Re: [PATCH 2/2] swiotlb: Enforce page alignment in swiotlb_alloc()

From: Christoph Hellwig <hch@lst.de>
To: "Petr Tesařík" <petr@tesarici.cz>
Cc: Christoph Hellwig <hch@lst.de>, Will Deacon <will@kernel.org>,
	linux-kernel@vger.kernel.org, kernel-team@android.com,
	iommu@lists.linux.dev,
	Marek Szyprowski <m.szyprowski@samsung.com>,
	Robin Murphy <robin.murphy@arm.com>,
	Petr Tesarik <petr.tesarik1@huawei-partners.com>,
	Dexuan Cui <decui@microsoft.com>
Subject: Re: [PATCH 2/2] swiotlb: Enforce page alignment in swiotlb_alloc()
Date: Mon, 29 Jan 2024 08:50:43 +0100	[thread overview]
Message-ID: <20240129075043.GA22019@lst.de> (raw)
In-Reply-To: <20240129084326.4667af6a@meshulam.tesarici.cz>

On Mon, Jan 29, 2024 at 08:43:26AM +0100, Petr Tesařík wrote:
> > So PFN_DOWN aligns the address and thus per se converting the unaligned
> > address isn't a problem.  That being said swiotlb obviously should never
> > allocate unaligned addresses, but the placement of this check feels
> > odd to me.  Also because it only catches swiotlb_alloc and not the
> > map side.
> 
> We may have to rethink how alignment constraints are interpreted. See
> also my reply to PATCH 1/2.
> 
> > Maybe just throw a WARN_ON_ONCE into slot_addr() ?
> 
> Yes.
> 
> Or, what if I write a KUnit test suite for swiotlb to combat this
> constant stream of various regressions?

Both sounds good to me.