linux-kernel - Re: [PATCH 2/2] dma: Add IOMMU static calls with clear default ops

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] [day] [month] [year] [list]

Message-ID: <20240713093259.GH1815706@unreal>
Date: Sat, 13 Jul 2024 12:32:59 +0300
From: Leon Romanovsky <leon@...nel.org>
To: Christoph Hellwig <hch@....de>
Cc: Robin Murphy <robin.murphy@....com>, Joerg Roedel <joro@...tes.org>,
	Will Deacon <will@...nel.org>,
	Marek Szyprowski <m.szyprowski@...sung.com>,
	Jason Gunthorpe <jgg@...dia.com>, linux-kernel@...r.kernel.org,
	iommu@...ts.linux.dev
Subject: Re: [PATCH 2/2] dma: Add IOMMU static calls with clear default ops

On Sat, Jul 13, 2024 at 07:18:13AM +0200, Christoph Hellwig wrote:
> On Fri, Jul 12, 2024 at 03:21:55PM +0100, Robin Murphy wrote:
> >> This is done to keep layering similar to existing in DMA subsystem. We
> >> have special files and calls to dma-direct, it looks natural to have
> >> special files and call to dma-iommu. It is not nice to call to drivers/iommu
> >> from kernel/dma/mapping.c
> >
> > That's where I firmly disagree. In the DMA API aspect, iommu-dma is exactly 
> > a peer of dma-direct,
> 
> Exactly.
> 
> > however it lives in drivers/iommu for practical 
> > reasons because it's more closely coupled to IOMMU API internals in all 
> > other aspects of its implementation.
> 
> TBH I think kernel/dma/ would be the better place.  But that's really the
> least my concernes at the moment.
> 
> But the important point is that I would really prefer to avoid another
> magic layer - just do direct calls like for dma-direct to keep it
> understandable (and probably faster).

I'll change my patch to be without extra layer, but is it unlikely that
it will make any difference in performance.

Thanks

>