Philip Mucci wrote:
Hi Stephane,
Did you happen to get this message?
I'm still having this issue with the latest CVS head. Is anyone else
seeing this on FC5/i386 (p6)?
Also, I have a suggestion for PFM debugging. Is it possible that debug=1
could not enable debugging in the context switch routines and we have a
separate debug option for that like overflow?
Seems like that would make a big difference in the run time of these
tests under debugging because my syslog is a busy puppy...
Regards,
Phil
Hi Phil,
A couple suggestions to help track down this problem. Could you compile
the code with "-save-temps" to generate the .s file from gcc? Run the
code in gdb to get a traceback of where exactly it is crashing?
-Will
On Mon, 2006-07-10 at 02:54 +0200, Philip Mucci wrote:
Ok, it appears to be a compiler bug...check this out:
[EMAIL PROTECTED] examples]$ gcc -g self.c ../lib/libpfm.a -I../include
detect_pmcs.c
[EMAIL PROTECTED] examples]$ ./a.out
pfmlib_i386_p6.c (pfm_i386_p6_dispatch_counters.211):
ev[0]=CPU_CLK_UNHALTED
pfmlib_i386_p6.c (pfm_i386_p6_dispatch_counters.211): ev[1]=INST_RETIRED
[perfsel0=0x510079 emask=0x79 umask=0x0 os=0 usr=1 en=1 int=1 inv=0
edge=0 cnt_mask=0] CPU_CLK_UNHALTED
[perfsel1=0x5100c0 emask=0xc0 umask=0x0 os=0 usr=1 en=1 int=1 inv=0
edge=0 cnt_mask=0] INST_RETIRED
PMD0 7490824041 CPU_CLK_UNHALTED
PMD1 9000008761 INST_RETIRED
[EMAIL PROTECTED] examples]$ gcc -O self.c ../lib/libpfm.a -I../include
detect_pmcs.c
[EMAIL PROTECTED] examples]$ ./a.out
pfmlib_i386_p6.c (pfm_i386_p6_dispatch_counters.211):
ev[0]=CPU_CLK_UNHALTED
pfmlib_i386_p6.c (pfm_i386_p6_dispatch_counters.211): ev[1]=INST_RETIRED
[perfsel0=0x510079 emask=0x79 umask=0x0 os=0 usr=1 en=1 int=1 inv=0
edge=0 cnt_mask=0] CPU_CLK_UNHALTED
[perfsel1=0x5100c0 emask=0xc0 umask=0x0 os=0 usr=1 en=1 int=1 inv=0
edge=0 cnt_mask=0] INST_RETIRED
PMD0 425 CPU_CLK_UNHALTED
PMD1 52 INST_RETIRED
Segmentation fault
[EMAIL PROTECTED] examples]$ gcc -O3 self.c ../lib/libpfm.a
-I../include detect_pmcs.c
[EMAIL PROTECTED] examples]$ ./a.out
pfmlib_i386_p6.c (pfm_i386_p6_dispatch_counters.211):
ev[0]=CPU_CLK_UNHALTED
pfmlib_i386_p6.c (pfm_i386_p6_dispatch_counters.211): ev[1]=INST_RETIRED
[perfsel0=0x510079 emask=0x79 umask=0x0 os=0 usr=1 en=1 int=1 inv=0
edge=0 cnt_mask=0] CPU_CLK_UNHALTED
[perfsel1=0x5100c0 emask=0xc0 umask=0x0 os=0 usr=1 en=1 int=1 inv=0
edge=0 cnt_mask=0] INST_RETIRED
pfm_write_pmcs error errno 22
[EMAIL PROTECTED] examples]$ gcc -v
Using built-in specs.
Target: i386-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada
--enable-java-awt=gtk --disable-dssi
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre
--with-cpu=generic --host=i386-redhat-linux
Thread model: posix
gcc version 4.1.1 20060525 (Red Hat 4.1.1-1)
So detect_pmcs appears to clobber something in the other code...
_______________________________________________
perfmon mailing list
[email protected]
http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
_______________________________________________
perfmon mailing list
[email protected]
http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
