linux-kernel - Re: [PATCH] dma-mapping: fix page attributes for dma_mmap

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: <20190801163457.GB26588@lst.de>
Date:   Thu, 1 Aug 2019 18:34:57 +0200
From:   Christoph Hellwig <hch@....de>
To:     Will Deacon <will@...nel.org>
Cc:     Christoph Hellwig <hch@....de>, iommu@...ts.linux-foundation.org,
        Shawn Anastasio <shawn@...stas.io>,
        Michael Ellerman <mpe@...erman.id.au>,
        Russell King <linux@...linux.org.uk>,
        Catalin Marinas <catalin.marinas@....com>,
        Robin Murphy <robin.murphy@....com>,
        linuxppc-dev@...ts.ozlabs.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] dma-mapping: fix page attributes for dma_mmap_*

On Thu, Aug 01, 2019 at 05:23:06PM +0100, Will Deacon wrote:
> > -	if (!dev_is_dma_coherent(dev) || (attrs & DMA_ATTR_WRITE_COMBINE))
> > -		return pgprot_writecombine(prot);
> > -	return prot;
> > +	return pgprot_writecombine(prot);
> >  }
> 
> Seems like a sensible cleanup to me:
> 
> Acked-by: Will Deacon <will@...nel.org>
> 
> Although arch_dma_mmap_pgprot() is a bit of a misnomer now that it only
> gets involved in the non-coherent case.

A better name is welcome.  My other idea would be to just remove it
entirely and do something like:

#ifndef pgprot_dmacoherent
#define pgprot_dmacoherent pgprot_noncached
#endif

pgprot_t dma_mmap_pgprot(struct device *dev, pgprot_t prot, unsigned long attrs)
{
	if (dev_is_dma_coherent(dev) || (attrs & DMA_ATTR_NON_CONSISTENT))
		return prot;
#ifdef pgprot_writecombine
	if (attrs & DMA_ATTR_WRITE_COMBINE)
		return pgprot_writecombine(prot);
#endif
	return pgprot_dmacoherent(prot);
}

But my worry is how this interacts with architectures that have an
uncached segment (mips, nios2, microblaze, extensa) where we'd have
the kernel access DMA_ATTR_WRITE_COMBINE mappigns using the uncached
segment, and userspace mmaps using pgprot_writecombine, which could
lead to aliasing issues.  But then again mips already supports
DMA_ATTR_WRITE_COMBINE, so this must be ok somehow.  I guess I'll
need to field that question to the relevant parties.