linux-kernel - Re: [PATCH 03/14] x86: use dma-direct

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]

Date:   Thu, 15 Mar 2018 12:53:48 +0100
From:   Christoph Hellwig <hch@....de>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     Christoph Hellwig <hch@....de>, x86@...nel.org,
        Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>,
        Tom Lendacky <thomas.lendacky@....com>,
        David Woodhouse <dwmw2@...radead.org>,
        Muli Ben-Yehuda <mulix@...ix.org>,
        Jon Mason <jdmason@...zu.us>, Joerg Roedel <joro@...tes.org>,
        iommu@...ts.linux-foundation.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 03/14] x86: use dma-direct

On Thu, Mar 15, 2018 at 09:56:13AM +0100, Thomas Gleixner wrote:
> On Wed, 14 Mar 2018, Christoph Hellwig wrote:
> 
> > The generic dma-direct implementation is now functionally equivalent to
> > the x86 nommu dma_map implementation, so switch over to using it.
> 
> Can you please convert the various drivers first and then remove the
> unused code?

Which various drivers?

> > Note that the various iommu drivers are switched from x86_dma_supported
> > to dma_direct_supported to provide identical functionality, although the
> > checks looks fairly questionable for at least some of them.
> 
> Can you please elaborate? From the above it's not clear which checks you
> are referring to. If you convert these drivers seperately then explicit
> information about your concerns wants to be in the changelogs.

This bit:

	/* Copied from i386. Doesn't make much sense, because it will
	   only work for pci_alloc_coherent.
	   The caller just has to use GFP_DMA in this case. */
	if (mask < DMA_BIT_MASK(24))
		return 0;

in x86_dma_supported, or the equivalent bit in dma_direct_supported.
Kept for bug to bug compatibility, but I guess I should reword or
just drop the changelog bit іf it causes confusion.