diff mbox series

[vhost] vdpa/mlx5: Fix oversized null mkey longer than 32bit

Message ID 20250217194443.145601-1-dtatulea@nvidia.com (mailing list archive)
State New
Headers show
Series [vhost] vdpa/mlx5: Fix oversized null mkey longer than 32bit | expand

Commit Message

Dragos Tatulea Feb. 17, 2025, 7:44 p.m. UTC
From: Si-Wei Liu <si-wei.liu@oracle.com>

create_user_mr() has correct code to count the number of null keys
used to fill in a hole for the memory map. However, fill_indir()
does not follow the same to cap the range up to the 1GB limit
correspondinly. Fill in more null keys for the gaps in between,
so that null keys are correctly populated.

Fixes: 94abbccdf291 ("vdpa/mlx5: Add shared memory registration code")
Cc: stable@vger.kernel.org
Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
---
 drivers/vdpa/mlx5/core/mr.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

Comments

Eugenio Perez Martin Feb. 18, 2025, 8:14 a.m. UTC | #1
On Mon, Feb 17, 2025 at 8:45 PM Dragos Tatulea <dtatulea@nvidia.com> wrote:
>
> From: Si-Wei Liu <si-wei.liu@oracle.com>
>
> create_user_mr() has correct code to count the number of null keys
> used to fill in a hole for the memory map. However, fill_indir()
> does not follow the same to cap the range up to the 1GB limit
> correspondinly.

s/correspondinly/correspondingly/g

Sounds to me the logic can be merged in a helper?

Either way,

Acked-by: Eugenio Pérez <eperezma@redhat.com>

Thanks!

> Fill in more null keys for the gaps in between,
> so that null keys are correctly populated.
>
> Fixes: 94abbccdf291 ("vdpa/mlx5: Add shared memory registration code")
> Cc: stable@vger.kernel.org
> Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
> Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
> ---
>  drivers/vdpa/mlx5/core/mr.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/vdpa/mlx5/core/mr.c b/drivers/vdpa/mlx5/core/mr.c
> index 8455f08f5d40..61424342c096 100644
> --- a/drivers/vdpa/mlx5/core/mr.c
> +++ b/drivers/vdpa/mlx5/core/mr.c
> @@ -190,9 +190,12 @@ static void fill_indir(struct mlx5_vdpa_dev *mvdev, struct mlx5_vdpa_mr *mkey, v
>                         klm->bcount = cpu_to_be32(klm_bcount(dmr->end - dmr->start));
>                         preve = dmr->end;
>                 } else {
> +                       u64 bcount = min_t(u64, dmr->start - preve, MAX_KLM_SIZE);
> +
>                         klm->key = cpu_to_be32(mvdev->res.null_mkey);
> -                       klm->bcount = cpu_to_be32(klm_bcount(dmr->start - preve));
> -                       preve = dmr->start;
> +                       klm->bcount = cpu_to_be32(klm_bcount(bcount));
> +                       preve += bcount;
> +
>                         goto again;
>                 }
>         }
> --
> 2.43.0
>
Dragos Tatulea Feb. 18, 2025, 7:43 p.m. UTC | #2
On 02/18, Eugenio Perez Martin wrote:
> On Mon, Feb 17, 2025 at 8:45 PM Dragos Tatulea <dtatulea@nvidia.com> wrote:
> >
> > From: Si-Wei Liu <si-wei.liu@oracle.com>
> >
> > create_user_mr() has correct code to count the number of null keys
> > used to fill in a hole for the memory map. However, fill_indir()
> > does not follow the same to cap the range up to the 1GB limit
> > correspondinly.
> 
> s/correspondinly/correspondingly/g
>
Will fix in v2.

> Sounds to me the logic can be merged in a helper?
>
Not sure if possible in a useful way: the logic in create_user_mr() is
different.

Also: this patch is kept small for stable tre.

> Either way,
> 
> Acked-by: Eugenio Pérez <eperezma@redhat.com>
> 
Thanks!

> 
> > Fill in more null keys for the gaps in between,
> > so that null keys are correctly populated.
> >
> > Fixes: 94abbccdf291 ("vdpa/mlx5: Add shared memory registration code")
> > Cc: stable@vger.kernel.org
> > Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
> > Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
> > ---
> >  drivers/vdpa/mlx5/core/mr.c | 7 +++++--
> >  1 file changed, 5 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/vdpa/mlx5/core/mr.c b/drivers/vdpa/mlx5/core/mr.c
> > index 8455f08f5d40..61424342c096 100644
> > --- a/drivers/vdpa/mlx5/core/mr.c
> > +++ b/drivers/vdpa/mlx5/core/mr.c
> > @@ -190,9 +190,12 @@ static void fill_indir(struct mlx5_vdpa_dev *mvdev, struct mlx5_vdpa_mr *mkey, v
> >                         klm->bcount = cpu_to_be32(klm_bcount(dmr->end - dmr->start));
> >                         preve = dmr->end;
> >                 } else {
> > +                       u64 bcount = min_t(u64, dmr->start - preve, MAX_KLM_SIZE);
> > +
> >                         klm->key = cpu_to_be32(mvdev->res.null_mkey);
> > -                       klm->bcount = cpu_to_be32(klm_bcount(dmr->start - preve));
> > -                       preve = dmr->start;
> > +                       klm->bcount = cpu_to_be32(klm_bcount(bcount));
> > +                       preve += bcount;
> > +
> >                         goto again;
> >                 }
> >         }
> > --
> > 2.43.0
> >
>
diff mbox series

Patch

diff --git a/drivers/vdpa/mlx5/core/mr.c b/drivers/vdpa/mlx5/core/mr.c
index 8455f08f5d40..61424342c096 100644
--- a/drivers/vdpa/mlx5/core/mr.c
+++ b/drivers/vdpa/mlx5/core/mr.c
@@ -190,9 +190,12 @@  static void fill_indir(struct mlx5_vdpa_dev *mvdev, struct mlx5_vdpa_mr *mkey, v
 			klm->bcount = cpu_to_be32(klm_bcount(dmr->end - dmr->start));
 			preve = dmr->end;
 		} else {
+			u64 bcount = min_t(u64, dmr->start - preve, MAX_KLM_SIZE);
+
 			klm->key = cpu_to_be32(mvdev->res.null_mkey);
-			klm->bcount = cpu_to_be32(klm_bcount(dmr->start - preve));
-			preve = dmr->start;
+			klm->bcount = cpu_to_be32(klm_bcount(bcount));
+			preve += bcount;
+
 			goto again;
 		}
 	}