aboutsummaryrefslogtreecommitdiff
path: root/tools/perf/util/trace-event-scripting.c
diff options
context:
space:
mode:
authorKonstantin Khlebnikov <[email protected]>2012-06-07 14:21:14 -0700
committerLinus Torvalds <[email protected]>2012-06-07 14:43:55 -0700
commit40af1bbdca47e5c8a2044039bb78ca8fd8b20f94 (patch)
tree70b5a21579f8b0137c09e54d523246d84be93430 /tools/perf/util/trace-event-scripting.c
parent39caa0916ef27cf1da5026eb708a2b8413156f75 (diff)
mm: correctly synchronize rss-counters at exit/exec
mm->rss_stat counters have per-task delta: task->rss_stat. Before changing task->mm pointer the kernel must flush this delta with sync_mm_rss(). do_exit() already calls sync_mm_rss() to flush the rss-counters before committing the rss statistics into task->signal->maxrss, taskstats, audit and other stuff. Unfortunately the kernel does this before calling mm_release(), which can call put_user() for processing task->clear_child_tid. So at this point we can trigger page-faults and task->rss_stat becomes non-zero again. As a result mm->rss_stat becomes inconsistent and check_mm() will print something like this: | BUG: Bad rss-counter state mm:ffff88020813c380 idx:1 val:-1 | BUG: Bad rss-counter state mm:ffff88020813c380 idx:2 val:1 This patch moves sync_mm_rss() into mm_release(), and moves mm_release() out of do_exit() and calls it earlier. After mm_release() there should be no pagefaults. [[email protected]: tweak comment] Signed-off-by: Konstantin Khlebnikov <[email protected]> Reported-by: Markus Trippelsdorf <[email protected]> Cc: Hugh Dickins <[email protected]> Cc: KAMEZAWA Hiroyuki <[email protected]> Cc: Oleg Nesterov <[email protected]> Cc: <[email protected]> [3.4.x] Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
Diffstat (limited to 'tools/perf/util/trace-event-scripting.c')
0 files changed, 0 insertions, 0 deletions