On Tue, 10 Aug 2010 20:00:23 +0300
Török Edwin <[email protected]> wrote:
> Disassembling linux_nat_core_of_thread_1 at the crash site I see this:
>
> 0x0000000000484d56 <+230>: callq 0x45c1d0 <make_cleanup>
> 0x0000000000484d5b <+235>: mov $0x28,%esi
> 0x0000000000484d60 <+240>: mov %rbp,%rdi
> 0x0000000000484d63 <+243>: callq 0x44dcc0 <str...@plt>
> 0x0000000000484d68 <+248>: mov $0x29,%esi
> 0x0000000000484d6d <+253>: mov %rax,%rdi
> 0x0000000000484d70 <+256>: callq 0x44dcc0 <str...@plt>
> => 0x0000000000484d75 <+261>: lea 0x2(%rax),%rdx
> 0x0000000000484d79 <+265>: movzbl 0x2(%rax),%eax
> 0x0000000000484d7d <+269>: cmp $0x20,%al
> 0x0000000000484d7f <+271>: jne 0x484d93
> <linux_nat_core_of_thread_1+291> 0x0000000000484d81 <+273>: nopl
> 0x0(%rax)
>
> And source has this:
> p = strchr (content, '(');
> p = strchr (p, ')') + 2; /* skip ")" and a whitespace. */
FWIW latest GDB CVS added proper NULL checks here:
http://sourceware.org/cgi-bin/cvsweb.cgi/src/gdb/linux-nat.c.diff?r1=1.169&r2=1.170&cvsroot=src&f=h
Thanks to tromey on #gdb for the hint.
I'm leaving this bug open, if you want to backport the patch.
Best regards,
--Edwin
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
