Messages in this thread | | | Date | Fri, 8 May 2009 00:14:44 -0400 (EDT) | From | Steven Rostedt <> | Subject | regression in TSC unstable? |
| |
The ring-buffer-benchmark module creates a producer and consumer and loops on do_gettimeofday until it hits about 10 seconds. Then it calculates the number of events recorded / time running.
After Ingo merged Linus's latest my tests when from 330ns per entry to 880ns. After bisecting it I came down to this change:
commit a71e4917dc0ebbcb5a0ecb7ca3486643c1c9a6e2 Author: Len Brown <len.brown@intel.com> Date: Tue Apr 21 00:50:11 2009 -0400
ACPI: idle: mark_tsc_unstable() at init-time, not run-time
I've never had a problem with the TSC with this box before. It passes the synchronization phase, and until this commit, the TSC never was marked as unstable.
Should it now be unstable? Or is this a false positive?
# cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 4 model name : Intel(R) Xeon(TM) CPU 2.80GHz stepping : 1 cpu MHz : 2793.272 cache size : 1024 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 1 apicid : 0 initial apicid : 0 fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc pebs bts pni dtes64 monitor ds_cpl cid cx16 xtpr bogomips : 5586.54 clflush size : 64 cache_alignment : 128 address sizes : 36 bits physical, 48 bits virtual power management:
Anything else I might need to know?
-- Steve
| |