linux-kernel - Re: [PATCH 3/4] Embed zone_id information within the zonelist->zones pointer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20070813234217.GI3406@bingen.suse.de>
Date:	Tue, 14 Aug 2007 01:42:17 +0200
From:	Andi Kleen <ak@...e.de>
To:	Christoph Lameter <clameter@....com>
Cc:	Andi Kleen <ak@...e.de>, Mel Gorman <mel@...net.ie>,
	Lee.Schermerhorn@...com, linux-kernel@...r.kernel.org,
	linux-mm@...ck.org
Subject: Re: [PATCH 3/4] Embed zone_id information within the zonelist->zones pointer

On Mon, Aug 13, 2007 at 03:22:25PM -0700, Christoph Lameter wrote:
> On Tue, 14 Aug 2007, Andi Kleen wrote:
> 
> > > x86_64 is the only platforms that uses ZONE_DMA32. Ia64 and other 64 bit 
> > > platforms use ZONE_DMA for <4GB allocs.
> > 
> > Yes, but ZONE_DMA32 == ZONE_DMA.
> 
> I am not sure what you mean by that. Ia64 ZONE_DMA == x86_84 ZONE_DMA32?

Hmm, when I wrote GFP_DMA32 it was a #define GFP_DMA32 GFP_DMA 
on ia64 so that drivers not need to ifdef.  Someone nasty
seems to have removed that too. I guess it would be best
to readd.

> 
> > Also when the slab users of GFP_DMA are all gone ia64 won't need
> > the slab support anymore. So either you change your ifdef in slub or 
> > switch to ZONE_DMA32 for IA64.
> 
> If you have gotten rid of all slab users of GFP_DMA (and also all arch 
> uses of it) then we can drop the code in SLAB.

No, e.g. s390 and some other architectures still use it.
You'll need to bug their respective maintainers.

> 1. Drop sl?b support for GFP_DMA.

Not yet.

> 
> 2. Drop GFP_DMA32 support.
> 
> Then we only allow page allocator allocs using GFP_DMA? That may be the 

Kind of yes.

> least invasive for arch code.

I would prefer for GFP_DMA to go away on x86 (but GFP_DMA32 stay). This way
we get clean compile errors instead of subtle breakage. Silently
changing the semantics would be bad.

But then it wouldn't make sense to have GFP_DMA on ia64 and GFP_DMA32
on x86. Since driver writers are more likely to test on x86
I would recommend ia64 having compatible semantics. It'll
save everybody trouble long term. This means it wouldn't 
help on IA64 machines that don't have a DMA zone -- they
would still need to validate drivers especially -- but at least
the others.

Also from my driver review driver authors often seem to have
trouble understanding what GFP_DMA really does. With GFP_DMA32 it 
is clearer that it applies to a address range and is not
some synonym for pci_map_*

-Andi

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/