| 1 |
On Sat, 29 Jul 2006 17:37:03 -0400 |
| 2 |
"Richard Hansen" <gentoo-a7x@×××××××××××.org> wrote: |
| 3 |
|
| 4 |
> After spending hours with gdb, I believe I have narrowed down the |
| 5 |
> problem to a setjmp() call in the Wireshark code. That call seems to |
| 6 |
> be trashing the stack, based on comparing local variable values |
| 7 |
> before and after the call. Is this a known limitation of SSP, or |
| 8 |
> should setjmp() work with SSP? |
| 9 |
|
| 10 |
Well, the setjmp() connection is new, but we do know that SSP in |
| 11 |
gcc-3.x sometimes generates bad code, and that sometimes gcc-3.x |
| 12 |
generates bad code regardless of SSP (where enabling SSP just causes |
| 13 |
the error to be detected -see bug #133301 for an example). However we |
| 14 |
haven't yet worked out what's going on. It would be worth building |
| 15 |
Wireshark with SSP off and running gdb in the same way you did before, |
| 16 |
breaking at the same place, to see if the same local variables are |
| 17 |
corrupted or not when SSP is disabled. |
| 18 |
|
| 19 |
-- |
| 20 |
Kevin F. Quinn |
Archives