Hi,
Could you confirm this bug is still appeared in glibc 2.3.2.ds1-13?
At Mon, 22 Mar 2004 22:58:50 +0900,
Dmitry Maksyoma wrote:
> Same here. ls -l, df, date, tar all segfault. Statically linked tar
> works ok, but not date.
>
> monster:/lib# strace date
> execve("/bin/date", ["date"], [/* 41 vars */]) = 0
> uname({sys="Linux", node="monster.amur.ru", ...}) = 0
> brk(0) = 0x80bbd88
> brk(0x80dcd88) = 0x80dcd88
> brk(0x80dd000) = 0x80dd000
> ipc_subcall(0, 0xbfffe728, 0xbfffe728, 0x16 <unfinished ...>
> +++ killed by SIGSEGV +++
>
> monster:/lib# /home/den/LIBC/strace date
> execve("/bin/date", ["date"], [/* 41 vars */]) = 0
> uname({sys="Linux", node="monster.amur.ru", ...}) = 0
> brk(0) = 0x80bbd88
> brk(0x80dcd88) = 0x80dcd88
> brk(0x80dd000) = 0x80dd000
> clock_gettime(0, <unfinished ...>
> +++ killed by SIGSEGV +++
>
> This is the third box that has been crippled. We've had to reinstall the
> system on the first two to get rid of this problem.
Interestingly, the first strace calls ipc_subcall, but the second
calls clock_gettime. Weird behavior. I have no idea to resolve it.
I think it's not glibc problem; instead, for example hardware, or
kernel. Please recheck it. If you have no objection, I'll close this
bug.
Regards,
-- gotom
