From: Andi Kleen (ak_at_suse.de)
Date: Thu Jan 03 2002 - 19:38:29 GMT
On Thu, Jan 03, 2002 at 10:26:13AM -0800, Bob Carragher wrote: > Well, under Solaris, I would guess that if a signal is being > generated to indicate thread mutex synchronisation, it is > being ignored by Ups. Certainly, I can set breakpoints in Solaris likely uses some other mechanism for its mutexes. > certain functions which can only be reached by "child" threads, > and have no problems with Ups not reaching them. Perhaps that > would be a sufficient "fix" for Ups for Linux -- i.e. either > have it ignore signal 32 or at least have the capability of > setting the action on such a signal? Looking at obj_signal.c you likely just need to extend the existing signal table a bit for this. -Andi (still hoping the ups will someday be GPLed)
This archive was generated by hypermail 2.1.4 : Wed Feb 13 2002 - 21:51:34 GMT