[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <92ebdc16-7938-462a-bb0e-4d5d368c5eb1@arm.com>
Date: Tue, 27 Aug 2024 13:19:13 +0100
From: Robin Murphy <robin.murphy@....com>
To: Christoph Hellwig <hch@....de>, iommu@...ts.linux.dev
Cc: "Martin K. Petersen" <martin.petersen@...cle.com>,
Marek Szyprowski <m.szyprowski@...sung.com>, linux-scsi@...r.kernel.org,
linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org,
dri-devel@...ts.freedesktop.org, dmaengine@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org, linux-mediatek@...ts.infradead.org,
linux-media@...r.kernel.org, linux-mmc@...r.kernel.org,
linux-stm32@...md-mailman.stormreply.com, linux-hyperv@...r.kernel.org,
netdev@...r.kernel.org
Subject: Re: [PATCH 1/4] scsi: check that busses support the DMA API before
setting dma parameters
On 24/08/2024 4:49 am, Christoph Hellwig wrote:
> We'll start throwing warnings soon when dma_set_seg_boundary and
> dma_set_max_seg_size are called on devices for buses that don't fully
> support the DMA API. Prepare for that by making the calls in the SCSI
> midlayer conditional.
Just thinking ahead, might it be worth a logical "are SG segment limits
relevant?" wrapper around the dev->dma_parms reference? Not a big deal
for now if we think this site is the only user, so either way,
Reviewed-by: Robin Murphy <robin.murphy@....com>
> Signed-off-by: Christoph Hellwig <hch@....de>
> ---
> drivers/scsi/scsi_lib.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
> index 3958a6d14bf457..7f0394c4492033 100644
> --- a/drivers/scsi/scsi_lib.c
> +++ b/drivers/scsi/scsi_lib.c
> @@ -1988,8 +1988,15 @@ void scsi_init_limits(struct Scsi_Host *shost, struct queue_limits *lim)
> if (shost->no_highmem)
> lim->features |= BLK_FEAT_BOUNCE_HIGH;
>
> - dma_set_seg_boundary(dev, shost->dma_boundary);
> - dma_set_max_seg_size(dev, shost->max_segment_size);
> + /*
> + * Propagate the DMA formation properties to the dma-mapping layer as
> + * a courtesy service to the LLDDs. This needs to check that the buses
> + * actually support the DMA API first, though.
> + */
> + if (dev->dma_parms) {
> + dma_set_seg_boundary(dev, shost->dma_boundary);
> + dma_set_max_seg_size(dev, shost->max_segment_size);
> + }
> }
> EXPORT_SYMBOL_GPL(scsi_init_limits);
>
Powered by blists - more mailing lists