Hi Dave,
It was noticed that when SPLIT_RSS_ACCOUNTING is enabled in kernel, the RSS
values shown by ps command is huge (negative value represented as ulong)
for certain tasks. Shown below is a ps output of such kind.
1416 1 3 db506ac0 IN 36235.9 936 4194128 debuggerd
1417 1 3 db500040 IN 0.0 16512 36 rild
1418 1 1 db500580 IN 0.1 47160 488 surfaceflinger
1419 1 0 db74d040 IN 1.1 458544 5148 zygote
* 1420 1 1 db5d3ac0 IN 36235.9 13868 4194280 drmserver*
1421 1 1 db5d3580 IN 0.0 54944 20 mediaserver
1422 1 0 db7c9580 IN 0.0 900 20 installd
* 1423 1 0 db7c9040 IN 36235.8 1828 4194112 keystore
1424 1 3 db5d8580 IN 36235.9 1152 4194276 akmdfs
1426 1 3 db7bd040 IN 36235.9 10880 4194212 glgps
1428 1 0 db61c040 IN 36235.9 2084 4194152 usb_portd
1429 1 0 db753ac0 IN 36235.9 868 4194148 atxd_proxy
1431 1 1 da6fbac0 UN 36235.9 6100 4194280 bkmgrd*
*
*
When SPLIT_RSS_ACCOUNTING is enabled, the rss values are stored in
task_struct of each thread, and is synced to the mm_struct.rss_stat only at
certain events. So during a crash it is likely that mm.rss_stat contains
old values which can even be negative.
I have prepared a patch (attached) to sync the task_struct.rss_stat with
mm.rss_stat, in get_task_mem_usage, when SPLIT_RSS_ACCOUNTING is enabled.
Please share your thoughts on this.
Thanks,
Vinayak