mirror of
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
synced 2025-08-05 16:54:27 +00:00
de3629baf5
Camm noticed that on parisc a SIGFPE exception will crash an application with
a second SIGFPE in the signal handler. Dave analyzed it, and it happens
because glibc uses a double-word floating-point store to atomically update
function descriptors. As a result of lazy binding, we hit a floating-point
store in fpe_func almost immediately.
When the T bit is set, an assist exception trap occurs when when the
co-processor encounters *any* floating-point instruction except for a double
store of register %fr0. The latter cancels all pending traps. Let's fix this
by clearing the Trap (T) bit in the FP status register before returning to the
signal handler in userspace.
The issue can be reproduced with this test program:
root@parisc:~# cat fpe.c
static void fpe_func(int sig, siginfo_t *i, void *v) {
sigset_t set;
sigemptyset(&set);
sigaddset(&set, SIGFPE);
sigprocmask(SIG_UNBLOCK, &set, NULL);
printf("GOT signal %d with si_code %ld\n", sig, i->si_code);
}
int main() {
struct sigaction action = {
.sa_sigaction = fpe_func,
.sa_flags = SA_RESTART|SA_SIGINFO };
sigaction(SIGFPE, &action, 0);
feenableexcept(FE_OVERFLOW);
return printf("%lf\n",1.7976931348623158E308*1.7976931348623158E308);
}
root@parisc:~# gcc fpe.c -lm
root@parisc:~# ./a.out
Floating point exception
root@parisc:~# strace -f ./a.out
execve("./a.out", ["./a.out"], 0xf9ac7034 /* 20 vars */) = 0
getrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=RLIM_INFINITY}) = 0
...
rt_sigaction(SIGFPE, {sa_handler=0x1110a, sa_mask=[], sa_flags=SA_RESTART|SA_SIGINFO}, NULL, 8) = 0
--- SIGFPE {si_signo=SIGFPE, si_code=FPE_FLTOVF, si_addr=0x1078f} ---
--- SIGFPE {si_signo=SIGFPE, si_code=FPE_FLTOVF, si_addr=0xf8f21237} ---
+++ killed by SIGFPE +++
Floating point exception
Signed-off-by: Helge Deller <deller@gmx.de>
Suggested-by: John David Anglin <dave.anglin@bell.net>
Reported-by: Camm Maguire <camm@maguirefamily.org>
Cc: stable@vger.kernel.org
|
||
|---|---|---|
| .. | ||
| cnv_float.h | ||
| dbl_float.h | ||
| decode_exc.c | ||
| denormal.c | ||
| dfadd.c | ||
| dfcmp.c | ||
| dfdiv.c | ||
| dfmpy.c | ||
| dfrem.c | ||
| dfsqrt.c | ||
| dfsub.c | ||
| driver.c | ||
| fcnvff.c | ||
| fcnvfu.c | ||
| fcnvfut.c | ||
| fcnvfx.c | ||
| fcnvfxt.c | ||
| fcnvuf.c | ||
| fcnvxf.c | ||
| float.h | ||
| fmpyfadd.c | ||
| fpbits.h | ||
| fpu.h | ||
| fpudispatch.c | ||
| frnd.c | ||
| hppa.h | ||
| Makefile | ||
| math-emu.h | ||
| README | ||
| sfadd.c | ||
| sfcmp.c | ||
| sfdiv.c | ||
| sfmpy.c | ||
| sfrem.c | ||
| sfsqrt.c | ||
| sfsub.c | ||
| sgl_float.h | ||
All files except driver.c are snapshots from the HP-UX kernel. They've been modified as little as possible. Even though they don't fit the Linux coding style, please leave them in their funny format just in case someone in the future, with access to HP-UX source code, is generous enough to update our copies with later changes from HP-UX -- it'll make their 'diff' job easier if our code is relatively unmodified. Required Disclaimer: Hewlett-Packard makes no implied or expressed warranties about this code nor any promises to maintain or test it in any way. This copy of this snapshot is no longer the property of Hewlett-Packard.