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.
|
|
|
|
This introduces a wrapper for fchdir(2), allowing a process to change
directory based on an open file descriptor.
The underlying function is available in libc crate since 0.2.20.
|
|
Baed on discussions on the related PR, sethostname now takes
an `S: AsRef<OsStr>` in order to allow for a greater range of inputs
that allow for a more fluid interface.
Signed-off-by: Paul Osborne <osbpau@gmail.com>
|
|
Previously gethostname just mutated a buffer. We now provide
a slightly more usable (but still allocation free) API that ensures
that the returned buffer is NUL-terminated. We give back a `&CStr`
instead of requiring that the user do all of the conversions from
`&[u8]` when we know we are dealing with a `&CStr`.
Signed-off-by: Paul Osborne <osbpau@gmail.com>
|
|
Although the underlying C API does take a pointer to a set of
characters, it is a requirement of almost every operating system
that these bytes not contain a premature NUL character or other
special characters. In other words, you want a `&str`. Changing
this to make the API make a bit more sense.
Signed-off-by: Paul Osborne <osbpau@gmail.com>
|
|
fix #480 and add simple test cases for that.
r? @fiveop
|
|
Fix ControlMessage::encode_into when encoding multiple messages
copy_bytes updates dst so that it points after the bytes that were just
copied into it. encode_into did not advance the buffer in the same way
when encoding the data.
See #473
|
|
fix build for eventfd feature
|
|
Always use libc's socketaddr_storage
Since rust-lang/rust#23425 is closed, we no longer need the special
case.
|
|
Since rust-lang/rust#23425 is closed, we no longer need the special
case.
|
|
|
|
|
|
|
|
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
|
|
Add support for system control sockets for XNU
I added support for macOS and iOS system sockets, which can be used to control the kernel as described [here](https://developer.apple.com/library/content/documentation/Darwin/Conceptual/NKEConceptual/control/control.html).
To do this, I had to add in support for `ioctl` on those platforms, so I added in `ioctl` support for all BSD-based platforms. The API seems to be the same between [xnu](https://opensource.apple.com/source/xnu/xnu-3248.60.10/bsd/sys/ioccom.h.auto.html), [FreeBSD](https://github.com/freebsd/freebsd/blob/master/sys/sys/ioccom.h), [NetBSD](https://ftp.netbsd.org/pub/NetBSD/NetBSD-current/src/sys/sys/ioccom.h), [OpenBSD](http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/sys/ioccom.h?rev=1.5&content-type=text/x-cvsweb-markup) and [Dragonfly BSD](http://gitweb.dragonflybsd.org/dragonfly.git/blob/HEAD:/sys/sys/ioccom.h).
I added a test that runs on macOS and iOS for the functionality. Let me know if I need to make any changes!
|
|
Ensure child stack passed to clone is 16 byte aligned.
The current implementation assumes that the array passed by the caller is word aligned (which I don't think Rust guarantees for [u8]) and a multiple of the word size.
|
|
Added getpgrp
|
|
|
|
|
|
|
|
|
|
|
|
|
|
r? @fiveop
|
|
|
|
copy_bytes updates dst so that it points after the bytes that were just
copied into it. encode_into did not advance the buffer in the same way
when encoding the data.
|
|
|
|
|
|
|
|
|
|
Timespec
Make TimeVal an opaque Newtype, and add another Newtype for Timespec.
|
|
Also, fix the TimeSpec::cmp and TimeVal::cmp methods, and fix some
formatting
|
|
|
|
Also, add trait TimeValLike, so some code can be shared between TimeSpec
and TimeVal.
|
|
|
|
By exposing its members, nix allowed users to create denormalized
TimeVals, which don't work with the derived() Eq and Ord methods.
Better to make TimeVal opaque, so it will always be normalized.
|
|
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
|
|
Change SigFlags into an enum.
Addresses #459.
This is a breaking change.
Should SigFlags be renamed to something more sensible?
|
|
Change KEvent to treat udata as an intptr_t instead of a uintptr_t.
This matches NetBSD's C definitions. Other operating systems define
it as void*, despite not really being a pointer, but none actually
define it as uintptr_t. Better to be right on NetBSD and wrong
everywhere else than wrong everywhere. Plus, it's what mio expects.
Please include this PR in nix 0.8.0
|
|
|
|
|
|
This matches NetBSD's C definitions. Other operating systems define
it as void*, despite not really being a pointer, but none actually
define it as uintptr_t. Better to be right on NetBSD and wrong
everywhere else than wrong everywhere. Plus, it's what mio expects.
|
|
|
|
|
|
|
|
|
|
|
|
|