Prev: [PATCHv5 00/11] g_multi & other improvements and cleanpus
Next: KVM: MMU: introduce gfn_to_page_atomic() and gfn_to_pfn_atomic()
From: Catalin Marinas on 16 Jun 2010 06:20
On 16 June 2010 09:52, Amit Agarwal <amit.agarr(a)gmail.com> wrote:
> I backported the latest kmemleak(2.6.34-rc1) patches to my 2.6.29-RT kernel
> for ARM target.
> I am facing the issue while inserting the kmemleak-test.ko module twice in
> my 29-RT kernel.
> Step I followed.
> {{{
> # mount -t debugfs nodev /sys/kernel/debug/ # cat /sys/kernel/debug/kmemleak
> # echo scan > /sys/kernel/debug/kmemleak # cat /sys/kernel/debug/kmemleak
> �# insmod kmemleak-test.ko
> �# rmmod kmemleak-test.ko
> # insmod kmemleak-test.ko
> # cat /sys/kernel/debug/kmemleak
> cat: can't open '/sys/kernel/debug/kmemleak': Device or resource busy
> }}}
> I have attached the�detail�log with the mail.
> I am not clear is this the behaviour of kmemleak

It works fine on the latest kernel, so it's not a kmemleak bug. Most
likely it's a back-porting issue. From the log, kmemleak disabled
itself because of errors (freeing some blocks of memory by a pointer
to somewhere inside the block rather than to the beginning).

All I can say is make sure that the kmemleak_*() callbacks are
correct. The log gives you the backtrace of where kmemleak_free() was
called from, so that should help.

--
Catalin
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 | 
Pages: 1
Prev: [PATCHv5 00/11] g_multi & other improvements and cleanpus
Next: KVM: MMU: introduce gfn_to_page_atomic() and gfn_to_pfn_atomic()