On Tue, Mar 10, 2026 at 5:07 AM Joshua Daley <[email protected]> wrote:
>
> On 3/8/2026 11:06 PM, Stefan Hajnoczi wrote:
> > On Thu, Feb 26, 2026 at 09:43:45PM +0100, Joshua Daley wrote:
> >> The last step of virtscsi_handle_event is to call virtscsi_kick_event,
> >> which calls INIT_WORK on it's own work item. INIT_WORK resets the
> >> work item's data bits to 0.
> >>
> >> If this occurs while the work item is being flushed by
> >> cancel_work_sync, then kernel/workqueue.c/work_offqd_enable triggers a
> >> kernel warning, as it expects the "disable" bit to be 1:
> >>
> >> [   21.450115] workqueue: work disable count underflowed
> >> [   21.450117] WARNING: CPU: 1 PID: 56 at kernel/workqueue.c:4328 
> >> enable_work+0x10a/0x120
> >> ...
> >> [   21.450171] Call Trace:
> >> [   21.450173]  [<000003db2e5bdc3e>] enable_work+0x10e/0x120
> >> [   21.450176] ([<000003db2e5bdc3a>] enable_work+0x10a/0x120)
> >> [   21.450178]  [<000003db2e5bdd86>] cancel_work_sync+0x86/0xa0
> >> [   21.450181]  [<000003daae97d9e4>] virtscsi_remove+0xb4/0xd0 
> >> [virtio_scsi]
> >> [   21.450184]  [<000003db2ef3b5ca>] virtio_dev_remove+0x6a/0xd0
> >> [   21.450186]  [<000003db2ef9106c>] 
> >> device_release_driver_internal+0x1ac/0x260
> >> [   21.450190]  [<000003db2ef8edc8>] bus_remove_device+0xf8/0x190
> >> [   21.450192]  [<000003db2ef88d72>] device_del+0x142/0x340
> >> [   21.450194]  [<000003db2ef88fa0>] device_unregister+0x30/0xa0
> >> [   21.450196]  [<000003db2ef3b2fa>] unregister_virtio_device+0x2a/0x40
> >>
> >> This warning may occur if a controller is detached immediately
> >> following a disk detach.
> >>
> >> Move the INIT_WORK call to prevent this. Don't re-init event list
> >> work items in virtscsi_kick_event, init them only once in
> >> virtscsi_init instead.
> >>
> >> Signed-off-by: Joshua Daley <[email protected]>
> >> ---
> >>   drivers/scsi/virtio_scsi.c | 6 +++++-
> >>   1 file changed, 5 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
> >> index 0ed8558dad72..173092931df6 100644
> >> --- a/drivers/scsi/virtio_scsi.c
> >> +++ b/drivers/scsi/virtio_scsi.c
> >> @@ -242,7 +242,6 @@ static int virtscsi_kick_event(struct virtio_scsi 
> >> *vscsi,
> >>      struct scatterlist sg;
> >>      unsigned long flags;
> >>
> >> -    INIT_WORK(&event_node->work, virtscsi_handle_event);
> >>      sg_init_one(&sg, event_node->event, sizeof(struct virtio_scsi_event));
> >>
> >>      spin_lock_irqsave(&vscsi->event_vq.vq_lock, flags);
> >> @@ -898,6 +897,11 @@ static int virtscsi_init(struct virtio_device *vdev,
> >>      virtscsi_config_set(vdev, cdb_size, VIRTIO_SCSI_CDB_SIZE);
> >>      virtscsi_config_set(vdev, sense_size, VIRTIO_SCSI_SENSE_SIZE);
> >>
> >> +    if (virtio_has_feature(vdev, VIRTIO_SCSI_F_HOTPLUG)) {
> >> +            for (i = 0; i < VIRTIO_SCSI_EVENT_LEN; i++)
> >> +                    INIT_WORK(&vscsi->event_list[i].work, 
> >> virtscsi_handle_event);
> >> +    }
> >
> > The eventq should be populated unconditionally so that non-hotplug
> > events are processed even when F_HOTPLUG is not negotiated. For example,
> > LUN capacity changes are reported via the VIRTIO_SCSI_T_PARAM_CHANGE
> > event. LUN capacity changes depend on F_CHANGE, not F_HOTPLUG.
> >
> > There is a related bug here: the other if (virtio_has_feature(vdev,
> > VIRTIO_SCSI_F_HOTPLUG)) conditionals in this file need to be revisited
> > so that LUN capacity changes are reported even when F_HOTPLUG is not
> > negotiated. You can test this bug with QEMU's -device
> > virtio-scsi-pci,hotplug=off parameter and the 'block_resize' QEMU
> > monitor command.
> >
> > Do you want to write a patch or do you want me to send a follow-up?
> >
> > Thanks,
> > Stefan
>
> I can write a patch. Thanks for your review.
>
> There are 3 other if (virtio_has_feature(vdev, VIRTIO_SCSI_F_HOTPLUG)) 
> conditionals in this file, for:
>
> 1. virtscsi_kick_event_all() called in virtscsi_probe()
> 2. virtscsi_cancel_event_work() called in virtscsi_remove()
> 3. virtscsi_kick_event_all() called in virtscsi_restore()
>
> Should the eventq be populated truly unconditionally? Then I would just 
> remove the conditions from
> these calls. Or would it be better to just change the conditions to also 
> check the other feature:
> if (... || virtio_has_feature(vdev, VIRTIO_SCSI_F_CHANGE))

Yes. The eventq is always present and the memory backing it is already
allocated. I think dropping the conditionals simplifies the driver and
avoids bugs like the F_CHANGE one without introducing downsides.

Stefan

Reply via email to