Re: hppa and sigaction

2005-05-31 Thread Martin Zobel-Helas
Hi Wesley, On Monday, 16 May 2005, you wrote: > On Sun, May 15, 2005 at 04:19:41PM -0400, Carlos O'Donell wrote: > > > ... or else another hppa machine which doesn't have this problem that is > > > open to debian developers? > > > > Yes, ask for a machine on [parisc-linux] lists.parisc-linux.org

Re: hppa and sigaction

2005-05-16 Thread Wesley W. Terpstra
On Sun, May 15, 2005 at 04:19:41PM -0400, Carlos O'Donell wrote: > > ... or else another hppa machine which doesn't have this problem that is > > open to debian developers? > > Yes, ask for a machine on [parisc-linux] lists.parisc-linux.org :) > There are a number available to testdrive with newer

Re: hppa and sigaction

2005-05-15 Thread Matt Taggart
"Wesley W. Terpstra" writes... > > > Is this "too old"? > > > > Yes. > > You need > 2.6.2 for this bug to be fixed on 64-bit kernels. > > Do you need to run a 64-bit smp kernel? > > I see. Thanks, that really helps; I was totally confused. =) > I don't have control over paer.debian.org. > > I

Re: hppa and sigaction

2005-05-15 Thread Carlos O'Donell
On Sun, May 15, 2005 at 12:33:04PM +0200, Wesley W. Terpstra wrote: > On Sun, May 15, 2005 at 02:44:13AM -0400, Carlos O'Donell wrote: > > > > If you are using an older 64-bit kernel all this information may be > > > > wrong, and require a kernel update. > > > > > > Linux paer 2.4.26-64-smp #1 SMP

Re: hppa and sigaction

2005-05-15 Thread Wesley W. Terpstra
On Sun, May 15, 2005 at 02:44:13AM -0400, Carlos O'Donell wrote: > > > If you are using an older 64-bit kernel all this information may be > > > wrong, and require a kernel update. > > > > Linux paer 2.4.26-64-smp #1 SMP Tue Aug 24 21:59:09 CEST 2004 parisc64 > > unknown > > > > Is this "too old

Re: hppa and sigaction

2005-05-15 Thread Carlos O'Donell
On Sat, May 14, 2005 at 11:48:46PM +0200, Wesley W. Terpstra wrote: > On Sat, May 14, 2005 at 02:46:23PM -0400, Carlos O'Donell wrote: > > The void* is ucontext_t. > > To get at sc_ioaq you must look in ucontext_t->uc_mcontext->sc_iaoq. > > This is what I've been doing, but it's always zero... >

Re: hppa and sigaction

2005-05-14 Thread Wesley W. Terpstra
On Sat, May 14, 2005 at 02:46:23PM -0400, Carlos O'Donell wrote: > The void* is ucontext_t. > To get at sc_ioaq you must look in ucontext_t->uc_mcontext->sc_iaoq. This is what I've been doing, but it's always zero... static void catcher (int sig, siginfo_t* sip, void* mystery) { ucontext_

Re: hppa and sigaction

2005-05-14 Thread Carlos O'Donell
On Sat, May 14, 2005 at 01:00:09PM +0200, Wesley W. Terpstra wrote: > When one sets an sa_sigaction with sigaction on parisc/hppa what is void*? > > I know it is sometimes ucontext_t or struct sigcontext depending on the > arch. However, I am trying to find the program counter prior to the signal

hppa and sigaction

2005-05-14 Thread Wesley W. Terpstra
When one sets an sa_sigaction with sigaction on parisc/hppa what is void*? I know it is sometimes ucontext_t or struct sigcontext depending on the arch. However, I am trying to find the program counter prior to the signal (for profiling) and with both structs, sc_ioaq is a constant. Am I going ab