lkml.org 
[lkml]   [2011]   [Sep]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH] kmemleak: Show where early_log issues come from
From
Date
On Wed, 2011-09-28 at 14:07 +0100, Catalin Marinas wrote:
> Based on initial patch by Steven Rostedt.
>
> Early kmemleak warnings did not show where the actual kmemleak API had
> been called from but rather just a backtrace to the kmemleak_init()
> function. By having all early kmemleak logs record the stack_trace, we
> can have kmemleak_init() write exactly where the problem occurred. This
> patch adds the setting of the kmemleak_warning variable every time a
> kmemleak warning is issued. The kmemleak_init() function checks this
> variable during early log replaying and prints the log trace if there
> was any warning.
>
> Reported-by: Steven Rostedt <rostedt@goodmis.org>
> Cc: Andrew Morton <akpm@google.com>
> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
> ---
>
> Steven, a simpler implementation of your idea. Basically we have
> kmemleak_warn() called in several places where errors actually occur. We
> can just reuse this for printing the early log information rather than
> changing the function prototypes to return an error code.

Acked-by: Steven Rostedt <rostedt@goodmis.org>

-- Steve




\
 
 \ /
  Last update: 2011-09-28 15:55    [from the cache]
©2003-2011 Jasper Spaans