Hi, we're using the MontaVista kernel (2.6.18), and I've been noticing that
our mmcqd process uses a lot of CPU time when we're writing:
Mem: 63372K used, 4360K free, 0K shrd, 208K buff, 40612K cached
Load average: 3.00 2.29 2.04 (Status: S=sleeping R=running, W=waiting)
PID USER STATUS RSS PPID %CPU %MEM COMMAND
214 root DW< 0 14 28.6 0.0 mmcqd
585 root S 3372 1 28.1 4.9 ............
530 root S 17028 1 28.0 25.0 .............
1243 root R 892 536 1.3 1.3 top
1209 root SW 0 14 0.1 0.0 pdflush
1201 root SW 0 14 0.1 0.0 pdflush
545 root S 2212 1 0.0 3.2 netconfigd
536 root S 940 1 0.0 1.3 sh
640 root S < 844 1 0.0 1.2 wdt.out
1 root S 584 0 0.0 0.8 init
1245 root S 580 530 0.0 0.8 sh
1246 root R 528 1245 0.0 0.7 mv
282 root S < 420 1 0.0 0.6 udevd
572 root S 228 1 0.0 0.3 udhcpc
76 root SW< 0 14 0.0 0.0 kswapd0
217 root SW< 0 14 0.0 0.0 rpciod/0
13 root SW< 0 1 0.0 0.0 khelper
11 root SW< 0 1 0.0 0.0 desched/0
12 root SW< 0 1 0.0 0.0 events/0
14 root SW< 0 1 0.0 0.0 kthread
2 root SW 0 1 0.0 0.0 posix_cpu_timer
My question is: has there been any change in the mainline kernel that might
address this sort of write performance? Or does anyone know why mmcqd
consumes so much CPU time in the MV kernel?
Any help is much appreciated,
Jeremy
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
