* Add more cfg aliases
Add cfg aliases for linux_android, bsd, and freebsdlike. Use them in
many places, though not everywhere they could theoretically be used.
Fixes#2188
* Use apple_targets in build.rs
Co-authored-by: SteveLauC <stevelauc@outlook.com>
* whitespace
Co-authored-by: SteveLauC <stevelauc@outlook.com>
* Define a "solarish" target alias.
* Describe cfg aliases in CONVENTIONS.md
* solarish in line 803
* solarish in line 845
* fix fmt
---------
Co-authored-by: SteveLauC <stevelauc@outlook.com>
This commit adds support for the signal timer mechanism in POSIX, the
mirror to timerfd on Linux.
Resolves#1424
Signed-off-by: Brian L. Troutwine <brian@troutwine.us>
parking_lot provides synchronization primitives which aren't
poisoned on panic. This makes it easier to determine which tests
are failing, as a test failure no longer causes all subsequent tests
using that mutex to fail.
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
test code breaks on fedora 33
```
$ cargo test
failures:
---- sys::test_socket::recvfrom::udp_offload::gro stdout ----
thread 'sys::test_socket::recvfrom::udp_offload::gro' panicked at 'called `Result::unwrap()` on an `Err` value: ParseError("Extra junk after valid version: _64")', test/sys/test_socket.rs:292:13
```
this is due underscore in release string( arch/x86_64), which is not supported by semver.
```
$ uname -r
5.8.14-300.fc33.x86_64
```
Replace the underscore with hypen to provide a consistent sematic.
Travis is now using Seccomp, and Docker's default Seccomp policy
disables execveat (though, weirdly, not fexecve). It also prohibits any
operations on AF_ALG sockets.
While I'm here, replace close/dup with dup2, which is more reliable.
Also, drop the fork mutex earlier. This way all of the exeve tests will
run, even if one fails.
https://docs.docker.com/engine/security/seccomp/
* kmod tests must run exclusively, because they load and unload a module
with a constant name.
* A few tests were doing some variant of chdir, but weren't taking the
CWD_MTX.
* The kmod tests read files by path relative to CWD, so they need the
CWD_MTX. But they don't need it exclusively, so convert the CWD_MTX
into an RwLock.
* Tests that do change the cwd need to change it back when they're done.
* On Linux, it requires the CAP_SYS_PACCT capability.
* Reenable the test on FreeBSD, because our FreeBSD CI environment is no
longer jailed (since we switched from BuildBot to CirrusCI), but check
at runtime whether the process is jailed.
* test_acct needs the FORK_MTX because it uses Command::new .
* Fix a race condition. acct(2) isn't synchronous. It starts a kernel
thread but does not wait for it to become ready. Fix it by running
the test command within the polling loop.
It's not sufficient to check for root privileges, because a
containerized process may have root's euid but still lack important
capabilities. Fix these tests by checking for the CAP_SYS_MOD
capability.
It's not sufficient to check for root privileges. In a container, the
euid may be root even though the user lacks some capabilities. Replace
this test's root check with a check for the CAP_NET_ADMIN capability
instead.
This is a lower-level interface than `std::fs::ReadDir`. Notable differences:
* can be opened from a file descriptor (as returned by `openat`, perhaps
before knowing if the path represents a file or directory). Uses
`fdopendir` for this, available on all Unix platforms as of
rust-lang/libc#1018.
* implements `AsRawFd`, so it can be passed to `fstat`, `openat`, etc.
* can be iterated through multiple times without closing and reopening the
file descriptor. Each iteration rewinds when finished.
* returns entries for `.` (current directory) and `..` (parent directory).
* returns entries' names as a `CStr` (no allocation or conversion beyond
whatever libc does).
820: Change AioCb to primarily use Bytes instead of Rc<[u8]> r=Susurrus a=asomers
`Rc<[u8]>` isn't a very good buffer type to use for aio. For one thing, it lacks interior mutability. For another, a single `Rc<[u8]>` can't be carved up into smaller buffers of the same type. `Bytes` and `BytesMut` fix both problems. This PR removes the ability to construct an `AioCb` from `Rc<[u8]>` and adds the ability to construct one from `Bytes`, `BytesMut`, or raw pointers (for consumers who need even more flexibility). At this stage, the PR has the following warts:
1. A hack is necessary to force small `Bytes` buffers to allocate on the heap. I plan to fix this with an enhancement to the bytes crate.
2. The `AioCb::buffer` method is necessary due to a deficiency in the tokio-core crate. Once I fix that, then only `AioCb::into_buffer`will need to be public.
It's not actually safe to read into an `Rc<[u8]>`. It only worked
because of a coincidental `unsafe` block. Replace that type with
`BytesMut` from the bytes crate. For consistency's sake, use `Bytes`
for writing too, and completely remove methods relating to `Rc<[u8]>`.
Note that the `AioCb` will actually own the `BytesMut` object. The
caller must call `into_buffer` to get it back once the I/O is complete.
Fixes#788
This was doing testing for errno constants and a few other
types that is no longer necessary now that these types are
all tested within the libc project itself.
