| Message ID | 1426081031-15296-1-git-send-email-michal.kazior@tieto.com (mailing list archive) |
|---|---|
| State | Not Applicable |
| Delegated to: | Kalle Valo |
| Headers | show |
Michal Kazior <michal.kazior@tieto.com> writes: > Some firmware revisions (tested with qca6174 > rm2.0-00088) deliver unsolicited unknown (kind of > garbled) HTC ep0 event to host in some cases. > > This issue was mainly observed with both qca988x > and qca6174 being installed on a single host > system. During driver probing if qca6174 booting > sequences were somehow deferred (e.g. by qca988x > implicitly making some resources busy presumably) > the unsolicited event would came around 1s after > ATH10K_HTC_MSG_READY_ID was delivered to host for > qca6174. > > The unsolicited event would confuse driver and > cause HTT initialization (and subsequently > driver probing) to fail. > > Make the ep0 event processing more robust. The > event will still be caught but instead will > only generate a warning now. > > Signed-off-by: Michal Kazior <michal.kazior@tieto.com> Thanks, applied.
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c index 2fd9e18..d33d5c4 100644 --- a/drivers/net/wireless/ath/ath10k/htc.c +++ b/drivers/net/wireless/ath/ath10k/htc.c @@ -414,7 +414,8 @@ static int ath10k_htc_rx_completion_handler(struct ath10k *ar, struct ath10k_htc_msg *msg = (struct ath10k_htc_msg *)skb->data; switch (__le16_to_cpu(msg->hdr.message_id)) { - default: + case ATH10K_HTC_MSG_READY_ID: + case ATH10K_HTC_MSG_CONNECT_SERVICE_RESP_ID: /* handle HTC control message */ if (completion_done(&htc->ctl_resp)) { /* @@ -438,6 +439,10 @@ static int ath10k_htc_rx_completion_handler(struct ath10k *ar, break; case ATH10K_HTC_MSG_SEND_SUSPEND_COMPLETE: htc->htc_ops.target_send_suspend_complete(ar); + break; + default: + ath10k_warn(ar, "ignoring unsolicited htc ep0 event\n"); + break; } goto out; }
Some firmware revisions (tested with qca6174 rm2.0-00088) deliver unsolicited unknown (kind of garbled) HTC ep0 event to host in some cases. This issue was mainly observed with both qca988x and qca6174 being installed on a single host system. During driver probing if qca6174 booting sequences were somehow deferred (e.g. by qca988x implicitly making some resources busy presumably) the unsolicited event would came around 1s after ATH10K_HTC_MSG_READY_ID was delivered to host for qca6174. The unsolicited event would confuse driver and cause HTT initialization (and subsequently driver probing) to fail. Make the ep0 event processing more robust. The event will still be caught but instead will only generate a warning now. Signed-off-by: Michal Kazior <michal.kazior@tieto.com> --- drivers/net/wireless/ath/ath10k/htc.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)