scsi_dma_map() was reporting a failure during boot on an AMD machine
with the IOMMU enabled.
scsi_dma_map failed: request for 36 bytes!
The issue was tracked down to a mistake in logic: should not return
an error if iommu_deferred_attach() returns zero.
Reported-by: Marshall Midden <[email protected]>
Fixes: dabb16f67215 ("iommu/dma: return error code from iommu_dma_map_sg()")
Link: https://lore.kernel.org/all/CAD2CkAWjS8=kKwEEN4cgVNjyFORUibzEiCUA-X+SMtbo0JoMmA@mail.gmail.com
Signed-off-by: Logan Gunthorpe <[email protected]>
Cc: Joerg Roedel <[email protected]>
Cc: Will Deacon <[email protected]>
---
drivers/iommu/dma-iommu.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
If possible, this fix should really get into Linus's tree before the
v5.15 release.
diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 896bea04c347..3e5a21b0bb24 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -1016,7 +1016,8 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg,
if (static_branch_unlikely(&iommu_deferred_attach_enabled)) {
ret = iommu_deferred_attach(dev, domain);
- goto out;
+ if (ret)
+ goto out;
}
if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
base-commit: 3906fe9bb7f1a2c8667ae54e967dc8690824f4ea
--
2.30.2
On 27/10/2021 18:47, Logan Gunthorpe wrote:
> scsi_dma_map() was reporting a failure during boot on an AMD machine
> with the IOMMU enabled.
>
> scsi_dma_map failed: request for 36 bytes!
>
> The issue was tracked down to a mistake in logic: should not return
> an error if iommu_deferred_attach() returns zero.
>
> Reported-by: Marshall Midden<[email protected]>
> Fixes: dabb16f67215 ("iommu/dma: return error code from iommu_dma_map_sg()")
> Link:https://lore.kernel.org/all/CAD2CkAWjS8=kKwEEN4cgVNjyFORUibzEiCUA-X+SMtbo0JoMmA@mail.gmail.com
> Signed-off-by: Logan Gunthorpe<[email protected]>
> Cc: Joerg Roedel<[email protected]>
> Cc: Will Deacon<[email protected]>
Seems reasonable, FWIW:
Reviewed-by: John Garry <[email protected]>