Received-SPF: pass (google.com: domain of linux-kernel+bounces-97389-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Fri, 8 Mar 2024 18:28:17 +0100
From: Petr =?UTF-8?B?VGVzYcWZw61r?= <petr@tesarici.cz>
To: Will Deacon <will@kernel.org>
Cc: linux-kernel@vger.kernel.org, kernel-team@android.com,
 iommu@lists.linux.dev, Christoph Hellwig <hch@lst.de>, Marek Szyprowski
 <m.szyprowski@samsung.com>, Robin Murphy <robin.murphy@arm.com>, Petr
 Tesarik <petr.tesarik1@huawei-partners.com>, Dexuan Cui
 <decui@microsoft.com>, Nicolin Chen <nicolinc@nvidia.com>, Michael Kelley
 <mhklinux@outlook.com>
Subject: Re: [PATCH v6 6/6] swiotlb: Reinstate page-alignment for mappings
 >= PAGE_SIZE
Message-ID: <20240308182817.7687329f@meshulam.tesarici.cz>
In-Reply-To: <20240308152829.25754-7-will@kernel.org>
References: <20240308152829.25754-1-will@kernel.org>
	<20240308152829.25754-7-will@kernel.org>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

On Fri,  8 Mar 2024 15:28:29 +0000
Will Deacon <will@kernel.org> wrote:

> For swiotlb allocations >= PAGE_SIZE, the slab search historically
> adjusted the stride to avoid checking unaligned slots. This had the
> side-effect of aligning large mapping requests to PAGE_SIZE, but that
> was broken by 0eee5ae10256 ("swiotlb: fix slot alignment checks").
> 
> Since this alignment could be relied upon drivers, reinstate PAGE_SIZE
> alignment for swiotlb mappings >= PAGE_SIZE.
> 
> Reported-by: Michael Kelley <mhklinux@outlook.com>
> Signed-off-by: Will Deacon <will@kernel.org>
> 
> ---
>  kernel/dma/swiotlb.c | 18 +++++++++++-------
>  1 file changed, 11 insertions(+), 7 deletions(-)
> 
> diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
> index c381a7ed718f..c5851034523f 100644
> --- a/kernel/dma/swiotlb.c
> +++ b/kernel/dma/swiotlb.c
> @@ -992,6 +992,17 @@ static int swiotlb_search_pool_area(struct device *dev, struct io_tlb_pool *pool
>  	BUG_ON(!nslots);
>  	BUG_ON(area_index >= pool->nareas);
>  
> +	/*
> +	 * Historically, swiotlb allocations >= PAGE_SIZE were guaranteed to be
> +	 * page-aligned in the absence of any other alignment requirements.
> +	 * 'alloc_align_mask' was later introduced to specify the alignment
> +	 * explicitly, however this is passed as zero for streaming mappings
> +	 * and so we preserve the old behaviour there in case any drivers are
> +	 * relying on it.
> +	 */
> +	if (!alloc_align_mask && !iotlb_align_mask && alloc_size >= PAGE_SIZE)
> +		alloc_align_mask = PAGE_SIZE - 1;
> +

This could be moved up the call chain to swiotlb_tbl_map_single(), but
since there is already a plan to modify the call signatures for the
next release, let's keep it here.

Reviewed-by: Petr Tesarik <petr.tesarik1@huawei-partners.com>

Petr T

>  	/*
>  	 * Ensure that the allocation is at least slot-aligned and update
>  	 * 'iotlb_align_mask' to ignore bits that will be preserved when
> @@ -1006,13 +1017,6 @@ static int swiotlb_search_pool_area(struct device *dev, struct io_tlb_pool *pool
>  	 */
>  	stride = get_max_slots(max(alloc_align_mask, iotlb_align_mask));
>  
> -	/*
> -	 * For allocations of PAGE_SIZE or larger only look for page aligned
> -	 * allocations.
> -	 */
> -	if (alloc_size >= PAGE_SIZE)
> -		stride = umax(stride, PAGE_SHIFT - IO_TLB_SHIFT + 1);
> -
>  	spin_lock_irqsave(&area->lock, flags);
>  	if (unlikely(nslots > pool->area_nslabs - area->used))
>  		goto not_found;