2021-10-13 17:03:09

by Logan Gunthorpe

Subject: [PATCH] RDMA/core: Set sgtable nents when using ib_dma_virt_map_sg()

ib_dma_map_sgtable_attrs() should be mapping the sgls and setting nents
but the ib_uses_virt_dma() path falls back to ib_dma_virt_map_sg()
which will not set the nents in the sgtable.

Check the return value (per the map_sg calling convention) and set
sgt->nents appropriately on success.

Link: https://lore.kernel.org/all/[email protected]/T/#u
Reported-by: Bart Van Assche <[email protected]>
Signed-off-by: Logan Gunthorpe <[email protected]>
Tested-by: Bart Van Assche <[email protected]>
---
include/rdma/ib_verbs.h | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 4b50d9a3018a..4ba642fc8a19 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -4097,8 +4097,13 @@ static inline int ib_dma_map_sgtable_attrs(struct ib_device *dev,
enum dma_data_direction direction,
unsigned long dma_attrs)
{
+ int nents;
+
if (ib_uses_virt_dma(dev)) {
- ib_dma_virt_map_sg(dev, sgt->sgl, sgt->orig_nents);
+ nents = ib_dma_virt_map_sg(dev, sgt->sgl, sgt->orig_nents);
+ if (!nents)
+ return -EIO;
+ sgt->nents = nents;
return 0;
}
return dma_map_sgtable(dev->dma_device, sgt, direction, dma_attrs);

base-commit: 2a152512a155aaf27c3e67834ffafaed9525a7b5
--
2.30.2

2021-10-13 18:37:14

by Bart Van Assche

Subject: Re: [PATCH] RDMA/core: Set sgtable nents when using ib_dma_virt_map_sg()

On 10/13/21 9:59 AM, Logan Gunthorpe wrote:
> ib_dma_map_sgtable_attrs() should be mapping the sgls and setting nents
> but the ib_uses_virt_dma() path falls back to ib_dma_virt_map_sg()
> which will not set the nents in the sgtable.
>
> Check the return value (per the map_sg calling convention) and set
> sgt->nents appropriately on success.
>
> Link: https://lore.kernel.org/all/[email protected]/T/#u
> Reported-by: Bart Van Assche <[email protected]>
> Signed-off-by: Logan Gunthorpe <[email protected]>
> Tested-by: Bart Van Assche <[email protected]>

Does this patch need a "Fixes:" tag?

Thanks,

Bart.

2021-10-13 18:40:23

by Logan Gunthorpe

Subject: Re: [PATCH] RDMA/core: Set sgtable nents when using ib_dma_virt_map_sg()

On 2021-10-13 12:33 p.m., Bart Van Assche wrote:
> On 10/13/21 9:59 AM, Logan Gunthorpe wrote:
>> ib_dma_map_sgtable_attrs() should be mapping the sgls and setting nents
>> but the ib_uses_virt_dma() path falls back to ib_dma_virt_map_sg()
>> which will not set the nents in the sgtable.
>>
>> Check the return value (per the map_sg calling convention) and set
>> sgt->nents appropriately on success.
>>
>> Link:
>> https://lore.kernel.org/all/[email protected]/T/#u
>>
>> Reported-by: Bart Van Assche <[email protected]>
>> Signed-off-by: Logan Gunthorpe <[email protected]>
>> Tested-by: Bart Van Assche <[email protected]>
>
> Does this patch need a "Fixes:" tag?

Right, yeah:

Fixes: 79fbd3e1241c ("RDMA: Use the sg_table directly and remove the
opencoded version from umem")

Logan

2021-10-13 19:48:33

by Jason Gunthorpe

Subject: Re: [PATCH] RDMA/core: Set sgtable nents when using ib_dma_virt_map_sg()

On Wed, Oct 13, 2021 at 10:59:42AM -0600, Logan Gunthorpe wrote:
> ib_dma_map_sgtable_attrs() should be mapping the sgls and setting nents
> but the ib_uses_virt_dma() path falls back to ib_dma_virt_map_sg()
> which will not set the nents in the sgtable.
>
> Check the return value (per the map_sg calling convention) and set
> sgt->nents appropriately on success.
>
> Link: https://lore.kernel.org/all/[email protected]/T/#u
> Reported-by: Bart Van Assche <[email protected]>
> Signed-off-by: Logan Gunthorpe <[email protected]>
> Tested-by: Bart Van Assche <[email protected]>
> ---
> include/rdma/ib_verbs.h | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)

Applied to for-next, thanks

Jason