I have not looked at the code for this in a long time, so I'm not sure why there's a 6k page count difference (roughly 6% difference in the kernel size as reported). I'll take a look at the code and try and sort out this discrepancy.
FWIW, I've always viewed using pp_kernel as a gauge for the amount of memory the kernel is using, not a precise metric. Watch this space... /jim Allan wrote: > Folks, > > I have the below ouput and I am trying to reconcile the ::memstat output back > to the kstats, within the Solaris Perf and Tools book it states that the > pp_kernel value is the amount of memory used if I run the following > > >> ::memstat >> > Page Summary Pages MB %Tot > ------------ ---------------- ---------------- ---- > Kernel 97685 763 38% > Anon 123733 966 48% > Exec and libs 5498 42 2% > Page cache 11476 89 4% > Free (cachelist) 5807 45 2% > Free (freelist) 15684 122 6% > > Total 259883 2030 > Physical 255439 1995 > > [EMAIL PROTECTED] kstat |grep pp_kernel > pp_kernel 103435 > [EMAIL PROTECTED] bc -l > [EMAIL PROTECTED] pagesize > 8192 > [EMAIL PROTECTED] bc -l > 103435*8192 > 847339520 > 847339520/1024 > 827480.00000000000000000000 > 827480.00000000000000000000/1024 > 808.08593750000000000000 > 808-763 > 45 > > > We can see the kernel is at 730mb is what I am seeing with the > pagesize*pp_kernel the kernel + free( cachelist ) or is this the kernel+ Exec > & Libs . > > I have seen issues with ::memstat not returing as the server is busy and I > was looking to use the kstats instead, so any help would be great. > > Thanks > > > This message posted from opensolaris.org > _______________________________________________ > perf-discuss mailing list > perf-discuss@opensolaris.org > _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
