Modeled after FreeBSD's `reallocf(3)`: a version of `realloc(3)` that
frees the memory when reallocation fails, simplifying error handling in
many cases.
PR-URL: https://github.com/libuv/libuv/pull/2735
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
PR-URL: https://github.com/libuv/libuv/pull/2013
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Signed-off-by: Irek Fakhrutdinov <ifakhrutdinov@rocketsoftware.com>
The os390 epoll_wait implementation uses poll() to detect events in
both file descriptors and the message queue used for file system
events. The only message queue ID is always placed at the end of the
array passed to the poll() call. When the poll() call returns all FDs
and the message queue are checked for events by iterating through that
array. In order to distinguish the message queue from the FDs its ID
value is compared with the only message queue ID we have and if it
matches the message queue handler function is called.
When the message queue ID is relatively small, it may have the same
value as the value of one of the file descriptors. If this happens,
the message queue handler is called for the matching file descriptor,
and this call fails. The file descriptor ends up being unhandled and
this makes the next poll() call to return immediately. Eventually this
will happen again and again, leading to an infinite busy loop and high
CPU usage.
To prevent the incorrect interpretation of file descriptors as the
message queue, a new field has been added to the epoll event struct.
This field is checked instead of the ID value and the message queue
handler function is never called for file descriptors.
PR-URL: https://github.com/libuv/libuv/pull/2013
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
The way libuv handled closing of `uv_signal_t` handles with pending
events introduced an infidelity where `uv_loop_alive()` returned false
while the signal handle was still in the closing-but-not-closed state.
Fixes: https://github.com/libuv/libuv/issues/2721
PR-URL: https://github.com/libuv/libuv/pull/2722
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
The wrong names and name sizes were being used with `sysctl(2)`, which
in particular made the call to get the machine's CPU speed always fail
with ENOTDIR.
PR-URL: https://github.com/libuv/libuv/pull/2685
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Previously, this field was being used uninitialized, leading to the
uv_write call returning failure. This only happens in the rare fallback
path where UV_HANDLE_EMULATE_IOCP is set, meaning that the OVERLAPPED
opened pipe already was associated with an IOCP.
Fixes: https://github.com/libuv/libuv/issues/629
PR-URL: https://github.com/libuv/libuv/pull/2620
Reviewed-By: Saúl Ibarra Corretgé <s@saghul.net>
This commits adds support for recvmmsg() and sendmmsg() extensions to
recvmsg() and sendmsg() that allows the caller to receive and send
multiple message from a socket using a single system call. This has
performance benefits for some applications.
Co-authored-by: Ondřej Surý <ondrej@sury.org>
Co-authored-by: Witold Kręcicki <wpk@culm.net>
PR-URL: https://github.com/libuv/libuv/pull/2532
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Saúl Ibarra Corretgé <s@saghul.net>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
It was checking that the destination buffer was big enough to hold
the total capacity of the process title (the total storage of argv)
when instead it should be checking that it's big enough to hold
the _current_ process title, which is normally much shorter.
Fixes: https://github.com/libuv/libuv/issues/2666
PR-URL: https://github.com/libuv/libuv/pull/2668
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Only update the `process_title` struct on success, otherwise it points
to memory libuv doesn't own. Notably, libuv only owns `argv[0]` _after_
copying the argv table.
PR-URL: https://github.com/libuv/libuv/pull/2668
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Current UNIX systems define various defaults for
TCP_KEEPINTVL and TCP_KEEPCNT, which makes the time
between TCP_KEEPIDLE delay is reached and timeout
effectively occurs unpredictable (Linux: /proc/sys
/net/ipv4/tcp_keepalive_intvl;tcp_keepalive_probes).
Do the following: set TCP_KEEPINTVL to 1 second (same
as Win32 default) and TCP_KEEPCNT to 10 times (same
as Win32 hardcoded value).
Fixes: https://github.com/libuv/libuv/issues/2664
PR-URL: https://github.com/libuv/libuv/pull/2669
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
This lets us get rid of the system call wrappers and simplify
the fs event watcher code a great deal.
All system calls were added well before 2.6.32, our baseline kernel.
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Assume the presence of the eventfd2() system call on Linux. It was added
in 2.6.27 and our baseline is 2.6.32.
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
This commit checks that setting the close-on-exec and non-blocking flags
on the file descriptors succeeds, and closes the file descriptors when
it fails.
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Assume the presence of the pipe2() system call. It was added
in FreeBSD 10 and Linux 2.6.27.
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
This commit checks that setting the close-on-exec flag on the file
descriptors succeeds, and closes the file descriptors when it fails.
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Assume the presence of the SOCK_CLOEXEC flag. It was added in FreeBSD 10
and before Linux 2.6.32.
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Assume the presence of the dup3() system call on Linux. It was added
in 2.6.27 and our baseline is 2.6.32.
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Assume the presence of the accept4() system call on Linux. It was added
in 2.6.28 and our baseline is 2.6.32. That lets us get rid of the system
call wrapper and the fallback logic in uv__accept().
PR-URL: https://github.com/libuv/libuv/pull/2665
Reviewed-By: Santiago Gimeno <santiago.gimeno@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
On IBMi PASE, listen(pipe_fd, backlog=0) leads to "Connection refused" error
PR-URL: https://github.com/libuv/libuv/pull/2641
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
`uv__open_cloexec()` already returns a libuv error code in case
of failure, and not `-1` like syscalls do.
Fixes: https://github.com/nodejs/help/issues/2099
PR-URL: https://github.com/libuv/libuv/pull/2645
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Jameson Nash <vtjnash@gmail.com>
Set this to NULL just before disposing it to make mistakes more
painfully obvious, hopefully.
PR-URL: https://github.com/libuv/libuv/pull/2626
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
The mutex also needs to protect the access to the state->loop variable,
since that's owned by the child thread and will be destroyed as soon as
it processes our message.
Fixes: https://github.com/libuv/libuv/issues/2625
PR-URL: https://github.com/libuv/libuv/pull/2626
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
On IBMi PASE we need to call Qp2getifaddrs() to get the network
interface configurations. And to call QDCRLIND to get the physical
addresses.
PR-URL: https://github.com/libuv/libuv/pull/2614
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
It's been reported that calling this function causes the Core Graphics
framework to start reporting bogus values.
Commit 565cdd16 ('Revert "darwin: speed up uv_set_process_title()"')
attempted to fix this but apparently merely postponed the moment
when `CGDisplayPixelsWide()` and friends start reporting bogus values.
The Chromium code base mentions that calling `SetApplicationIsDaemon()`
prevents the HIServices framework from terminating the process when it
can't connect to launchservicesd.
Libuv itself doesn't use HIServices but it's possible that the libuv
user does. If said user doesn't call `SetApplicationIsDaemon()`, it's
possible this commit introduces an observable change in behavior.
The `SetApplicationIsDaemon()` call was introduced in commit 08e0e63f
("darwin: avoid calling GetCurrentProcess") from October 2013 to work
around a bug in macos 10.9 where the Activity Monitor showed the program
as "Not responding."
Fixes: https://github.com/libuv/libuv/issues/2566 (for real, hopefully)
Fixes: https://github.com/nodejs/node/issues/31328
PR-URL: https://github.com/libuv/libuv/pull/2593
Refs: https://cs.chromium.org/chromium/src/sandbox/mac/system_services.cc?l=26&rcl=a06d2fe5a279ddecd358d919d461080e2c53c92e
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Since this re-open logic is supposed to be transparent to the parent
program, this flag helps ensure that there are no unintended
side-effects. This flag is already the default (and only possible
option) on FreeBSD. Normally, Linux will also act as if you passed
NOCTTY. However, it appears that mac OS sometimes strips the parent
process of its controlling terminal if you don't pass this flag.
Fixes: https://github.com/libuv/libuv/issues/973
Fixes: https://github.com/JuliaLang/julia/issues/16398
PR-URL: https://github.com/libuv/libuv/pull/2619
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
This assert was stronger than necessary (we assert the actual condition
later in this function). And it's vulnerable to a race condition
occurring in uv__work_done where we drain the work queue but still have
a stale message on this object.
Fixes: https://github.com/libuv/libuv/issues/2610
PR-URL: https://github.com/libuv/libuv/pull/2615
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
This code already had special handling for decoding utf-8
characters correctly for the UCS-2 conhost
(trailing incomplete bytes and characters > 0xFFFF).
Rather than trying to duplicate that, we can simply
delete the duplicate code-path and selectively disable
the parts that do not apply.
PR-URL: https://github.com/libuv/libuv/pull/1965
Ref: https://github.com/JuliaLang/julia/issues/27267
Co-authored-by: Mustafa M. <mus-m@outlook.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
uv_fs_copyfile() calls fchmod() to change the target file's permissions
to the source file's permissions but that operation errors with EPERM on
CIFS shares unless they are mounted with the "noperm" option.
Since UNIX-style permissions don't make sense for CIFS anyway, let's
handle the error in libuv by recognizing that it's a CIFS share and
continuing when that is the case.
The same logic probably applies to (a subset of) FUSE file systems but
those haven't been whitelisted yet.
Fixes: https://github.com/libuv/libuv/issues/2596
Refs: https://github.com/nodejs/node/issues/31170
PR-URL: https://github.com/libuv/libuv/pull/2597
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Saúl Ibarra Corretgé <saghul@gmail.com>
On IBMi, msg_controllen of struct msghdr can not exceed 256.
The default UV__CMSG_FD_COUNT value 64 leads EINVAL error.
Set UV__CMSG_FD_COUNT to 60 on IBMi to resolve the issue.
PR-URL: https://github.com/libuv/libuv/pull/2591
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
