Because of the possible failure of the dma_supported(), the
dma_set_mask_and_coherent() may return error num.
Therefore, it should be better to check it and return the error if
fails.
Fixes: 1c894225bf5b ("iommu/ipmmu-vmsa: IPMMU device is 40-bit bus master")
Signed-off-by: Jiasheng Jiang <[email protected]>
---
drivers/iommu/ipmmu-vmsa.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c
index aaa6a4d59057..7df5da44a004 100644
--- a/drivers/iommu/ipmmu-vmsa.c
+++ b/drivers/iommu/ipmmu-vmsa.c
@@ -1003,7 +1003,9 @@ static int ipmmu_probe(struct platform_device *pdev)
bitmap_zero(mmu->ctx, IPMMU_CTX_MAX);
mmu->features = of_device_get_match_data(&pdev->dev);
memset(mmu->utlb_ctx, IPMMU_CTX_INVALID, mmu->features->num_utlbs);
- dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
+ ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
+ if (ret)
+ return ret;
/* Map I/O memory and request IRQ. */
res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
--
2.25.1
Adding more potential reviewers.
On Thu, Jan 06, 2022 at 10:43:02AM +0800, Jiasheng Jiang wrote:
> Because of the possible failure of the dma_supported(), the
> dma_set_mask_and_coherent() may return error num.
> Therefore, it should be better to check it and return the error if
> fails.
>
> Fixes: 1c894225bf5b ("iommu/ipmmu-vmsa: IPMMU device is 40-bit bus master")
> Signed-off-by: Jiasheng Jiang <[email protected]>
> ---
> drivers/iommu/ipmmu-vmsa.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c
> index aaa6a4d59057..7df5da44a004 100644
> --- a/drivers/iommu/ipmmu-vmsa.c
> +++ b/drivers/iommu/ipmmu-vmsa.c
> @@ -1003,7 +1003,9 @@ static int ipmmu_probe(struct platform_device *pdev)
> bitmap_zero(mmu->ctx, IPMMU_CTX_MAX);
> mmu->features = of_device_get_match_data(&pdev->dev);
> memset(mmu->utlb_ctx, IPMMU_CTX_INVALID, mmu->features->num_utlbs);
> - dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
> + ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
> + if (ret)
> + return ret;
>
> /* Map I/O memory and request IRQ. */
> res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> --
> 2.25.1
On Mon, Feb 14, 2022 at 12:32:21PM +0000, Robin Murphy wrote:
> In this particular case it cannot fail on any system the driver actually
> runs on - it's a platform device so the dma_mask pointer is always
> initialised, then dma_direct_supported() on arm64 will always return true
> for any mask wider than 32 bits, while arm_dma_supported() will also always
> pass since a 32-bit system cannot have memory above 40 bits either.
>
> There's no great harm in adding the check for the sake of consistency, I
> guess, but it's purely cosmetic and not fixing anything.
Okay, thanks Robin and Nikita for looking to quickly into this. I will
apply it for 5.18 just to be on the safe side if the above assumptions
change.
Regards,
Joerg
> Adding more potential reviewers.
>
> On Thu, Jan 06, 2022 at 10:43:02AM +0800, Jiasheng Jiang wrote:
>> Because of the possible failure of the dma_supported(), the
>> dma_set_mask_and_coherent() may return error num.
>> Therefore, it should be better to check it and return the error if
>> fails.
Indeed, most uses of dma_set_mask_and_coherent() in the kernel do check return value.
Reviewed-by: Nikita Yushchenko <[email protected]>
On 2022-02-14 11:43, Joerg Roedel wrote:
> Adding more potential reviewers.
>
> On Thu, Jan 06, 2022 at 10:43:02AM +0800, Jiasheng Jiang wrote:
>> Because of the possible failure of the dma_supported(), the
>> dma_set_mask_and_coherent() may return error num.
>> Therefore, it should be better to check it and return the error if
>> fails.
In this particular case it cannot fail on any system the driver actually
runs on - it's a platform device so the dma_mask pointer is always
initialised, then dma_direct_supported() on arm64 will always return
true for any mask wider than 32 bits, while arm_dma_supported() will
also always pass since a 32-bit system cannot have memory above 40 bits
either.
There's no great harm in adding the check for the sake of consistency, I
guess, but it's purely cosmetic and not fixing anything.
Thanks,
Robin.
>> Fixes: 1c894225bf5b ("iommu/ipmmu-vmsa: IPMMU device is 40-bit bus master")
>> Signed-off-by: Jiasheng Jiang <[email protected]>
>> ---
>> drivers/iommu/ipmmu-vmsa.c | 4 +++-
>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c
>> index aaa6a4d59057..7df5da44a004 100644
>> --- a/drivers/iommu/ipmmu-vmsa.c
>> +++ b/drivers/iommu/ipmmu-vmsa.c
>> @@ -1003,7 +1003,9 @@ static int ipmmu_probe(struct platform_device *pdev)
>> bitmap_zero(mmu->ctx, IPMMU_CTX_MAX);
>> mmu->features = of_device_get_match_data(&pdev->dev);
>> memset(mmu->utlb_ctx, IPMMU_CTX_INVALID, mmu->features->num_utlbs);
>> - dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
>> + ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
>> + if (ret)
>> + return ret;
>>
>> /* Map I/O memory and request IRQ. */
>> res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> --
>> 2.25.1