2024-01-17 19:23:43

by Steven Sistare

Subject: [PATCH V1] vdpa_sim: reset must not run

vdpasim_do_reset sets running to true, which is wrong, as it allows
vdpasim_kick_vq to post work requests before the device has been
configured. To fix, do not set running until VIRTIO_CONFIG_S_FEATURES_OK
is set.

Fixes: 0c89e2a3a9d0 ("vdpa_sim: Implement suspend vdpa op")
Signed-off-by: Steve Sistare <[email protected]>
Reviewed-by: Eugenio Pérez <[email protected]>
---
drivers/vdpa/vdpa_sim/vdpa_sim.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
index be2925d0d283..6304cb0b4770 100644
--- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
+++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
@@ -160,7 +160,7 @@ static void vdpasim_do_reset(struct vdpasim *vdpasim, u32 flags)
}
}

- vdpasim->running = true;
+ vdpasim->running = false;
spin_unlock(&vdpasim->iommu_lock);

vdpasim->features = 0;
@@ -483,6 +483,7 @@ static void vdpasim_set_status(struct vdpa_device *vdpa, u8 status)

mutex_lock(&vdpasim->mutex);
vdpasim->status = status;
+ vdpasim->running = (status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
mutex_unlock(&vdpasim->mutex);
}

--
2.39.3

2024-01-22 04:13:19

by Jason Wang

Subject: Re: [PATCH V1] vdpa_sim: reset must not run

On Thu, Jan 18, 2024 at 3:23 AM Steve Sistare <[email protected]> wrote:
>
> vdpasim_do_reset sets running to true, which is wrong, as it allows
> vdpasim_kick_vq to post work requests before the device has been
> configured. To fix, do not set running until VIRTIO_CONFIG_S_FEATURES_OK
> is set.
>
> Fixes: 0c89e2a3a9d0 ("vdpa_sim: Implement suspend vdpa op")
> Signed-off-by: Steve Sistare <[email protected]>
> Reviewed-by: Eugenio Pérez <[email protected]>

Acked-by: Jason Wang <[email protected]>

Thanks

2024-01-22 10:22:10

by Stefano Garzarella

Subject: Re: [PATCH V1] vdpa_sim: reset must not run

On Wed, Jan 17, 2024 at 11:23:23AM -0800, Steve Sistare wrote:
>vdpasim_do_reset sets running to true, which is wrong, as it allows
>vdpasim_kick_vq to post work requests before the device has been
>configured. To fix, do not set running until VIRTIO_CONFIG_S_FEATURES_OK
>is set.
>
>Fixes: 0c89e2a3a9d0 ("vdpa_sim: Implement suspend vdpa op")
>Signed-off-by: Steve Sistare <[email protected]>
>Reviewed-by: Eugenio P?rez <[email protected]>
>---
> drivers/vdpa/vdpa_sim/vdpa_sim.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
>diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>index be2925d0d283..6304cb0b4770 100644
>--- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
>+++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>@@ -160,7 +160,7 @@ static void vdpasim_do_reset(struct vdpasim *vdpasim, u32 flags)
> }
> }
>
>- vdpasim->running = true;
>+ vdpasim->running = false;
> spin_unlock(&vdpasim->iommu_lock);
>
> vdpasim->features = 0;
>@@ -483,6 +483,7 @@ static void vdpasim_set_status(struct vdpa_device *vdpa, u8 status)
>
> mutex_lock(&vdpasim->mutex);
> vdpasim->status = status;
>+ vdpasim->running = (status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
> mutex_unlock(&vdpasim->mutex);

Should we do something similar also in vdpasim_resume() ?

I mean something like this:

diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
index be2925d0d283..55e4633d5442 100644
--- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
+++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
@@ -520,7 +520,7 @@ static int vdpasim_resume(struct vdpa_device *vdpa)
int i;

mutex_lock(&vdpasim->mutex);
- vdpasim->running = true;
+ vdpasim->running = (vdpasim->status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;

if (vdpasim->pending_kick) {
/* Process pending descriptors */

Thanks,
Stefano

2024-01-22 10:48:15

by Eugenio Perez Martin

Subject: Re: [PATCH V1] vdpa_sim: reset must not run

On Mon, Jan 22, 2024 at 11:22 AM Stefano Garzarella <[email protected]> wrote:
>
> On Wed, Jan 17, 2024 at 11:23:23AM -0800, Steve Sistare wrote:
> >vdpasim_do_reset sets running to true, which is wrong, as it allows
> >vdpasim_kick_vq to post work requests before the device has been
> >configured. To fix, do not set running until VIRTIO_CONFIG_S_FEATURES_OK
> >is set.
> >
> >Fixes: 0c89e2a3a9d0 ("vdpa_sim: Implement suspend vdpa op")
> >Signed-off-by: Steve Sistare <[email protected]>
> >Reviewed-by: Eugenio Pérez <[email protected]>
> >---
> > drivers/vdpa/vdpa_sim/vdpa_sim.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> >diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >index be2925d0d283..6304cb0b4770 100644
> >--- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >+++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >@@ -160,7 +160,7 @@ static void vdpasim_do_reset(struct vdpasim *vdpasim, u32 flags)
> > }
> > }
> >
> >- vdpasim->running = true;
> >+ vdpasim->running = false;
> > spin_unlock(&vdpasim->iommu_lock);
> >
> > vdpasim->features = 0;
> >@@ -483,6 +483,7 @@ static void vdpasim_set_status(struct vdpa_device *vdpa, u8 status)
> >
> > mutex_lock(&vdpasim->mutex);
> > vdpasim->status = status;
> >+ vdpasim->running = (status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
> > mutex_unlock(&vdpasim->mutex);
>
> Should we do something similar also in vdpasim_resume() ?
>
> I mean something like this:
>
> diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> index be2925d0d283..55e4633d5442 100644
> --- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
> +++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> @@ -520,7 +520,7 @@ static int vdpasim_resume(struct vdpa_device *vdpa)
> int i;
>
> mutex_lock(&vdpasim->mutex);
> - vdpasim->running = true;
> + vdpasim->running = (vdpasim->status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
>
> if (vdpasim->pending_kick) {
> /* Process pending descriptors */
>
> Thanks,
> Stefano
>

The suspend and resume operation should not be called before
DRIVER_OK, so maybe we should add that protection at
drivers/vhost/vdpa.c actually?

Thanks!

2024-01-22 10:59:49

by Stefano Garzarella

Subject: Re: Re: [PATCH V1] vdpa_sim: reset must not run

On Mon, Jan 22, 2024 at 11:47:22AM +0100, Eugenio Perez Martin wrote:
>On Mon, Jan 22, 2024 at 11:22 AM Stefano Garzarella <[email protected]> wrote:
>>
>> On Wed, Jan 17, 2024 at 11:23:23AM -0800, Steve Sistare wrote:
>> >vdpasim_do_reset sets running to true, which is wrong, as it allows
>> >vdpasim_kick_vq to post work requests before the device has been
>> >configured. To fix, do not set running until VIRTIO_CONFIG_S_FEATURES_OK
>> >is set.
>> >
>> >Fixes: 0c89e2a3a9d0 ("vdpa_sim: Implement suspend vdpa op")
>> >Signed-off-by: Steve Sistare <[email protected]>
>> >Reviewed-by: Eugenio Pérez <[email protected]>
>> >---
>> > drivers/vdpa/vdpa_sim/vdpa_sim.c | 3 ++-
>> > 1 file changed, 2 insertions(+), 1 deletion(-)
>> >
>> >diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>> >index be2925d0d283..6304cb0b4770 100644
>> >--- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
>> >+++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>> >@@ -160,7 +160,7 @@ static void vdpasim_do_reset(struct vdpasim *vdpasim, u32 flags)
>> > }
>> > }
>> >
>> >- vdpasim->running = true;
>> >+ vdpasim->running = false;
>> > spin_unlock(&vdpasim->iommu_lock);
>> >
>> > vdpasim->features = 0;
>> >@@ -483,6 +483,7 @@ static void vdpasim_set_status(struct vdpa_device *vdpa, u8 status)
>> >
>> > mutex_lock(&vdpasim->mutex);
>> > vdpasim->status = status;
>> >+ vdpasim->running = (status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
>> > mutex_unlock(&vdpasim->mutex);
>>
>> Should we do something similar also in vdpasim_resume() ?
>>
>> I mean something like this:
>>
>> diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>> index be2925d0d283..55e4633d5442 100644
>> --- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
>> +++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>> @@ -520,7 +520,7 @@ static int vdpasim_resume(struct vdpa_device *vdpa)
>> int i;
>>
>> mutex_lock(&vdpasim->mutex);
>> - vdpasim->running = true;
>> + vdpasim->running = (vdpasim->status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
>>
>> if (vdpasim->pending_kick) {
>> /* Process pending descriptors */
>>
>> Thanks,
>> Stefano
>>
>
>The suspend and resume operation should not be called before
>DRIVER_OK, so maybe we should add that protection at
>drivers/vhost/vdpa.c actually?

Yeah, I think so!

Anyway, IMHO we should at least return an error in vdpa_sim if
vdpasim_suspend/resume are called before DRIVER_OK (in another patch of
course).

Stefano

2024-02-09 22:37:13

by Steven Sistare

Subject: Re: [PATCH V1] vdpa_sim: reset must not run

On 1/22/2024 5:59 AM, Stefano Garzarella wrote:
> On Mon, Jan 22, 2024 at 11:47:22AM +0100, Eugenio Perez Martin wrote:
>> On Mon, Jan 22, 2024 at 11:22 AM Stefano Garzarella <[email protected]> wrote:
>>>
>>> On Wed, Jan 17, 2024 at 11:23:23AM -0800, Steve Sistare wrote:
>>> >vdpasim_do_reset sets running to true, which is wrong, as it allows
>>> >vdpasim_kick_vq to post work requests before the device has been
>>> >configured.  To fix, do not set running until VIRTIO_CONFIG_S_FEATURES_OK
>>> >is set.
>>> >
>>> >Fixes: 0c89e2a3a9d0 ("vdpa_sim: Implement suspend vdpa op")
>>> >Signed-off-by: Steve Sistare <[email protected]>
>>> >Reviewed-by: Eugenio Pérez <[email protected]>
>>> >---
>>> > drivers/vdpa/vdpa_sim/vdpa_sim.c | 3 ++-
>>> > 1 file changed, 2 insertions(+), 1 deletion(-)
>>> >
>>> >diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>>> >index be2925d0d283..6304cb0b4770 100644
>>> >--- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
>>> >+++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>>> >@@ -160,7 +160,7 @@ static void vdpasim_do_reset(struct vdpasim *vdpasim, u32 flags)
>>> >               }
>>> >       }
>>> >
>>> >-      vdpasim->running = true;
>>> >+      vdpasim->running = false;
>>> >       spin_unlock(&vdpasim->iommu_lock);
>>> >
>>> >       vdpasim->features = 0;
>>> >@@ -483,6 +483,7 @@ static void vdpasim_set_status(struct vdpa_device *vdpa, u8 status)
>>> >
>>> >       mutex_lock(&vdpasim->mutex);
>>> >       vdpasim->status = status;
>>> >+      vdpasim->running = (status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
>>> >       mutex_unlock(&vdpasim->mutex);
>>>
>>> Should we do something similar also in vdpasim_resume() ?
>>>
>>> I mean something like this:
>>>
>>> diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>>> index be2925d0d283..55e4633d5442 100644
>>> --- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
>>> +++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
>>> @@ -520,7 +520,7 @@ static int vdpasim_resume(struct vdpa_device *vdpa)
>>>          int i;
>>>
>>>          mutex_lock(&vdpasim->mutex);
>>> -       vdpasim->running = true;
>>> +       vdpasim->running = (vdpasim->status & VIRTIO_CONFIG_S_FEATURES_OK) != 0;
>>>
>>>          if (vdpasim->pending_kick) {
>>>                  /* Process pending descriptors */
>>>
>>> Thanks,
>>> Stefano
>>>
>>
>> The suspend and resume operation should not be called before
>> DRIVER_OK, so maybe we should add that protection at
>> drivers/vhost/vdpa.c actually?
>
> Yeah, I think so!
>
> Anyway, IMHO we should at least return an error in vdpa_sim if vdpasim_suspend/resume are called before DRIVER_OK (in another patch of course).

I submitted "vdpa: suspend and resume require DRIVER_OK" to check this in vdpa.c so there
is no need to check it in the leaf drivers.

I also submitted V2 of this patch, "vdpa_sim: reset must not run".
It checks for DRIVER_OK, instead of FEATURES_OK.

- Steve