RE: [PATCH] virtio_pci: Wait for legacy device to be reset

From: Angus Chen
Date: Tue Apr 11 2023 - 02:39:27 EST


Hi mst.

> -----Original Message-----
> From: Michael S. Tsirkin <mst@xxxxxxxxxx>
> Sent: Tuesday, April 11, 2023 2:30 PM
> To: Angus Chen <angus.chen@xxxxxxxxxxxxxxx>
> Cc: jasowang@xxxxxxxxxx; virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH] virtio_pci: Wait for legacy device to be reset
>
> On Tue, Apr 11, 2023 at 09:38:32AM +0800, Angus Chen wrote:
> > We read the status of device after reset,
> > It is not guaranteed that the device be reseted successfully.
>
> Sorry not guaranteed by what? I am guessing you have a legacy device
> that does not reset fully on write, and you need to wait?
When the card not finished reset, the read only return the middle state of card.
>
> > We can use a while loop to make sure that,like the modern device did.
> > The spec is not request it ,but it work.
> >
> > Signed-off-by: Angus Chen <angus.chen@xxxxxxxxxxxxxxx>
>
> Generally I don't much like touching legacy, no telling what
> that will do. Case in point, is your device a pure
> legacy device or a transitional device?
Yes.,we have a real card which is use vitio spec.
Thank you.
>
> > ---
> > drivers/virtio/virtio_pci_legacy.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/virtio/virtio_pci_legacy.c b/drivers/virtio/virtio_pci_legacy.c
> > index 2257f1b3d8ae..f2d241563e4f 100644
> > --- a/drivers/virtio/virtio_pci_legacy.c
> > +++ b/drivers/virtio/virtio_pci_legacy.c
> > @@ -14,6 +14,7 @@
> > * Michael S. Tsirkin <mst@xxxxxxxxxx>
> > */
> >
> > +#include <linux/delay.h>
> > #include "linux/virtio_pci_legacy.h"
> > #include "virtio_pci_common.h"
> >
> > @@ -97,7 +98,8 @@ static void vp_reset(struct virtio_device *vdev)
> > vp_legacy_set_status(&vp_dev->ldev, 0);
> > /* Flush out the status write, and flush in device writes,
> > * including MSi-X interrupts, if any. */
> > - vp_legacy_get_status(&vp_dev->ldev);
> > + while (vp_legacy_get_status(&vp_dev->ldev))
> > + msleep(1);
>
> The problem with this is that it will break surprise
> removal even worse than it's already broken.
>
>
> > /* Flush pending VQ/configuration callbacks. */
> > vp_synchronize_vectors(vdev);
> > }
> > --
> > 2.25.1