2023-02-06 12:20:39

by Eli Cohen

Subject: [PATCH 2/2] vdpa/mlx5: Initialize CVQ iotlb spinlock

Initialize itolb spinlock.

Fixes: 5262912ef3cf ("vdpa/mlx5: Add support for control VQ and MAC setting")
Signed-off-by: Eli Cohen <[email protected]>
---
drivers/vdpa/mlx5/core/resources.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/drivers/vdpa/mlx5/core/resources.c b/drivers/vdpa/mlx5/core/resources.c
index 45ad41287a31..d5a59c9035fb 100644
--- a/drivers/vdpa/mlx5/core/resources.c
+++ b/drivers/vdpa/mlx5/core/resources.c
@@ -233,6 +233,7 @@ static int init_ctrl_vq(struct mlx5_vdpa_dev *mvdev)
if (!mvdev->cvq.iotlb)
return -ENOMEM;

+ spin_lock_init(&mvdev->cvq.iommu_lock);
vringh_set_iotlb(&mvdev->cvq.vring, mvdev->cvq.iotlb, &mvdev->cvq.iommu_lock);

return 0;
--
2.38.1

2023-02-13 12:20:08

by Michael S. Tsirkin

Subject: Re: [PATCH 2/2] vdpa/mlx5: Initialize CVQ iotlb spinlock

On Mon, Feb 06, 2023 at 02:20:16PM +0200, Eli Cohen wrote:
> Initialize itolb spinlock.
>
> Fixes: 5262912ef3cf ("vdpa/mlx5: Add support for control VQ and MAC setting")
> Signed-off-by: Eli Cohen <[email protected]>

threading was broken here but whatevs.

> ---
> drivers/vdpa/mlx5/core/resources.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/vdpa/mlx5/core/resources.c b/drivers/vdpa/mlx5/core/resources.c
> index 45ad41287a31..d5a59c9035fb 100644
> --- a/drivers/vdpa/mlx5/core/resources.c
> +++ b/drivers/vdpa/mlx5/core/resources.c
> @@ -233,6 +233,7 @@ static int init_ctrl_vq(struct mlx5_vdpa_dev *mvdev)
> if (!mvdev->cvq.iotlb)
> return -ENOMEM;
>
> + spin_lock_init(&mvdev->cvq.iommu_lock);
> vringh_set_iotlb(&mvdev->cvq.vring, mvdev->cvq.iotlb, &mvdev->cvq.iommu_lock);
>
> return 0;
> --
> 2.38.1

2023-02-13 12:24:58

by Eli Cohen

Subject: Re: [PATCH 2/2] vdpa/mlx5: Initialize CVQ iotlb spinlock

On 13/02/2023 14:19, Michael S. Tsirkin wrote:
> On Mon, Feb 06, 2023 at 02:20:16PM +0200, Eli Cohen wrote:
>> Initialize itolb spinlock.
>>
>> Fixes: 5262912ef3cf ("vdpa/mlx5: Add support for control VQ and MAC setting")
>> Signed-off-by: Eli Cohen <[email protected]>
> threading was broken here but whatevs.
What exactly do you mean by "threading was broken". Is it because I sent
two unrelated fixes together appearing as patch 1/2 and 2/2?
>
>> ---
>> drivers/vdpa/mlx5/core/resources.c | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/drivers/vdpa/mlx5/core/resources.c b/drivers/vdpa/mlx5/core/resources.c
>> index 45ad41287a31..d5a59c9035fb 100644
>> --- a/drivers/vdpa/mlx5/core/resources.c
>> +++ b/drivers/vdpa/mlx5/core/resources.c
>> @@ -233,6 +233,7 @@ static int init_ctrl_vq(struct mlx5_vdpa_dev *mvdev)
>> if (!mvdev->cvq.iotlb)
>> return -ENOMEM;
>>
>> + spin_lock_init(&mvdev->cvq.iommu_lock);
>> vringh_set_iotlb(&mvdev->cvq.vring, mvdev->cvq.iotlb, &mvdev->cvq.iommu_lock);
>>
>> return 0;
>> --
>> 2.38.1

2023-02-13 12:27:07

by Michael S. Tsirkin

Subject: Re: [PATCH 2/2] vdpa/mlx5: Initialize CVQ iotlb spinlock

On Mon, Feb 13, 2023 at 02:24:40PM +0200, Eli Cohen wrote:
>
> On 13/02/2023 14:19, Michael S. Tsirkin wrote:
> > On Mon, Feb 06, 2023 at 02:20:16PM +0200, Eli Cohen wrote:
> > > Initialize itolb spinlock.
> > >
> > > Fixes: 5262912ef3cf ("vdpa/mlx5: Add support for control VQ and MAC setting")
> > > Signed-off-by: Eli Cohen <[email protected]>
> > threading was broken here but whatevs.
> What exactly do you mean by "threading was broken". Is it because I sent two
> unrelated fixes together appearing as patch 1/2 and 2/2?

exactly. if I see 2/2 I expect a thread so I can find 1/2
and preferably a cover letter explaining what unifies all
these patches.
If they are unrelated no need to number them.

> >
> > > ---
> > > drivers/vdpa/mlx5/core/resources.c | 1 +
> > > 1 file changed, 1 insertion(+)
> > >
> > > diff --git a/drivers/vdpa/mlx5/core/resources.c b/drivers/vdpa/mlx5/core/resources.c
> > > index 45ad41287a31..d5a59c9035fb 100644
> > > --- a/drivers/vdpa/mlx5/core/resources.c
> > > +++ b/drivers/vdpa/mlx5/core/resources.c
> > > @@ -233,6 +233,7 @@ static int init_ctrl_vq(struct mlx5_vdpa_dev *mvdev)
> > > if (!mvdev->cvq.iotlb)
> > > return -ENOMEM;
> > > + spin_lock_init(&mvdev->cvq.iommu_lock);
> > > vringh_set_iotlb(&mvdev->cvq.vring, mvdev->cvq.iotlb, &mvdev->cvq.iommu_lock);
> > > return 0;
> > > --
> > > 2.38.1