| Message ID | 1525094357-24782-2-git-send-email-sgruszka@redhat.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 811a3991510735566b66069fdd4ce3ce33a2ec18 |
| Delegated to: | Kalle Valo |
| Headers | show |
Hi On Tue, May 01, 2018 at 06:27:17AM -0600, Kofi Agor wrote: > I haven't had a chance to test your patch yet, but does it resolve the tx > queue flush issues? What platform are you developing/testing on? I'm not sure about the issue, but most likely the patch does not fix it. I tested the patch on RT3062 pcie device an RT3071 usb dongle. Regards Stanislaw
does it fix mt7620 bug? On 02/05/2018, Stanislaw Gruszka <sgruszka@redhat.com> wrote: > Hi > > On Tue, May 01, 2018 at 06:27:17AM -0600, Kofi Agor wrote: >> I haven't had a chance to test your patch yet, but does it resolve the tx >> queue flush issues? What platform are you developing/testing on? > > I'm not sure about the issue, but most likely the patch > does not fix it. I tested the patch on RT3062 pcie device an > RT3071 usb dongle. > > Regards > Stanislaw >
On Wed, Jun 20, 2018 at 06:22:45PM +0200, Tom Psyborg wrote: > does it fix mt7620 bug? It is possible that patch fixes some bug that happen also on mt7620, but is was not that I found a reproducible bug and patch fixed it. Stanislaw > On 02/05/2018, Stanislaw Gruszka <sgruszka@redhat.com> wrote: > > Hi > > > > On Tue, May 01, 2018 at 06:27:17AM -0600, Kofi Agor wrote: > >> I haven't had a chance to test your patch yet, but does it resolve the tx > >> queue flush issues? What platform are you developing/testing on? > > > > I'm not sure about the issue, but most likely the patch > > does not fix it. I tested the patch on RT3062 pcie device an > > RT3071 usb dongle. > > > > Regards > > Stanislaw > >
diff --git a/drivers/net/wireless/ralink/rt2x00/rt2800mmio.c b/drivers/net/wireless/ralink/rt2x00/rt2800mmio.c index 1123e2bed803..e1a7ed7e4892 100644 --- a/drivers/net/wireless/ralink/rt2x00/rt2800mmio.c +++ b/drivers/net/wireless/ralink/rt2x00/rt2800mmio.c @@ -600,6 +600,7 @@ void rt2800mmio_kick_queue(struct data_queue *queue) case QID_AC_VI: case QID_AC_BE: case QID_AC_BK: + WARN_ON_ONCE(rt2x00queue_empty(queue)); entry = rt2x00queue_get_entry(queue, Q_INDEX); rt2x00mmio_register_write(rt2x00dev, TX_CTX_IDX(queue->qid), entry->entry_idx); diff --git a/drivers/net/wireless/ralink/rt2x00/rt2x00mac.c b/drivers/net/wireless/ralink/rt2x00/rt2x00mac.c index a971bc7a6b63..c380c1f56ba6 100644 --- a/drivers/net/wireless/ralink/rt2x00/rt2x00mac.c +++ b/drivers/net/wireless/ralink/rt2x00/rt2x00mac.c @@ -739,8 +739,7 @@ void rt2x00mac_flush(struct ieee80211_hw *hw, struct ieee80211_vif *vif, return; tx_queue_for_each(rt2x00dev, queue) - if (!rt2x00queue_empty(queue)) - rt2x00queue_flush_queue(queue, drop); + rt2x00queue_flush_queue(queue, drop); } EXPORT_SYMBOL_GPL(rt2x00mac_flush); diff --git a/drivers/net/wireless/ralink/rt2x00/rt2x00queue.c b/drivers/net/wireless/ralink/rt2x00/rt2x00queue.c index a6884e73d2ab..7c1f8f561d4a 100644 --- a/drivers/net/wireless/ralink/rt2x00/rt2x00queue.c +++ b/drivers/net/wireless/ralink/rt2x00/rt2x00queue.c @@ -1000,6 +1000,8 @@ void rt2x00queue_flush_queue(struct data_queue *queue, bool drop) (queue->qid == QID_AC_BE) || (queue->qid == QID_AC_BK); + if (rt2x00queue_empty(queue)) + return; /* * If we are not supposed to drop any pending
We have check if queue is not empty when start flushing queues on by mac80211 callback, but we also can start flushing queues by internal driver calls. So move check into rt2x00queue_flush_queue() to assure we do not flush empty queue anytime. Additionally add warning if we start to kick empty queue as in such situation we set wrong index in the HW queue, what can confuse the HW and have various negative consequences. Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> --- drivers/net/wireless/ralink/rt2x00/rt2800mmio.c | 1 + drivers/net/wireless/ralink/rt2x00/rt2x00mac.c | 3 +-- drivers/net/wireless/ralink/rt2x00/rt2x00queue.c | 2 ++ 3 files changed, 4 insertions(+), 2 deletions(-)