linux-kernel - Re: [PATCH v2 1/1] iommu/dma: Use DMA ops setter instead of direct assignment

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 for Android: free password hash cracker in your pocket

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

Message-ID: <YgOrpC/Wg2MA7eRy@smile.fi.intel.com>
Date:   Wed, 9 Feb 2022 13:55:16 +0200
From:   Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To:     Christoph Hellwig <hch@...radead.org>
Cc:     Robin Murphy <robin.murphy@....com>,
        Joerg Roedel <jroedel@...e.de>,
        iommu@...ts.linux-foundation.org, linux-kernel@...r.kernel.org,
        Will Deacon <will@...nel.org>
Subject: Re: [PATCH v2 1/1] iommu/dma: Use DMA ops setter instead of direct
 assignment

On Tue, Feb 08, 2022 at 11:06:56PM -0800, Christoph Hellwig wrote:
> On Mon, Feb 07, 2022 at 03:55:32PM +0000, Robin Murphy wrote:
> > On 2022-02-07 14:13, Andy Shevchenko wrote:
> > > Use DMA ops setter instead of direct assignment. Even we know that
> > > this module doesn't perform access to the dma_ops member of struct device,
> > > it's better to use setter to avoid potential problems in the future.
> > 
> > What potential problems are you imagining? This whole file is a DMA ops
> > implementation, not driver code (and definitely not a module); if anyone
> > removes the "select DMA_OPS" from CONFIG_IOMMU_DMA they deserve whatever
> > breakage they get.
> > 
> > I concur that there's no major harm in using the helper here, but I also see
> > no point in pretending that there's any value to abstracting the operation
> > in this particular context.
> 
> Yeah.  Killing off the the wrapper is actually on my todo list, mostly
> because it leads to people doing completely broken things like the VDPA
> private dma ops that should not exist.

Let's abandon this change. (I see that it's kinda 50/50 of the users
with API and without)

-- 
With Best Regards,
Andy Shevchenko