Re: /proc/meminfo question

[Vlastimil Babka <vbabka@suse.cz>]

On 04/04/2016 09:22 AM, Ming Lin wrote:
> Hi,
>
> I'm debugging a memory leak in an internal driver.
> There is ~2G leak after stress tests.
>
> If I read below /proc/meminfo correctly, it doesn't show where the 2G
> memory is possibly used.
>
> Buffers + Cached + Active*/Inactive* + Slab* only about 300M.
>
> Is there other statistics not shown in /proc/meminfo or am I missing
> obvious things?

Nope. Statistics exist only for the common cases. Having e.g. each 
driver add own line there wouldn't be feasible.

However I already considered and would welcome an explicit 
"Unaccounted:" line in /prec/meminfo, so that the amount of possibly 
leaked memory is immediately visible and one doesn't need to remember 
which lines are best to sum up and whether there were subtle changes 
depending on kernel version.

> root@target:~# page-types
>               flags	page-count       MB  symbolic-flags			long-symbolic-flags

For debugging such leaks I suggest you try the page_owner functionality 
instead.

Vlastimil

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>