On 02/18/2015 03:57 PM, Eugene Rudoy wrote:
> hmm, I'm still having the same behavior as without the patch - a lot
> of unexpected breakpoints at exactly the same places. S. attached
> logs.
Might have something to do with difference kernel versions. On my test machine, 
the signal event is immediate followed by an rt_sigreturn, where as the logs 
you posted show a sigreturn. rt_sigreturn and sigreturn have different syscall 
numbers; perhaps we need to check for both.

Just to be clear, this change is not related to tracing PLT-less binaries. The 
following patch is against master.

2015-02-19
* sysdeps/linux-gnu/mips/trace.c(syscall_p):
- If top of callstack was sigreturn or rt_sigreturn, skip check for preceeding 
syscall instruction.
- Return the appropriate syscall number in *sysnum

Regards,
Faraz Shahbazker

_______________________________________________
Ltrace-devel mailing list
[email protected]
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/ltrace-devel

Reply via email to