On Wed, Nov 29, 2006 at 05:04:38PM +0100, Philip J. Mucci wrote: > Hi Stefane, > > Well, I never ran these test cases before now.... > > It looks like when I run either rtop or syst, I'm seeing an error > because Perfmon2 seems to think that there's other contexts running when > in fact there isn't. Something isn't getting cleaned up, because I can > increase the number of 'conflicting sessions' simply by running 'self' a > number of times. It seems every run increases this value. > Are you running UP? Could you try compiling SMP and runing UP?
I'll see if I can reproduce in UP mode. > For example: > > sicily ~ # dmesg | grep conflict > perfmon: pfm_reserve_session.250: CPU0 [2549]: system wide imppossible, > 8 conflictingtask_sessions > sicily ~ # ~phil/libpfm/examples/self > [CP0_25_0(pmc0)=0x18 event_mask=0x0 usr=1 os=0 sup=0 exl=0 int=1] CYCLES > [CP0_25_0(pmd0)] > [CP0_25_2(pmc1)=0x118 event_mask=0x8 usr=1 os=0 sup=0 exl=0 int=1] > INSNS_COMPLETE > [CP0_25_2(pmd1)] > PMD0 21139069063 CYCLES > PMD1 15000092402 INSNS_COMPLETE > sicily ~ # ~phil/libpfm/examples/syst > pfm_load_context error errno 16 > sicily ~ # dmesg | grep conflict > perfmon: pfm_reserve_session.250: CPU0 [2563]: system wide imppossible, > 9 conflictingtask_sessions > > Any ideas? Let me know where it should be fixed because I'll fix it in > my code base also. > > Here is more complete log of earlier failures. > > Phil > > > Nov 29 11:34:03 localhost perfmon: pfm_alloc_fd.739: CPU0 [2501]: new > inode ino=3890 @98000000014d3f30 > Nov 29 11:34:03 localhost perfmon: pfm_find_set.499: CPU0 [2501]: > looking for set=0 > Nov 29 11:34:03 localhost perfmon: pfm_find_set.571: CPU0 [2501]: > set_id=0 size=2584 view=980000000fb3df70 remap=0 mmap_offs=0 > Nov 29 11:34:03 localhost perfmon: pfm_init_evtset.468: CPU0 [2501]: > set0 pmc0=0x10 > Nov 29 11:34:03 localhost perfmon: pfm_init_evtset.468: CPU0 [2501]: > set0 pmc1=0x10 > Nov 29 11:34:03 localhost perfmon: __pfm_create_context.1625: CPU0 > [2501]: ctx=980000000f50e000 flags=0x2 system=1 notify_block=0 no_msg=0 > use_fmt=0 remap=0 ctx_fd=3 mode=0 > Nov 29 11:34:03 localhost perfmon: pfm_check_task_state.192: CPU0 > [2501]: state=1 check_mask=0x0 > Nov 29 11:34:03 localhost perfmon: __pfm_getinfo_evtsets.676: CPU0 > [2501]: set0 flags=0x0 eff_usec=0 runs=0 > Nov 29 11:34:03 localhost perfmon: pfm_check_task_state.192: CPU0 > [2501]: state=1 check_mask=0x1 > Nov 29 11:34:03 localhost perfmon: pfm_find_set.499: CPU0 [2501]: > looking for set=0 > Nov 29 11:34:03 localhost perfmon: __pfm_write_pmcs.465: CPU0 [2501]: > set0 pmc0=0x19 a_pmu=0 u_pmcs=0x1 nu_pmcs=1 > Nov 29 11:34:03 localhost perfmon: __pfm_write_pmcs.465: CPU0 [2501]: > set0 pmc1=0x119 a_pmu=0 u_pmcs=0x3 nu_pmcs=2 > Nov 29 11:34:03 localhost perfmon: pfm_check_task_state.192: CPU0 > [2501]: state=1 check_mask=0x1 > Nov 29 11:34:03 localhost perfmon: pfm_find_set.499: CPU0 [2501]: > looking for set=0 > Nov 29 11:34:03 localhost perfmon: __pfm_write_pmds.299: CPU0 [2501]: > set0 pmd0=0x0 flags=0x0 a_pmu=0 hw_pmd=0x0 ctx_pmd=0x0 s_reset=0x0 > l_reset=0x0 u_pmds=0x1 nu_pmds=1 s_pmds=0x0 r_pmds=0x0 o_pmds=0x0 > o_thres=0 compat=0 eventid=0 > Nov 29 11:34:03 localhost perfmon: __pfm_write_pmds.299: CPU0 [2501]: > set0 pmd1=0x0 flags=0x0 a_pmu=0 hw_pmd=0x0 ctx_pmd=0x0 s_reset=0x0 > l_reset=0x0 u_pmds=0x3 nu_pmds=2 s_pmds=0x0 r_pmds=0x0 o_pmds=0x0 > o_thres=0 compat=0 eventid=0 > Nov 29 11:34:03 localhost perfmon: pfm_check_task_state.192: CPU0 > [2501]: state=1 check_mask=0x2 > Nov 29 11:34:03 localhost perfmon: pfm_find_set.499: CPU0 [2501]: > looking for set=0 > Nov 29 11:34:03 localhost perfmon: pfm_prepare_sets.112: CPU0 [2501]: > set0 sw_next=0 > Nov 29 11:34:03 localhost perfmon: __pfm_load_context.1245: CPU0 [2501]: > load_pid=0 set=0 set_flags=0x0 > Nov 29 11:34:03 localhost perfmon: pfm_reserve_session.241: CPU0 [2501]: > in sys_sessions=0 task_sessions=2 syswide=1 cpu=0 > Nov 29 11:34:03 localhost perfmon: pfm_reserve_session.250: CPU0 [2501]: > system wide imppossible, 2 conflictingtask_sessions > Nov 29 11:34:03 localhost > Nov 29 11:34:03 localhost perfmon: __pfm_close.531: CPU0 [2501]: state=1 > Nov 29 11:34:03 localhost perfmon: pfm_context_free.116: CPU0 [2501]: > free ctx @980000000f50e000 > -- -Stephane _______________________________________________ perfmon mailing list [email protected] http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
