linux-kernel - Re: [RFC/RFT PATCH 1/2] dma-contiguous: Simplify dma_*_from

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: <20190430152058.GC25447@lst.de>
Date:   Tue, 30 Apr 2019 17:20:58 +0200
From:   Christoph Hellwig <hch@....de>
To:     Robin Murphy <robin.murphy@....com>
Cc:     Nicolin Chen <nicoleotsuka@...il.com>, hch@....de,
        m.szyprowski@...sung.com, vdumpa@...dia.com, linux@...linux.org.uk,
        catalin.marinas@....com, will.deacon@....com, chris@...kel.net,
        jcmvbkbc@...il.com, joro@...tes.org, dwmw2@...radead.org,
        tony@...mide.com, akpm@...ux-foundation.org, sfr@...b.auug.org.au,
        treding@...dia.com, keescook@...omium.org, iamjoonsoo.kim@....com,
        wsa+renesas@...g-engineering.com,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        linux-xtensa@...ux-xtensa.org, iommu@...ts.linux-foundation.org
Subject: Re: [RFC/RFT PATCH 1/2] dma-contiguous: Simplify
 dma_*_from_contiguous() function calls

On Tue, Apr 30, 2019 at 01:52:26PM +0100, Robin Murphy wrote:
> As Catalin pointed out before, many of the users above may still have 
> implicit assumptions about the default CMA area, i.e. that this won't 
> return something above the limit they originally passed to 
> dma_contiguous_reserve(). I'm not sure how straightforward that is to 
> resolve - at the very least we may have to monkey around with GFP_DMA{32} 
> flags based on where dma_contiguous_default_area lies :(

Or just convert the callers one by one.  The two most interesting ones
are dma-direct which always check addressability after the allocation,
and dma-iommu, which doesn't care.  dma-iommu.c and intel-iommu.c also
don't care, but should use dma-iommu by next merge window anyway,
which leaves arm which is so complicated that we better don't touch
it for now, and xtensa, which I hope to switch to dma-direct in the
next merge window or two.