* Fix the value of optlen as passed to libc::setsockopt
Ever since the sockopt macro was written in
979faf0976 we've been passing the wrong
size for u8 and usize sockopts.
* Add tests for sockopt::IpMulticastTtl #2073
On Linux and Android, sockopt::IpMulticastTtl can be set for both IPv4
and IPv6 sockets, but not on FreeBSD.
---------
Co-authored-by: Alan Somers <asomers@gmail.com>
Co-authored-by: Simon B. Gasse <sgasse@users.noreply.github.com>
From [tcp](https://man7.org/linux/man-pages/man7/tcp.7.html) man page:
```
TCP_FASTOPEN_CONNECT (since Linux 4.11)
This option enables an alternative way to perform Fast
Open on the active side (client). When this option is
enabled, connect(2) would behave differently depending on
if a Fast Open cookie is available for the destination.
If a cookie is not available (i.e. first contact to the
destination), connect(2) behaves as usual by sending a SYN
immediately, except the SYN would include an empty Fast
Open cookie option to solicit a cookie.
If a cookie is available, connect(2) would return 0
immediately but the SYN transmission is deferred. A
subsequent write(2) or sendmsg(2) would trigger a SYN with
data plus cookie in the Fast Open option. In other words,
the actual connect operation is deferred until data is
supplied.
Note: While this option is designed for convenience,
enabling it does change the behaviors and certain system
calls might set different errno values. With cookie
present, write(2) or sendmsg(2) must be called right after
connect(2) in order to send out SYN+data to complete 3WHS
and establish connection. Calling read(2) right after
connect(2) without write(2) will cause the blocking socket
to be blocked forever.
The application should either set TCP_FASTOPEN_CONNECT
socket option before write(2) or sendmsg(2), or call
write(2) or sendmsg(2) with MSG_FASTOPEN flag directly,
instead of both on the same connection.
Here is the typical call flow with this new option:
s = socket();
setsockopt(s, IPPROTO_TCP, TCP_FASTOPEN_CONNECT, 1, ...);
connect(s);
write(s); /* write() should always follow connect()
* in order to trigger SYN to go out. */
read(s)/write(s);
/* ... */
close(s);
```
* Added object for automatic unlock on drop.
* Added appropriate changelog file.
* Fixed doc error on `Flock`.
* Requested changes to make `fcntl::flock` private and OwnedFd instead of RawFd.
* Indent fix.
* Removed doc links to private item, updated imports.
* More import fixes.
* Format changes.
* Remove unused import for redox.
* Added `Flockable` trait per discussions, added tests for `Flock`.
* Simplified flock error conditionals.
* Added missing cfg flags for `Flockable`.
* Added missing cfg flags for impl of `Flockable` on `File`.
* Update src/fcntl.rs
Co-authored-by: SteveLauC <stevelauc@outlook.com>
* Updated `Flockable` comment per suggestion.
* Finalized `FlockArg` as enum and removed TODO accordingly, removed flock wrapper entirely.
* Implemented `Flockable` for `OwnedFd` as well.
* Fixed linting error.
* Updated changelog accordingly.
* Corrected errno logic in `Flock::lock`.
* Properly dropped inner type for `Flock`.
* Replaced `ManuallyDrop` with `Option` as inner `Flock` type to avoid double-free.
* Fixed linting errors.
* Removed unnecessary cfg condition, updated documentation.
* Modified Flock behavior for drop() and unlock().
* Reverted changes to original `flock()` and `FlockArg` for deprecation period, added EINVAL case if the unlock operation is given to `Flock::lock()`.
* Refactored `Flock` to wrap T directly and avoid a double-free after `unlock()` is used.
* Fixed linting errors.
* More linting fixes.
* Fixed example code for `Flock::unlock()`.
* Made requested changes.
* Removed duplicate import.
* Format change.
---------
Co-authored-by: SteveLauC <stevelauc@outlook.com>
From [man page](https://man7.org/linux/man-pages/man7/ip.7.html):
```
IP_BIND_ADDRESS_NO_PORT (since Linux 4.2)
Inform the kernel to not reserve an ephemeral port when
using bind(2) with a port number of 0. The port will
later be automatically chosen at connect(2) time, in a way
that allows sharing a source port as long as the 4-tuple
is unique.
```
* 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>
* Added FreeBSD's SCM_REALTIME and SCM_MONOTONIC into sys::socket::ControlMessageOwned.
* Creating a SocketTimestamp enum for the SO_TS_CLOCK setsockopt for FreeBSD.
* Fixing whitespace
* Fixing CI error on cfg attributes
* Removing legacy doc attributes
* Formatting cleanup
* Updating changelog
* Adding tests for new TsClock setsockopt enum and the two new packet timestamp control messages for FreeBSD
* Replacing an assert_eq with an assert in new tests.
* Removing qemu ignore for new FreeBSD tests
* Giving new FreeBSD timestamping tests each a unique socket
* Updating monotonic packet timestamp test to account for monotonicity
* Moving test ports again to line up with changes in #2196
* Attempting checks again
* Wrapping ptr::read_unaligned calls in unsafe blocks
* Add fanotify API wrappers
* Review/fix enum comments
* Rename has_compile_version to check_version
* Fix lint unsafe_block_in_unsafe_fn
* Rename fanotify OFlags to EventFFlags
* Use existing function at_rawfd from fcntl
* Add missing feature guard for docs
* Change FanotifyEvent struct to a simple wrapper over libc structure
* Change FanotifyResponse struct to a simple wrapper over libc structure
* Cast pointer with function 'cast' instead of 'as'
* Add FAN_REPORT_PIDFD and FAN_REPORT_TID fanotify init flags
* refactor: change dirfd arg of xxxat() to Option<fd>
* changelog
* gate at_rawfd() with features fs or process
* fix import
* fix test on Android
* simplify at_rawfd()
* update changelog
* Relax lifetime requirements for PollFd::new
Fixes#2118
* Take BorrowedFd as the argument for PollFd::new
&AsFd didn't work because there are 'static types, like std::fs::File,
which implement AsFd. The created BorrowedFd type within the
PollFd::new method would have a very brief lifetime, but the PhantomData
would capture the lifetime of the std::fs::File. Taking BorrowFd<'fd>
argument makes the lifetime explicit.
* fix legacy comment
---------
Co-authored-by: Steve Lau <stevelauc@outlook.com>
* Relax lifetime requirements for FdSet::{insert, remove, contains}
Fixes#2130
* Take BorrowedFd as the argument for FdSet::{insert, remove, contains}
&AsFd doesn't work because there are 'static types, like std::fs::File,
which implement AsFd.
* fix changelog & remove unused entries
* fix wrong PR number
---------
Co-authored-by: Steve Lau <stevelauc@outlook.com>
* Cast pointers more carefully
Prefer ptr::{cast, cast_const, cast_mut} over `as`. The latter makes it
too easy to accidentally change both the type and mutability when
changing only one was intended.
This exercise caught an unintended mutability cast in one function, the
BSD version of sendfile. In this case there's no UB because it isn't
possible for anything else to get a reference to the data that was
incorrectly cast.
There was also a type cast that wasn't guaranteed to be correct (but
probably was) due to memory layout guarantees in if_nametoindex.
* Remove unnecessary cast in UnixCredentials::groups
* unistd: make getpeereid take AsFd rather than a raw FD
`AsFd` returns a `BorrowedFd` that is correct by construction,
meaning that this API rejects more error states at the type
level (such as passing in `fd = -1`).
* CHANGELOG: record changes
* unistd: pass AsFd impl by value
* unistd: more AsFd usage
* CHANGELOG: update changes
* Changed `name` parameter of `mq_open` and `mq_unlink` to be generic over `NixPath`.
* Updated with PR link.
* Fixed docs example code.
* Fixed docs example code for real this time.
