Re: [PATCH]: Clean up of __alloc_pages

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Rohit Seth wrote:
On Sat, 2005-10-29 at 12:33 +1000, Nick Piggin wrote:

If you don't do this, then a GFP_HIGH allocator can allocate right
down to its limit before it kicks kswapd, then it either will fail or
will have to do direct reclaim.



You are right if there are only GFP_HIGH requests coming in then the
allocation will go down to (min - min/2) before kicking in kswapd.
Though if the requester is not ready to wait, there is another good shot
at allocation succeed before we get into direct reclaim (and this is
happening based on can_try_harder flag).


Still, it is a change in behaviour that I would rather not introduce
with a cleanup patch (and is something we don't want to introduce anyway).

So if you could fix that up it would be good.

How about moving the zone_statistics up into the 'if (page)'
test of get_page_from_freelist? This way we don't have to
evaluate page_zone().



Let us keep this as is for now.  Will revisit once after the
pcp_prefer_allocation patches get in place.

Well page_zone is yet another cacheline that doesn't need to be touched,
and that is introduced by this patch. But the line is likely to be hot,
and get_page_from_freelist does not have the required 'zonelist' which
I didn't notice before.

So OK, revisit this later.

--
SUSE Labs, Novell Inc.

Send instant messages to your online friends http://au.messenger.yahoo.com -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Stuff]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]     [Linux Resources]
  Powered by Linux