When multiple threads are performing USB transmission, musb->lock will be
unlocked when musb_giveback is executed. At this time, qh may be released
in the dequeue process in other threads, resulting in a wild pointer, so
it needs to be here get qh again, and judge whether qh is NULL, and when
dequeue, you need to set qh to NULL.
Fixes: dbac5d07d13e ("usb: musb: host: don't start next rx urb if current one failed")
Signed-off-by: Xingxing Luo <[email protected]>
---
drivers/usb/musb/musb_host.c | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/drivers/usb/musb/musb_host.c b/drivers/usb/musb/musb_host.c
index a02c29216955..9df27db5847a 100644
--- a/drivers/usb/musb/musb_host.c
+++ b/drivers/usb/musb/musb_host.c
@@ -321,10 +321,16 @@ static void musb_advance_schedule(struct musb *musb, struct urb *urb,
musb_giveback(musb, urb, status);
qh->is_ready = ready;
+ /*
+ * musb->lock had been unlocked in musb_giveback, so somtimes qh
+ * may freed, need get it again
+ */
+ qh = musb_ep_get_qh(hw_ep, is_in);
+
/* reclaim resources (and bandwidth) ASAP; deschedule it, and
* invalidate qh as soon as list_empty(&hep->urb_list)
*/
- if (list_empty(&qh->hep->urb_list)) {
+ if (qh != NULL && list_empty(&qh->hep->urb_list)) {
struct list_head *head;
struct dma_controller *dma = musb->dma_controller;
@@ -2398,6 +2404,7 @@ static int musb_urb_dequeue(struct usb_hcd *hcd, struct urb *urb, int status)
* and its URB list has emptied, recycle this qh.
*/
if (ready && list_empty(&qh->hep->urb_list)) {
+ musb_ep_set_qh(qh->hw_ep, is_in, NULL);
qh->hep->hcpriv = NULL;
list_del(&qh->ring);
kfree(qh);
--
2.17.1
Hello!
On 9/14/23 4:56 AM, Xingxing Luo wrote:
> When multiple threads are performing USB transmission, musb->lock will be
> unlocked when musb_giveback is executed. At this time, qh may be released
> in the dequeue process in other threads, resulting in a wild pointer, so
> it needs to be here get qh again, and judge whether qh is NULL, and when
> dequeue, you need to set qh to NULL.
>
> Fixes: dbac5d07d13e ("usb: musb: host: don't start next rx urb if current one failed")
> Signed-off-by: Xingxing Luo <[email protected]>
> ---
> drivers/usb/musb/musb_host.c | 9 ++++++++-
> 1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/musb/musb_host.c b/drivers/usb/musb/musb_host.c
> index a02c29216955..9df27db5847a 100644
> --- a/drivers/usb/musb/musb_host.c
> +++ b/drivers/usb/musb/musb_host.c
> @@ -321,10 +321,16 @@ static void musb_advance_schedule(struct musb *musb, struct urb *urb,
> musb_giveback(musb, urb, status);
> qh->is_ready = ready;
>
> + /*
> + * musb->lock had been unlocked in musb_giveback, so somtimes qh
Sometimes?
> + * may freed, need get it again
> + */
> + qh = musb_ep_get_qh(hw_ep, is_in);
> +
> /* reclaim resources (and bandwidth) ASAP; deschedule it, and
> * invalidate qh as soon as list_empty(&hep->urb_list)
> */
> - if (list_empty(&qh->hep->urb_list)) {
> + if (qh != NULL && list_empty(&qh->hep->urb_list)) {
Just qh, perhaps?
[...]
MBR, Sergey
On 9/14/23 1:06 PM, Sergey Shtylyov wrote:
[...]
>> When multiple threads are performing USB transmission, musb->lock will be
>> unlocked when musb_giveback is executed. At this time, qh may be released
>> in the dequeue process in other threads, resulting in a wild pointer, so
>> it needs to be here get qh again, and judge whether qh is NULL, and when
>> dequeue, you need to set qh to NULL.
>>
>> Fixes: dbac5d07d13e ("usb: musb: host: don't start next rx urb if current one failed")
>> Signed-off-by: Xingxing Luo <[email protected]>
>> ---
>> drivers/usb/musb/musb_host.c | 9 ++++++++-
>> 1 file changed, 8 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/usb/musb/musb_host.c b/drivers/usb/musb/musb_host.c
>> index a02c29216955..9df27db5847a 100644
>> --- a/drivers/usb/musb/musb_host.c
>> +++ b/drivers/usb/musb/musb_host.c
>> @@ -321,10 +321,16 @@ static void musb_advance_schedule(struct musb *musb, struct urb *urb,
>> musb_giveback(musb, urb, status);
>> qh->is_ready = ready;
>>
>> + /*
>> + * musb->lock had been unlocked in musb_giveback, so somtimes qh
>
> Sometimes?
>
>> + * may freed, need get it again
+ * may be freed, need to get it again
Overlooked it in the 1st review, sorry...
[...]
MBR, Sergey
Add more.
On Mon, Sep 18, 2023 at 1:22 PM xingxing luo <[email protected]> wrote:
>
> On Fri, Sep 15, 2023 at 4:48 PM Sergey Shtylyov <[email protected]> wrote:
> >
> > On 9/15/23 5:59 AM, xingxing luo wrote:
> > [...]
> >
> > >>> When multiple threads are performing USB transmission, musb->lock will be
> > >>> unlocked when musb_giveback is executed. At this time, qh may be released
> > >>> in the dequeue process in other threads, resulting in a wild pointer, so
> > >>> it needs to be here get qh again, and judge whether qh is NULL, and when
> > >>> dequeue, you need to set qh to NULL.
> > >>>
> > >>> Fixes: dbac5d07d13e ("usb: musb: host: don't start next rx urb if current one failed")
> > >>> Signed-off-by: Xingxing Luo <[email protected]>
> > >>> ---
> > >>> drivers/usb/musb/musb_host.c | 9 ++++++++-
> > >>> 1 file changed, 8 insertions(+), 1 deletion(-)
> > >>>
> > >>> diff --git a/drivers/usb/musb/musb_host.c b/drivers/usb/musb/musb_host.c
> > >>> index a02c29216955..9df27db5847a 100644
> > >>> --- a/drivers/usb/musb/musb_host.c
> > >>> +++ b/drivers/usb/musb/musb_host.c
> > >>> @@ -321,10 +321,16 @@ static void musb_advance_schedule(struct musb *musb, struct urb *urb,
> > >>> musb_giveback(musb, urb, status);
> > >>> qh->is_ready = ready;
> > >>>
> > >>> + /*
> > >>> + * musb->lock had been unlocked in musb_giveback, so somtimes qh
> > >>
> > >> Sometimes?
> >
> > You have a typo...
> >
> > >>
> > >>> + * may freed, need get it again
> > >>> + */
> > >>> + qh = musb_ep_get_qh(hw_ep, is_in);
> > >>> +
> > >>> /* reclaim resources (and bandwidth) ASAP; deschedule it, and
> > >>> * invalidate qh as soon as list_empty(&hep->urb_list)
> > >>> */
> > >>> - if (list_empty(&qh->hep->urb_list)) {
> > >>> + if (qh != NULL && list_empty(&qh->hep->urb_list)) {
> > >>
> > >> Just qh, perhaps?
> > >
> > > Could you elaborate a little more?
> > > Thanks.
> >
> > Just 'qh' gives you the same as 'qh != NULL'.
>
> Ok, I will address this in the next version.
>
> >
> > [...]
> >
> > MBR, Sergey
B.R Xingxing.Luo