| Message ID | 20230809042738.13394-3-quic_hprem@quicinc.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | Johannes Berg |
| Headers | show |
| Series | wifi: ath12k: Ignore fragments from uninitialized peer in dp | expand |
On Wed, 2023-08-09 at 09:57 +0530, Harshitha Prem wrote: > > During the crash, the fragment id of that packet was 0 and > __fls returned a junk value. Hence, it was identified that "__fls" > would have an undefined behavior if the argument is passed as "0". > Therefore, add changes to handle the same. Umm. That makes it sound like you are surprised by this behaviour of __fls() and expected something else?! Please go read the documentation, and then rewrite the commit message. You really shouldn't be surprised by this. johannes
On Wed, 2023-08-09 at 14:17 +0200, Johannes Berg wrote: > On Wed, 2023-08-09 at 09:57 +0530, Harshitha Prem wrote: > > > > During the crash, the fragment id of that packet was 0 and > > __fls returned a junk value. Hence, it was identified that "__fls" > > would have an undefined behavior if the argument is passed as "0". > > Therefore, add changes to handle the same. > > Umm. That makes it sound like you are surprised by this behaviour of > __fls() and expected something else?! Please go read the documentation, > and then rewrite the commit message. You really shouldn't be surprised > by this. > Also, btw, "Fixes:" tags would be nice for fixes. I'm sure Kalle has brought that up a million times before ;) johannes
On 8/9/2023 5:47 PM, Johannes Berg wrote: > On Wed, 2023-08-09 at 09:57 +0530, Harshitha Prem wrote: >> >> During the crash, the fragment id of that packet was 0 and >> __fls returned a junk value. Hence, it was identified that "__fls" >> would have an undefined behavior if the argument is passed as "0". >> Therefore, add changes to handle the same. > > Umm. That makes it sound like you are surprised by this behaviour of > __fls() and expected something else?! Please go read the documentation, > and then rewrite the commit message. You really shouldn't be surprised > by this. > > johannes Thanks Johannes for reviewing the changes. Sure, I will update the commit message in next patch version. Thanks, Harshitha
diff --git a/drivers/net/wireless/ath/ath12k/dp_rx.c b/drivers/net/wireless/ath/ath12k/dp_rx.c index 5923c7c9eaff..ea8dfea22be6 100644 --- a/drivers/net/wireless/ath/ath12k/dp_rx.c +++ b/drivers/net/wireless/ath/ath12k/dp_rx.c @@ -3238,7 +3238,7 @@ static int ath12k_dp_rx_frag_h_mpdu(struct ath12k *ar, goto out_unlock; } - if (frag_no > __fls(rx_tid->rx_frag_bitmap)) + if ((!rx_tid->rx_frag_bitmap || frag_no > __fls(rx_tid->rx_frag_bitmap))) __skb_queue_tail(&rx_tid->rx_frags, msdu); else ath12k_dp_rx_h_sort_frags(ab, &rx_tid->rx_frags, msdu);
When max virtual ap interfaces are configured in all the bands with ACS and hostapd restart is done every 60s, a crash is observed at random times because of handling the uninitialized peer fragments. During the crash, the fragment id of that packet was 0 and __fls returned a junk value. Hence, it was identified that "__fls" would have an undefined behavior if the argument is passed as "0". Therefore, add changes to handle the same. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1 Signed-off-by: Harshitha Prem <quic_hprem@quicinc.com> --- drivers/net/wireless/ath/ath12k/dp_rx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)