Le 22/01/2022 à 19:05, Sinan Kaya a écrit :
On 1/15/2022 2:35 AM, Christophe JAILLET wrote:
As stated in [1], dma_set_mask() with a 64-bit mask never fails if
dev->dma_mask is non-NULL.
So, if it fails, the 32 bits case will also fail for the same reason.
Simplify code and remove some dead code accordingly.
[1]:
https://lore.kernel.org/linux-kernel/YL3vSPK5DXTNvgdx@xxxxxxxxxxxxx/#t
Can we please document this?
Hi, the patch has been applied, but [1] is sometimes given as an
explanation link.
CJ
[1]:
https://lists.linuxfoundation.org/pipermail/iommu/2019-February/033674.html
Usual practice was to try allocating 64 bit DMA if possible and fallback
to 32 bits.
Signed-off-by: Christophe JAILLET <christophe.jaillet@xxxxxxxxxx>
---
v2: have the subject and updated driver match
---
drivers/dma/qcom/hidma.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/dma/qcom/hidma.c b/drivers/dma/qcom/hidma.c
index 65d054bb11aa..51587cf8196b 100644
--- a/drivers/dma/qcom/hidma.c
+++ b/drivers/dma/qcom/hidma.c
@@ -838,9 +838,7 @@ static int hidma_probe(struct platform_device *pdev)
rc = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
if (rc) {
dev_warn(&pdev->dev, "unable to set coherent mask to 64");
- rc = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32));
- if (rc)
- goto dmafree;
+ goto dmafree;
}
dmadev->lldev = hidma_ll_init(dmadev->ddev.dev,
