linux-toradex.git/mm/kmemleak.c, branch v2.6.32.7

kmemleak: Check for NULL pointer returned by create_object()

2009-10-09T20:28:47+00:00

This patch adds NULL pointer checking in the early_alloc() function.

Reported-by: Paul E. McKenney 
Signed-off-by: Catalin Marinas 
Acked-by: Pekka Enberg 
Acked-by: Paul E. McKenney 
Signed-off-by: Linus Torvalds

kmemleak: Use GFP_ATOMIC for early_alloc().

2009-10-09T20:28:47+00:00

We can't use GFP_KERNEL inside rcu_read_lock().

Signed-off-by: Tetsuo Handa 
Signed-off-by: Catalin Marinas 
Signed-off-by: Linus Torvalds

kmemleak: Improve the "Early log buffer exceeded" error message

2009-09-11T09:42:09+00:00

Based on a suggestion from Jaswinder, clarify what the user would need
to do to avoid this error message from kmemleak.

Reported-by: Jaswinder Singh Rajput 
Signed-off-by: Catalin Marinas

kmemleak: fix sparse warning for static declarations

2009-09-08T16:34:07+00:00

This fixes these sparse warnings:

mm/kmemleak.c:1179:6: warning: symbol 'start_scan_thread' was not declared. Should it be static?
mm/kmemleak.c:1194:6: warning: symbol 'stop_scan_thread' was not declared. Should it be static?

Acked-by: Pekka Enberg 
Signed-off-by: Luis R. Rodriguez 
Signed-off-by: Catalin Marinas

kmemleak: fix sparse warning over overshadowed flags

2009-09-08T16:34:06+00:00

A secondary irq_save is not required as a locking before it was
already disabling irqs.

This fixes this sparse warning:
mm/kmemleak.c:512:31: warning: symbol 'flags' shadows an earlier one
mm/kmemleak.c:448:23: originally declared here

Signed-off-by: Luis R. Rodriguez 
Acked-by: Pekka Enberg 
Signed-off-by: Catalin Marinas

kmemleak: move common painting code together

2009-09-08T16:22:20+00:00

When painting grey or black we do the same thing, bring
this together into a helper and identify coloring grey or
black explicitly with defines. This makes this a little
easier to read.

Signed-off-by: Luis R. Rodriguez 
Signed-off-by: Catalin Marinas

kmemleak: add clear command support

2009-09-08T15:36:08+00:00

In an ideal world your kmemleak output will be small, when its
not (usually during initial bootup) you can use the clear command
to ingore previously reported and unreferenced kmemleak objects. We
do this by painting all currently reported unreferenced objects grey.
We paint them grey instead of black to allow future scans on the same
objects as such objects could still potentially reference newly
allocated objects in the future.

To test a critical section on demand with a clean
/sys/kernel/debug/kmemleak you can do:

echo clear > /sys/kernel/debug/kmemleak
        test your kernel or modules
echo scan > /sys/kernel/debug/kmemleak

Then as usual to get your report with:

cat /sys/kernel/debug/kmemleak

Signed-off-by: Luis R. Rodriguez 
Signed-off-by: Catalin Marinas

kmemleak: use bool for true/false questions

2009-09-08T15:34:50+00:00

Acked-by: Pekka Enberg 
Signed-off-by: Luis R. Rodriguez 
Signed-off-by: Catalin Marinas

kmemleak: Do no create the clean-up thread during kmemleak_disable()

2009-09-08T15:31:15+00:00

The kmemleak_disable() function could be called from various contexts
including IRQ. It creates a clean-up thread but the kthread_create()
function has restrictions on which contexts it can be called from,
mainly because of the kthread_create_lock. The patch changes the
kmemleak clean-up thread to a workqueue.

Signed-off-by: Catalin Marinas 
Reported-by: Eric Paris

kmemleak: Scan all thread stacks

2009-09-04T15:05:55+00:00

This patch changes the for_each_process() loop with the
do_each_thread()/while_each_thread() pair.

Signed-off-by: Catalin Marinas