| Age | Commit message (Collapse) | Author |
|---|
|
Add POSIX AIO support
POSIX AIO is a standard for asynchronous file I/O. Read, write, and
fsync operations can all take place in the background, with completion
notification delivered by a signal, by a new thread, by kqueue, or not
at all.
The SigEvent class, used for AIO notifications among other things, is
also added.
|
|
|
|
|
|
|
|
POSIX AIO is a standard for asynchronous file I/O. Read, write, and
fsync operations can all take place in the background, with completion
notification delivered by a signal, by a new thread, by kqueue, or not
at all. This commit supports all standard AIO functions. However,
lio_listio is disabled on macos because it doesn't seem to work, even
though the syscall is present.
The SigEvent class, used for AIO notifications among other things, is
also added.
Also, impl AsRef for TimeVal and TimeSpec
|
|
|
|
|
|
r? @fiveop
|
|
|
|
|
|
Also, add trait TimeValLike, so some code can be shared between TimeSpec
and TimeVal.
|
|
test_getsockname used an IPv4 socket and assumed that localhost was
"127.0.0.1". But that assumption doesn't hold on IPv6-only hosts or on
shared-IP FreeBSD jails. Unfortunately, the Rust standard library
doesn't provide a good way to resolve localhost. So change the test to
use a unix-domain socket instead.
|
|
There were multiple errors regarding Unix domain sockets:
* UnixAddr::path assumed that gethostbyname and similar functions would
include the terminating null as part of len. That is not universally
true. In fact, POSIX only guarantees that len will be at least large
enough to store the non-null-terminated path. So it could be larger or
smaller than nix was assuming. Since abstract sockets' paths are not
strings, we can't modify gethostbyname. Instead, I implemented the fix in
UnixAddr::path and UnixAddr::new. I clarified the documentation too.
* SockAddr::as_ffi_pair contained a Linuxism.
* sockaddr_storage_to_addr forgot to adjust sun_len when creating a UnixAddr
|
|
Sometimes, on a heavily laden system, select would timeout and the test would
fail. Fix it by lengthening the timeout to 10s.
|
|
Get rid of a few test compilation warnings
Cleans up after 0fa7250b9575a2746519a854d0138c4c8255f109 and 40351db00da6ee8c904f47599ee692a85e3d96ef.
|
|
|
|
|
|
|
|
unistd: Redesign the enum returned by fork()
This changes the name of the enum returned by `fork()` to `ForkResult`,
and changes the `Parent` variant to be struct-like.
The result can be matched like
use nix::unistd::ForkResult::*;
match fork().unwrap() {
Parent { child } => { ... }
Child => { ... }
}
using the shorthand matching syntax for struct-like enum variants.
This is a breaking change.
|
|
This commit changes the name of the enum returned by `fork()` to
`ForkResult`, and changes the `Parent` variant to be struct-like.
The result can be matched like
use nix::unistd::ForkResult::*;
match fork().unwrap() {
Parent { child } => { ... }
Child => { ... }
}
using the shorthand matching syntax for struct-like enum variants.
This is a breaking change.
|
|
Fixes #329
|
|
Replace a busy loop with a call to `pause(2)`.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The returned length of AF_UNIX sockaddrs is significant, and generally
does not match the length of the entire structure. For filesystem
sockets, this is ignorable because the path is also NUL-terminated, but
for unbound sockets (e.g., a socketpair) or abstract-namespace
sockets (a Linux extension where the address is an arbitrary
bytestring), we need to keep track of the length.
Fixes #177. Also add a UnixAddr::new_abstract function and some better
handling of abstract-namespace socket addresses to fix #169.
|
|
|
|
|
|
The best specification for control message layout appears to be
[RFC 2292, section 4](https://tools.ietf.org/html/rfc2292#section-4),
despite this not being a wire protocol. These definitions have also been
checked against glibc 2.19 <bits/socket.h> and Linux 4.0
<linux/socket.h>, and tested on Debian 8.1 and FreeBSD 10.2 x86_64.
The API differs a bit from the cmsg(3) API for type-safety reasons (and
also because the cmsg(3) API is terrible). See test/sys/test_socket.rs
for an example.
Only supports SCM_RIGHTS at the moment.
Fixes #88.
|
|
This reverts commit 046af7d1ba82506f9bc48e62ac0584361025fc02.
|
|
|
|
The best specification for control message layout appears to be
[RFC 2292, section 4](https://tools.ietf.org/html/rfc2292#section-4),
despite this not being a wire protocol. These definitions have also been
checked against glibc 2.19 <bits/socket.h> and Linux 4.0
<linux/socket.h>, and tested on Debian 8.1 and FreeBSD 10.2 x86_64.
The API differs a bit from the cmsg(4) API for type-safety reasons (and
also because the cmsg(4) API is terrible). See test/sys/test_socket.rs
for an example.
Only supports SCM_RIGHTS at the moment.
Fixes #88.
|
|
This change also adds macro usages in the tests. Nothing is asserted
but the use of the macros provides a basic compile-time check that
is otherwise missing.
Signed-off-by: Paul Osborne <osbpau@gmail.com>
|
|
|
|
|
|
* Fixed an unused_import error in `termios.rs` for Android.
* Fixed undefined references to `preadv` and `pwritev` for Android -
At least they don't exist from API level 3 to 21.
* Fixed the uid > 0 and gid > 0 checks in `stat`'s tests - Running the
tests by root is possible, especially when running on a rooted Android
device.
Those changes made rust-nix buildable (again) on Android. All the tests
passed as well.
|
|
* Extend the enums in WaitStatus to include all process states (signaled,
stopped, exited, continued).
* Decode status from waitpid
* Return appropate WaitStatus
* Update tests to use the new WaitStatus
* Add new tests for specific status values
|
|
|
|
Fix the boundary condition so that we never call
Rng::gen_range(64, 64).
Helps #144
|
|
|
|
|
|
This commit provides a new implementation for ioctl that is much
more generic, allowing for clients to use send any manner of ioctl
requests at special files. The implementation provides two main features
that help to raise the level of abstraction over that provided by libc.
1. The module now provides functions that provide the same functionality
as the linux kernel _IO* macros. These are used frequently in the
linux kernel for building ops for ioctls. The use of these helper
functions are not required.
2. Functions are provided for the 3 main types of ioctl usage patterns
(read, write, and execute). For many subystems, the read() call
which returns a Result<T> and the write calls taking a &T provide
a nice interface.
All of the methods wrapping ioctl are unsafe and will probably need to
remain that way unless knowledge of the semantics of every possible
ioctl call are added to the nix library. The best that exists for
ioctls are some conventions around the op, but even these conventions
are really only used for newer devices added to the kernel.
This change resolves #108
|
|
|
|
|
|
|
|
|
|
|
