On 07/31/2012 06:42 PM, Christoph Lameter wrote:

On Tue, 31 Jul 2012, Glauber Costa wrote:

quoted

Since you said you had reworked this, I'll just stop looking for now.
But would you please make sure that this following use case is well
tested before you send?

1) After machine is up, create a bogus cache
2) free that cache right away.
3) Create two more caches.

The creation of the second cache fails, because
kmem_cache_alloc(kmem_cache, x) returns bad values. Those bad values can
take multiple forms, but the most common is a value that is equal to an
already assigned value.

If you enable debugging you will see those issues right away and do not
need to infer from other problems that there is an issue in the
allocators.

Yes, but since deleting caches is not a common operation in the kernel,
you will have to force somehow.

I didn't see anything with debugging enabled, simply because there is no
problem at all before you remove the first cache.

--
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>