| Age | Commit message (Collapse) | Author |
|---|
|
Avoid calling select with empty sets which hangs the process
This makes uClibc-ng act like glibc and musl
Without this fix the test_poll of python3 testsuite hangs forever
Scenario of the issue:
If you call poll with only invalid file descriptors, like in python3
testsuite
(https://github.com/python/cpython/blob/master/Lib/test/test_poll.py#L83)
You will go through uClibc poll emulation code, which is based on
select syscall.
Your first call to select will fail, it will return -1 and errno will be
set to EBADF: https://github.com/wbx-github/uclibc-ng/blob/master/libc/sysdeps/linux/common/poll.c#L120
Then you will go through the for loop which tests individually each file descriptor by calling
select on each one: https://github.com/wbx-github/uclibc-ng/blob/master/libc/sysdeps/linux/common/poll.c#L163
each call will also return -1 with errno being equal to EBADF.
Therefore all pollfd will have the POLLNVAL flag in their respective revents field.
And, the most important, rset/wset/xset will stay empty.
Then the for loop ends, the "continue" makes the while loop run again.
The following select() is run again: https://github.com/wbx-github/uclibc-ng/blob/master/libc/sysdeps/linux/common/poll.c#L120
But this time the sets are empty.
If the poll was called with timeout set to -1, this select will hang forever because there is no timeout
and the sets are empty so no event will ever wake it up.
test program:
int main(void)
{
struct pollfd pfd;
int ret;
int pipe_fds[2];
pipe(pipe_fds);
close(pipe_fds[0]);
close(pipe_fds[1]);
pfd.fd = pipe_fds[0];
pfd.events = POLLIN | POLLOUT | POLLPRI;
pfd.revents = 0;
ret = poll(&pfd, 1, -1);
printf("ret: %d\n", ret);
if (ret < 0)
printf("error: %s", strerror(errno));
else {
puts("revents: ");
if (pfd.revents & POLLERR)
printf(" POLLERR");
if (pfd.revents & POLLHUP)
printf(" POLLHUP");
if (pfd.revents & POLLNVAL)
printf(" POLLNVAL");
puts("");
}
return 0;
}
This hangs on uClibc-ng aarch64 and Kalray's arch (kv3) but does the following on musl and glibc:
"
ret: 1
revents:
POLLNVAL
"
strace output of this program with uClibc *without* the patch applied:
pselect6(4, [3], [3], [3], NULL, NULL) = -1 EBADF (Bad file descriptor)
pselect6(4, [3], [3], [3], {tv_sec=0, tv_nsec=0}, NULL) = -1 EBADF (Bad file descriptor)
pselect6(0, 0x7ffffffb80, 0x7ffffffb68, 0x7ffffffb50, NULL, NULL
(never finishes)
strace output of this program with uClibc *with* the patch applied:
pselect6(4, [3], [3], [3], NULL, NULL) = -1 EBADF (Bad file descriptor)
pselect6(4, [3], [3], [3], {tv_sec=0, tv_nsec=0}, NULL) = -1 EBADF (Bad file descriptor)
write(1, "ret: 1\n", 7ret: 1
) = 7
write(1, "revents: \n", 10revents:
) = 10
write(1, " POLLNVAL\n", 10 POLLNVAL
) = 10
exit_group(0) = ?
+++ exited with 0 +++
|
|
This matches a similar change made to glibc.
No functional changes here.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
use __SYSCALL_SIGSET_T_SIZE
use non-cancellable select in fallback
Signed-off-by: Peter S. Mazinger <ps.m@gmx.net>
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
|
|
While there, remove libc_hidden_protos
Signed-off-by: Peter S. Mazinger <ps.m@gmx.net>
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
|
|
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
Conflicts:
Makefile.in
extra/Configs/Config.in
libc/sysdeps/linux/common/bits/kernel-features.h
libc/sysdeps/linux/common/poll.c
libc/sysdeps/linux/common/sysdep.h
libc/sysdeps/linux/sh/sysdep.h
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
fixes bug #253
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
|
|
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
Now they are only enabled if linuxthreads.old are selected.
|
|
Appears to build fine (several .configs tried)
|
|
|
|
|
|
like
o UCLIBC_HAS_GNU_ERROR
o UCLIBC_HAS_BSD_ERR
o UCLIBC_HAS_PTY
o UCLIBC_HAS_GETPT (1)
o UCLIBC_SYSCALL_STUBS
o UCLIBC_SYSCALL_STUB_WARNING
o UCLIBC_LINUX_SPECIFIC (2)
o UCLIBC_BSD_SPECIFIC (3)
o UCLIBC_NTP_LEGACY (4)
o UCLIBC_SV4_DEPRECATED (5)
o UCLIBC_HAVE_REALTIME (6)
o UCLIBC_HAVE_ADVANCED_REALTIME (7)
o UCLIBC_HAVE_EPOLL (8)
o UCLIBC_HAVE_XATTR (9)
o UCLIBC_HAVE_PROFILING (10)
(1) make non-standard getpt optional and implement standard posix_openpt
(2) fstatfs(), inotify_*(), ioperm(), iopl(), madvise(), modify_ldt(),
personality()
ppoll(), setresuid()
(3) mincore(), getdomainname(), setdomainname()
(4) ntp_adjtime(), ntp_gettime() aliases
(5) ustat() [use statfs(2) in your code instead]
(6) All marked as "(REALTIME)" in SUSv3
(7) All marked as "(ADVANCED REALTIME)" in SUSv3
(8) epoll_create(), epoll_ctl(), epoll_wait()
(9) all Extended Attributes
(10) helpers for gcc's -finstrument-functions
- Fixes _dl_exit()
- Implements sleep(3) for !UCLIBC_HAVE_REALTIME
- Implements usleep(3) for !UCLIBC_HAVE_REALTIME
- adds #warning about incorrect posix_fadvise{,64}()
- removes unused and unwanted uselib()
Net outcome is that an allnoconfig with HAVE_SHARED is now about 88k instead
of formerly 130k.
|
|
in string.h and strings.h. This caught unguarded string ops in
libc/inet/ethers.c __ether_line_w() function.
I will wait for fallout reports for a week or so,
then continue converting more libc_hidden_proto's.
|
|
Jean-Christian de Rivaz writes:
The attached patch solve an issue I faced while using the libdbus-glib
waiting for a D-Bus message or the end of a glib timer at the same time.
This specific case of use generate a poll call with a zero timeout. On
platformes with the glibc a zero timeout poll return immetiately even if
there is no file descriptor event. But on platformes with uClibc a zero
timeout poll block until a file descriptor event occurs.
|
|
I actually suspect this code into the file uClibc/libc/sysdeps/linux/common/poll.c:
tval.tv_nsec = (timeout % 1000) *1000; <==== make only usec!
From milisecond this really needs a * 1000000 to make nanosecond. Without this
a 1100 milisecond timeout is converted into a 1 seconde and 100 microsecond
timeout! This can explain the weird result of the test code.
|
|
|
|
applications stop using _syscall#() and use syscall() instead. Cleanup
internal handling of syscall includes to use the correct header file.
|
|
|
|
|
|
missing headers, other jump relocs removed
|
|
|
|
|
|
|
|
|
|
is not a SuSv3 symbol). Rather than using __bzero internally per Alexandre's
original patch, use memset instead.
|
|
The #ifdef __NR_poll test was failing because it was done before any includes.
Hence, the emulation was always being used.
NOTE: The emulation fails a couple of tests in test_poll.py!
|
|
|
|
select for older 2.0 kernels where poll is missing.
|
