp, because no
temporary file cleaner should run in there.
* Skip mount tests on Cirrus. They fail for an unknown reason.
Issue #1351
* Skip the AF_ALG tests on Cirrus-CI
Issue #1352
|
|
|
|
There were two problems discovered with the `recvmmsg(2)` implementation
that this changeset attempts to fix:
1. As mentioned in nix-rust/issues/1325, `recvmmsg(2)` can return fewer
messages than requested, and
2. Passing the return value of `recvmmsg(2)` as the number of bytes in
the messages received is incorrect.
This changeset incorporates the proposed fix from nix-rust/issues/1325,
as well as passing the correct value (`mmsghdr.msg_len`) for the number
of bytes in a given message.
|
|
See https://github.com/rust-lang/libc/issues/1848 in which this type is
changing from i32 to i64; the change is being announced via this
deprecation.
|
|
See https://github.com/rust-lang/libc/issues/1501 in which this type's
trait implementations are being removed; the change is being announced
via this deprecation.
|
|
Without this commit, nix doesn't currently compile on OpenBSD. Some keepalive
settings can be set globally, but not per-socket (see e.g.
https://bugzilla.mozilla.org/show_bug.cgi?id=970550 for some digging into this).
Since it seems that NetBSD and DragonflyBSD have these settings, it makes more
sense to only exclude them on OpenBSD rather than include them on (pretty much
every) other operating systems.
|
|
|
|
|
|
1303: Renable the ScmTimestamp test on FreeBSD i386 r=posborne a=asomers
The relevant bug was fixed in 11.3-RELEASE
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=222039
Co-authored-by: Alan Somers <asomers@gmail.com>
|
|
Fix deprecation warning from libc update.
|
|
The relevant bug was fixed in 11.3-RELEASE
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=222039
|
|
1296: DragonFlyBSD: use __errno_location now provided by the libc crate r=asomers a=cmusser
Dragonfly recently (in Dfly 5.8) added an `__errno_location()` function to make it easy to access the thread-local `errno` variable. This has been exposed in rust-libc as of 0.2.77. This PR uses that functionality instead of the locally compiled C language shim, which the PR removes. One issue is backwards compatibilty. It requires 0.2.77 libc and DragonFly 5.8. Not sure how to gracefully handle older DragonFly versions, although I'm also not sure doing so is worth it.
Co-authored-by: Chuck Musser <cmusser@sonic.net>
|
|
1300: Add PTRACE_SYSEMU and PTRACE_SYSEMU_SINGLESTEP support r=asomers a=voidc
Closes #1249.
I think @jabedude was working on this, but as there was no progress since May I went ahead and implemented it myself.
I'm not completely sure about the cfg gates. Could we enable the functions for more targets?
I'm also open for suggestions of better names for the new functions.
Co-authored-by: Dominik Stolz <d.stolz@tum.de>
|
|
1293: Mark nix::unistd::fork as unsafe. r=asomers a=vi
Fix tests. No change in documentation.
Resolves #1030.
Don't forget to bump major version number to `0.19`.
Co-authored-by: Vitaly _Vi Shukela <vi0oss@gmail.com>
|
|
|
|
|
|
This obviates the need for the C language shim for this variable,
which this commit removes.
|
|
Fix tests. No change in documentation.
Resolves #1030.
|
|
(https://github.com/rust-lang/libc/pull/1870)
Denied lint being overridden instead of unsafe block removed
to preserve compatibility with old `libc` versions.
|
|
|
|
These were previously missing on musl and mips targets because of
missing definitions in the libc crate.
|
|
1278: Allow both &[CString] and &[&CStr] for sequence args in unistd r=asomers a=youknowone
I am new to this project, so I wonder this kind of change is acceptable for the project or not.
I agree taking `&[&CStr]` is the ideal way for the functions. But unfortunately, when users generate `CString`s from `&str`, they may get `&[CString]` rather than `&[&CStr]`
Unlike single `CString` to `&CStr` argument, there is no easy way to convert `&[CString]` to `&[&CStr]` without creating new vector.
By changing the type from `&[&CStr]` to `&[AsRef<CStr>]`, the args now can take both `&[CString]` and `&[&CStr]` without changing that much from nix.
Co-authored-by: Jeong YunWon <jeong@youknowone.org>
|