Hi Mathieu, On 9/30/22 09:50, Shengjiu Wang wrote: > The workqueue may execute late even after remoteproc is stopped or > stopping, some resources (rpmsg device and endpoint) have been > released in rproc_stop_subdevices(), then rproc_vq_interrupt() > accessing these resources will cause kennel dump. > > Call trace: > virtqueue_add_split+0x1ac/0x560 > virtqueue_add_inbuf+0x4c/0x60 > rpmsg_recv_done+0x15c/0x294 > vring_interrupt+0x6c/0xa4 > rproc_vq_interrupt+0x30/0x50 > imx_dsp_rproc_vq_work+0x24/0x40 [imx_dsp_rproc] > process_one_work+0x1d0/0x354 > worker_thread+0x13c/0x470 > kthread+0x154/0x160 > ret_from_fork+0x10/0x20 > > Add mutex protection in imx_dsp_rproc_vq_work(), if the state is > not running, then just skip calling rproc_vq_interrupt(). > > Also the flush workqueue operation can't be added in rproc stop > for the same reason. The call sequence is > > rproc_shutdown > -> rproc_stop > ->rproc_stop_subdevices > ->rproc->ops->stop() > ->imx_dsp_rproc_stop > ->flush_work > -> rproc_vq_interrupt > > The resource needed by rproc_vq_interrupt has been released in > rproc_stop_subdevices, so flush_work is not safe to be called in > imx_dsp_rproc_stop. > > Fixes: ec0e5549f358 ("remoteproc: imx_dsp_rproc: Add remoteproc driver for DSP on i.MX") > Signed-off-by: Shengjiu Wang <shengjiu.wang@xxxxxxx> > Reviewed-by: Peng Fan <peng.fan@xxxxxxx> > --- > changes in v2: > - update commit message > > drivers/remoteproc/imx_dsp_rproc.c | 12 +++++++++--- > 1 file changed, 9 insertions(+), 3 deletions(-) > > diff --git a/drivers/remoteproc/imx_dsp_rproc.c b/drivers/remoteproc/imx_dsp_rproc.c > index 899aa8dd12f0..95da1cbefacf 100644 > --- a/drivers/remoteproc/imx_dsp_rproc.c > +++ b/drivers/remoteproc/imx_dsp_rproc.c > @@ -347,9 +347,6 @@ static int imx_dsp_rproc_stop(struct rproc *rproc) > struct device *dev = rproc->dev.parent; > int ret = 0; > > - /* Make sure work is finished */ > - flush_work(&priv->rproc_work); > - > if (rproc->state == RPROC_CRASHED) { > priv->flags &= ~REMOTE_IS_READY; > return 0; > @@ -432,9 +429,18 @@ static void imx_dsp_rproc_vq_work(struct work_struct *work) > { > struct imx_dsp_rproc *priv = container_of(work, struct imx_dsp_rproc, > rproc_work); > + struct rproc *rproc = priv->rproc; > + > + mutex_lock(&rproc->lock); > + > + if (rproc->state != RPROC_RUNNING) > + goto unlock_mutex; > > rproc_vq_interrupt(priv->rproc, 0); > rproc_vq_interrupt(priv->rproc, 1); > + > +unlock_mutex: > + mutex_unlock(&rproc->lock); > } I guess that this fix the issue in case of the use of a work queue for the notification. As I reproduced the issue on stm32 only one time without fix, difficult for me to test the fix... I also have a look on some other drivers It seems not applicable for some other drivers such as TI k3 R5[1] that call virtio under interrupt context ( if I have well understood). An alternative to try to fix that in a more generic way could be to call rproc_remove_virtio_dev in the subdev->unprepare, instead of subdev->stop. This should postpone the virtio unregister after the stop of the coprocessor. Of course it is only a suggestion/area of improvement, this fix LGTM. [1]https://elixir.bootlin.com/linux/v6.0/source/drivers/remoteproc/ti_k3_r5_remoteproc.c#L180 Regards, Arnaud > > /**