Date: Sun, 11 Jul 2010 15:35:00 -0700 From: Garrett Cooper <yanegomi@gmail.com> To: Kostik Belousov <kostikbel@gmail.com> Cc: hackers@freebsd.org Subject: Re: *sigpause hanging on 8.x+ Message-ID: <AANLkTik79TZiA0CLubDgVpp8hY4M_C9LnQIIOKmMQ1In@mail.gmail.com> In-Reply-To: <20100711214016.GR2408@deviant.kiev.zoral.com.ua> References: <AANLkTikfxu3BjQE4nnNJ-VeNSDy9vhfXEhY1XQ4jWtY4@mail.gmail.com> <20100711210843.GQ2408@deviant.kiev.zoral.com.ua> <AANLkTimjk7oJKlB7YCrZimMsTQtvKIxipZtfBfpIdSz3@mail.gmail.com> <20100711214016.GR2408@deviant.kiev.zoral.com.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
2010/7/11 Kostik Belousov <kostikbel@gmail.com>:
> On Sun, Jul 11, 2010 at 02:30:01PM -0700, Garrett Cooper wrote:
>> On Sun, Jul 11, 2010 at 2:08 PM, Kostik Belousov <kostikbel@gmail.com> w=
rote:
>> > On Sun, Jul 11, 2010 at 12:39:39PM -0700, Garrett Cooper wrote:
>> >> So, long story short... I've basically ported the open posix testsuit=
e
>> >> to FreeBSD, and one of the tests tests out sigpause. Unfortunately th=
e
>> >> sucker hangs on my dev box at home.
>> >>
>> >> I've written a short testcase that demonstrates this. It prints out:
>> >>
>> >> $ ~/test_sigpause
>> >> 0
>> >>
>> >> And proceeds to be unresponsive to signals (except SIGSTOP / SIGKILL,
>> >> as expected).
>> >>
>> >> When I monkey around with libc's compat4.3 stuff a bit, this is what =
comes up:
>> >>
>> >> $ env LD_LIBRARY_PATH=3D$PWD:/usr/src/lib/libc/../libthr ~/test_sigpa=
use
>> >> 0
>> >> before sigemptyset
>> >> before _sigsuspend
>> >>
>> >> So it's getting stuck after calling _sigsuspend.
>> >>
>> >> I tried the same thing on a i386 8-STABLE VM and it hangs as well.
>> >>
>> >> I tried applying similar printfs in libthr but it's not hitting that
>> >> code at all (it's now responding to SIGTERM though, which is
>> >> interesting, but not too interesting to me).
>> >>
>> >> I also wrote similar code that exercised the functionality in
>> >> sigsuspend, by calling sigprocmask beforehand, and it works.
>> >>
>> >> Thoughts?
>> >>
>> >> -Garrett
>> >>
>> >> Dev machine:
>> >> FreeBSD bayonetta.local 9.0-CURRENT FreeBSD 9.0-CURRENT #1
>> >> r206173:209901M: Sun Jul 11 04:18:42 PDT 2010
>> >> root@:/usr/obj/usr/src/sys/BAYONETTA =A0amd64
>> >> VM:
>> >> FreeBSD starr-bastion.localdomain 8.0-STABLE FreeBSD 8.0-STABLE #0
>> >> r207913: Tue May 11 06:21:57 UTC 2010
>> >> root@starr-bastion.localdomain:/usr/obj/usr/src/sys/GENERIC =A0i386
>> >>
>> >> Index: compat-43/sigcompat.c
>> >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>> >> --- compat-43/sigcompat.c =A0 =A0 (revision 206173)
>> >> +++ compat-43/sigcompat.c =A0 =A0 (working copy)
>> >> @@ -36,6 +36,7 @@
>> >> =A0#include "namespace.h"
>> >> =A0#include <sys/param.h>
>> >> =A0#include <signal.h>
>> >> +#include <stdio.h>
>> >> =A0#include <string.h>
>> >> =A0#include "un-namespace.h"
>> >> =A0#include "libc_private.h"
>> >> @@ -102,7 +103,9 @@
>> >> =A0{
>> >> =A0 =A0 =A0 sigset_t set;
>> >>
>> >> + =A0 =A0 printf("before sigemptyset\n");
>> >> =A0 =A0 =A0 sigemptyset(&set);
>> >> + =A0 =A0 printf("before _sigsuspend\n");
>> >> =A0 =A0 =A0 set.__bits[0] =3D mask;
>> >> =A0 =A0 =A0 return (_sigsuspend(&set));
>> >> =A0}
>> >> @@ -111,10 +114,16 @@
>> >> =A0xsi_sigpause(int sig)
>> >> =A0{
>> >> =A0 =A0 =A0 sigset_t set;
>> >> + =A0 =A0 int rc;
>> >>
>> >> + =A0 =A0 printf("before sigemptyset\n");
>> >> =A0 =A0 =A0 sigemptyset(&set);
>> >> + =A0 =A0 printf("before sigaddset\n");
>> >> =A0 =A0 =A0 sigaddset(&set, sig);
>> >> - =A0 =A0 return (_sigsuspend(&set));
>> >> + =A0 =A0 printf("before _sigsuspend\n");
>> >> + =A0 =A0 rc =3D (_sigsuspend(&set));
>> >> + =A0 =A0 printf("after _sigsuspend\n");
>> >> + =A0 =A0 return rc;
>> >> =A0}
>> >>
>> >> =A0int
>> >>
>> >> $ cat ~/test_sigpause.c
>> >> #include <signal.h>
>> >> #include <stdio.h>
>> >>
>> >> int
>> >> main (void)
>> >> {
>> >> =A0 =A0 =A0 =A0 printf("0\n");
>> >> =A0 =A0 =A0 =A0 fflush(stdout);
>> >> =A0 =A0 =A0 =A0 (void) sigpause(1);
>> >> =A0 =A0 =A0 =A0 return 0;
>> >> }
>> >> $ cat ~/test_sigsuspend.c
>> >> #include <err.h>
>> >> #include <signal.h>
>> >>
>> >> int
>> >> main (void)
>> >> {
>> >> =A0 =A0 =A0 =A0 sigset_t oset;
>> >> =A0 =A0 =A0 =A0 sigset_t nset;
>> >> =A0 =A0 =A0 =A0 if (sigprocmask(1, &nset, &oset) =3D=3D -1)
>> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 err(1, "sigprocmask(-1, &nset, &oset)=
");
>> >> =A0 =A0 =A0 =A0 if (sigprocmask(-1, &nset, &oset) =3D=3D -1)
>> >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 err(1, "sigprocmask(-1, &nset, &oset)=
");
>> >> =A0 =A0 =A0 =A0 return (sigsuspend(&nset));
>> >> }
>> >
>> > It seems I got a sigmask for sigpause inside the xsi_sigpause() backwa=
rd.
>> > On the other hand, I do not understand what is your issue with sigpaus=
e().
>>
>> The negative testcase from the open posix testsuite was setup so that
>> setting sigpause(-1) would return -1 with EINVAL, according to the
>> sig* manpages (-1 is an invalid signal of course). That isn't being
>> triggered with either function today.
>>
>> 0 seems a bit wonky too (it's an invalid signal number).
>>
>> My bet is that values greater than SIGRTMAX aren't interpreted properly =
either.
>
> I will add these checks, thanks.
Much obliged :)... FWIW sigprocmask fails to do the right thing in
detecting the signal number:
$ ~/test_sigprocmask
signo =3D -1 result not sane (0 !=3D -1, errno: 0 !=3D EINVAL)
signo =3D 0 result not sane (0 !=3D -1, errno: 0 !=3D EINVAL)
signo =3D 1 result sane
signo =3D 9 result sane
signo =3D 17 result sane
signo =3D 65 result sane
signo =3D 64 result sane
signo =3D 66 result not sane (0 !=3D -1, errno: 0 !=3D EINVAL)
Would this fix that?
Index: sys/kern/kern_sig.c
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
--- sys/kern/kern_sig.c (revision 206173)
+++ sys/kern/kern_sig.c (working copy)
@@ -988,6 +988,9 @@
struct proc *p;
int error;
+ if (!_SIG_VALID(how))
+ return (-EINVAL);
+
p =3D td->td_proc;
if (!(flags & SIGPROCMASK_PROC_LOCKED))
PROC_LOCK(p);
I'll look for more low-hanging fruit.
>> > diff --git a/lib/libc/compat-43/sigcompat.c b/lib/libc/compat-43/sigco=
mpat.c
>> > index c3ba30a..bab9d5c 100644
>> > --- a/lib/libc/compat-43/sigcompat.c
>> > +++ b/lib/libc/compat-43/sigcompat.c
>> > @@ -111,9 +111,12 @@ int
>> > =A0xsi_sigpause(int sig)
>> > =A0{
>> > =A0 =A0 =A0 =A0sigset_t set;
>> > + =A0 =A0 =A0 int error;
>> >
>> > - =A0 =A0 =A0 sigemptyset(&set);
>> > - =A0 =A0 =A0 sigaddset(&set, sig);
>> > + =A0 =A0 =A0 error =3D _sigprocmask(SIG_BLOCK, NULL, &set);
>> > + =A0 =A0 =A0 if (error !=3D 0)
>> > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return (error);
>> > + =A0 =A0 =A0 sigdelset(&set, sig);
>> > =A0 =A0 =A0 =A0return (_sigsuspend(&set));
>> > =A0}
>>
>> Doesn't this violate the restore clause noted in the manpage?
>>
>> =A0 =A0 =A0The xsi_sigpause() function removes sig from the signal mask =
of the call-
>> =A0 =A0 =A0ing process and suspend the calling process until a signal is=
received.
>> =A0 =A0 =A0The xsi_sigpause() function restores the signal mask of the p=
rocess to
>> =A0 =A0 =A0its original state before returning.
>>
>> So if I had a sigset defined above with sig, then redefined it, I
>> would be whacking the previous handler by passing in NULL to
>> _sigprocmask, correct? If so, sigpause has issues too in its
>> implementation.
> No, not correct. Read the description of sigsuspend.
Yeah, I was wrong here:
The sigsuspend() system call temporarily changes the blocked signal ma=
sk
to the set to which sigmask points, and then waits for a signal to
arrive; on return the previous set of masked signals is restored. The
signal mask set is usually empty to indicate that all signals are to b=
e
unblocked for the duration of the call.
> Also note that the sigprocmask call does not change process mask.
Not so sure about this though:
The sigprocmask() system call examines and/or changes the current sign=
al
mask (those signals that are blocked from delivery). Signals are bloc=
ked
if they are members of the current signal mask set.
>> There's also some interesting SIGDELSET action going on in libthr's
>> copy of _sigsuspend's with SIGCANCEL (apparently that's the unofficial
>> alias for SIGRTMIN as defined by libthr), but that's a sidenote for
>> the actual issue seen here.
Here's the test app I wrote and executed above, just for future referen=
ce:
Thanks!
-Garrett
$ cat ~/test_sigprocmask.c
#include <errno.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#define TEST_SIGPROCMASK_POS(signo) do { \
printf("signo =3D %d ", signo); \
rc =3D sigprocmask(-1, NULL, &oset); \
if (rc !=3D 0) { \
printf("result not sane (%d !=3D 0, errno: %d)\n", \
rc, errno); \
} else \
printf("result sane\n"); \
} while (0)
#define TEST_SIGPROCMASK_NEG(signo) do { \
printf("signo =3D %d ", signo); \
rc =3D sigprocmask(-1, NULL, &oset); \
if (rc !=3D -1 || errno !=3D EINVAL) { \
printf("result not sane (%d !=3D -1, " \
"errno: %d !=3D EINVAL)\n", \
rc, errno); \
} else \
printf("result sane\n"); \
} while (0)
int
main(void)
{
sigset_t oset;
int rc;
TEST_SIGPROCMASK_NEG(-1);
TEST_SIGPROCMASK_NEG(0);
TEST_SIGPROCMASK_POS(SIGHUP);
/* The system quietly disallows SIGKILL or SIGSTOP to be blocked. */
TEST_SIGPROCMASK_POS(SIGKILL);
TEST_SIGPROCMASK_POS(SIGSTOP);
TEST_SIGPROCMASK_POS(SIGRTMIN);
TEST_SIGPROCMASK_POS(SIGRTMIN-1);
TEST_SIGPROCMASK_NEG(SIGRTMIN+1);
return (0);
}
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTik79TZiA0CLubDgVpp8hY4M_C9LnQIIOKmMQ1In>