| Message ID | 20220908173618.155291-2-hdegoede@redhat.com |
|---|---|
| State | New |
| Headers | show |
| Series | rt2x00 WA for mac80211 core (BUG?) submitting skb-s with queue-ids >= hw.queues | expand |
On Thu, Sep 08, 2022 at 07:36:18PM +0200, Hans de Goede wrote: > Even though ieee80211_hw.queues is set to 2, the ralink rt2x00 driver > is seeing tx skbs submitted to it with the queue-id set to 2 / set to > IEEE80211_AC_BE on a rt2500 card when associating with an access-point. I'm impressed you have still working rt2500 card :-) > This causes rt2x00queue_get_tx_queue() to return NULL and the following > error to be logged: "ieee80211 phy0: rt2x00mac_tx: Error - Attempt to > send packet over invalid queue 2", after which association with the AP > fails. > > This patch works around this by mapping QID_AC_BE and QID_AC_BK > to QID_AC_VI when there are only 2 tx_queues. > > Signed-off-by: Hans de Goede <hdegoede@redhat.com> Acked-by: Stanislaw Gruszka <stf_xl@wp.pl>
Hi, On 9/15/22 20:56, Stanislaw Gruszka wrote: > On Thu, Sep 08, 2022 at 07:36:18PM +0200, Hans de Goede wrote: >> Even though ieee80211_hw.queues is set to 2, the ralink rt2x00 driver >> is seeing tx skbs submitted to it with the queue-id set to 2 / set to >> IEEE80211_AC_BE on a rt2500 card when associating with an access-point. > > I'm impressed you have still working rt2500 card :-) I recently become the owner of a Turion64 based MSI S270 laptop to test some kernel code I was refactoring on. This has a mini-pci rt2500 card, which indeed still works after this fix. The thing has 2G RAM and even runs a Fedora 37 beta but the CPU is quite slow and gets pretty toasty. >> This causes rt2x00queue_get_tx_queue() to return NULL and the following >> error to be logged: "ieee80211 phy0: rt2x00mac_tx: Error - Attempt to >> send packet over invalid queue 2", after which association with the AP >> fails. >> >> This patch works around this by mapping QID_AC_BE and QID_AC_BK >> to QID_AC_VI when there are only 2 tx_queues. >> >> Signed-off-by: Hans de Goede <hdegoede@redhat.com> > > Acked-by: Stanislaw Gruszka <stf_xl@wp.pl> Thanks. Regards, Hans
diff --git a/drivers/net/wireless/ralink/rt2x00/rt2x00.h b/drivers/net/wireless/ralink/rt2x00/rt2x00.h index 8f5772b98f58..07a6a5a9ce13 100644 --- a/drivers/net/wireless/ralink/rt2x00/rt2x00.h +++ b/drivers/net/wireless/ralink/rt2x00/rt2x00.h @@ -1309,8 +1309,11 @@ void rt2x00queue_unmap_skb(struct queue_entry *entry); */ static inline struct data_queue * rt2x00queue_get_tx_queue(struct rt2x00_dev *rt2x00dev, - const enum data_queue_qid queue) + enum data_queue_qid queue) { + if (queue >= rt2x00dev->ops->tx_queues && queue < IEEE80211_NUM_ACS) + queue = rt2x00dev->ops->tx_queues - 1; + if (queue < rt2x00dev->ops->tx_queues && rt2x00dev->tx) return &rt2x00dev->tx[queue];
Even though ieee80211_hw.queues is set to 2, the ralink rt2x00 driver is seeing tx skbs submitted to it with the queue-id set to 2 / set to IEEE80211_AC_BE on a rt2500 card when associating with an access-point. This causes rt2x00queue_get_tx_queue() to return NULL and the following error to be logged: "ieee80211 phy0: rt2x00mac_tx: Error - Attempt to send packet over invalid queue 2", after which association with the AP fails. This patch works around this by mapping QID_AC_BE and QID_AC_BK to QID_AC_VI when there are only 2 tx_queues. Signed-off-by: Hans de Goede <hdegoede@redhat.com> --- The net/mac80211 code has checks for local->hw.queues < IEEE80211_NUM_ACS in various places and returns queue-id 0 when this condition is true. So it looks like the rt2x00 driver receiving a queue-id of 2 / IEEE80211_AC_BE might actually be a bug in the mac80211 core code, I will send out a separate email about this. --- drivers/net/wireless/ralink/rt2x00/rt2x00.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)