Re: [GIT PULL v2] Early SLAB fixes for 2.6.31

From: Pekka Enberg
Date: Mon Jun 15 2009 - 14:20:04 EST

Hi Linus,

On Mon, 15 Jun 2009, Pekka Enberg wrote:
>> Actually, there's a slight complication here. If I push gfp mask to
>> __might_sleep(), lockdep_trace_alloc() and so on, the mask is
>> effective _everywhere_ even outside of slab. Yes, it makes sense if we
>> push the masking right down to the page allocator but I wonder if
>> that's something we want to do at this point?

On Mon, Jun 15, 2009 at 8:15 PM, Linus
Torvalds<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> This actually doesn't sound like a complication to me, but a potential
> cleanup.
> Right now we already have that magic "system_state" test in __might_sleep.
> Maybe we could get rid of that, and replace it with that test for gpf
> bits. So we'd have just _one_ magic special case, and it's directly
> related to memory allocation (which is really the reason for that system
> state thing too).
> But maybe we have other reasons for that system_state special case, that
> are independent. I have not checked.

I'll double-check this but I think we can do that. The only
problematic one is __lockdep_trace_alloc() which is used by the page
allocator and the slab allocator but we only want to do masking in the
latter one. But I guess we can just introduce a
lockdep_trace_slab_alloc() or something that does it before calling
the internal function.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at