Hi David,
On Mon, 23 Aug 2010, Pekka Enberg wrote:
>> > Hows about you add a helper function
>> >
>> > void *[kmalloc|alloc_page]_retrying_forever_because_i_suck(lots of args)
>> >
>> > then convert the callsites to use that, then nuke __GFP_NOFAIL?
>>
>> I'd prefer killing off __GFP_NOFAIL properly :-)
On Mon, Aug 23, 2010 at 11:13 PM, David Rientjes <rientjes@xxxxxxxxxx> wrote:
> And how is this not done properly if it's not even needed for any of the
> allocations in this patchset since the page allocator loops forever for
> their orders and context? (This is why we don't need to add __GFP_NOWARN
> in its place.)
>
> This is a cleanup patchset to remove the unnecessary use of __GFP_NOFAIL,
> there _are_ GFP_KERNEL | __GFP_NOFAIL allocations that need to be
> addressed in phase three.
My point is that I don't think Andrew's helper will change all that
much. Fixing the actual callers is the hard part and I don't see your
patches helping that either. Hiding the looping in filesystem code is
only going to make problematic callers harder to find (e.g
kmem_alloc() in XFS code).
FWIW, I'd just add a nasty WARN_ON in the page allocator and put a big
fat "fix your shit" comment on top of it to motivate people to fix
their code.
Pekka
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
