| Message ID | 20241008053744.1070-3-quic_kangyang@quicinc.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | wifi: ath11k: fix full monitor mode for QCN9074 | expand |
On 10/7/2024 10:37 PM, Kang Yang wrote: > ath11k_hal_srng_* should be used with srng->lock to protect srng data. > > For ath11k_dp_rx_mon_dest_process() and ath11k_dp_full_mon_process_rx(), > they use ath11k_hal_srng_* for many times but never call srng->lock. > > So when running (full) monitor mode, warning will occur: > RIP: 0010:ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] > Call Trace: > ? ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] > ath11k_dp_rx_process_mon_status+0xc45/0x1190 [ath11k] > ? idr_alloc_u32+0x97/0xd0 > ath11k_dp_rx_process_mon_rings+0x32a/0x550 [ath11k] > ath11k_dp_service_srng+0x289/0x5a0 [ath11k] > ath11k_pcic_ext_grp_napi_poll+0x30/0xd0 [ath11k] > __napi_poll+0x30/0x1f0 > net_rx_action+0x198/0x320 > __do_softirq+0xdd/0x319 > > So add srng->lock for them to avoid such warnings. > > Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30 > Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1 > > Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices") > Signed-off-by: Kang Yang <quic_kangyang@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> (note that since this has a locking change Kalle wants to review/approve so I've assigned this series back to him in patchwork)
Kang Yang <quic_kangyang@quicinc.com> writes: > ath11k_hal_srng_* should be used with srng->lock to protect srng data. > > For ath11k_dp_rx_mon_dest_process() and ath11k_dp_full_mon_process_rx(), > they use ath11k_hal_srng_* for many times but never call srng->lock. > > So when running (full) monitor mode, warning will occur: > RIP: 0010:ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] > Call Trace: > ? ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] > ath11k_dp_rx_process_mon_status+0xc45/0x1190 [ath11k] > ? idr_alloc_u32+0x97/0xd0 > ath11k_dp_rx_process_mon_rings+0x32a/0x550 [ath11k] > ath11k_dp_service_srng+0x289/0x5a0 [ath11k] > ath11k_pcic_ext_grp_napi_poll+0x30/0xd0 [ath11k] > __napi_poll+0x30/0x1f0 > net_rx_action+0x198/0x320 > __do_softirq+0xdd/0x319 > > So add srng->lock for them to avoid such warnings. > > Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30 > Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1 > > Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices") > Signed-off-by: Kang Yang <quic_kangyang@quicinc.com> [...] > @@ -5607,7 +5609,7 @@ static int ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, > struct hal_sw_mon_ring_entries *sw_mon_entries; > struct ath11k_pdev_mon_stats *rx_mon_stats; > struct sk_buff *head_msdu, *tail_msdu; > - void *mon_dst_srng = &ar->ab->hal.srng_list[dp->rxdma_mon_dst_ring.ring_id]; > + struct hal_srng *mon_dst_srng; > void *ring_entry; > u32 rx_bufs_used = 0, mpdu_rx_bufs_used; > int quota = 0, ret; > @@ -5623,6 +5625,9 @@ static int ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, > goto reap_status_ring; > } > > + mon_dst_srng = &ar->ab->hal.srng_list[dp->rxdma_mon_dst_ring.ring_id]; > + spin_lock_bh(&mon_dst_srng->lock); Why initialise mon_dst_srng differently? The commit message mentions nothing about this change.
On 12/12/2024 10:53 PM, Kalle Valo wrote: > Kang Yang <quic_kangyang@quicinc.com> writes: > >> ath11k_hal_srng_* should be used with srng->lock to protect srng data. >> >> For ath11k_dp_rx_mon_dest_process() and ath11k_dp_full_mon_process_rx(), >> they use ath11k_hal_srng_* for many times but never call srng->lock. >> >> So when running (full) monitor mode, warning will occur: >> RIP: 0010:ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] >> Call Trace: >> ? ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] >> ath11k_dp_rx_process_mon_status+0xc45/0x1190 [ath11k] >> ? idr_alloc_u32+0x97/0xd0 >> ath11k_dp_rx_process_mon_rings+0x32a/0x550 [ath11k] >> ath11k_dp_service_srng+0x289/0x5a0 [ath11k] >> ath11k_pcic_ext_grp_napi_poll+0x30/0xd0 [ath11k] >> __napi_poll+0x30/0x1f0 >> net_rx_action+0x198/0x320 >> __do_softirq+0xdd/0x319 >> >> So add srng->lock for them to avoid such warnings. >> >> Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30 >> Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1 >> >> Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices") >> Signed-off-by: Kang Yang <quic_kangyang@quicinc.com> > > [...] > >> @@ -5607,7 +5609,7 @@ static int ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, >> struct hal_sw_mon_ring_entries *sw_mon_entries; >> struct ath11k_pdev_mon_stats *rx_mon_stats; >> struct sk_buff *head_msdu, *tail_msdu; >> - void *mon_dst_srng = &ar->ab->hal.srng_list[dp->rxdma_mon_dst_ring.ring_id]; >> + struct hal_srng *mon_dst_srng; >> void *ring_entry; >> u32 rx_bufs_used = 0, mpdu_rx_bufs_used; >> int quota = 0, ret; >> @@ -5623,6 +5625,9 @@ static int ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, >> goto reap_status_ring; >> } >> >> + mon_dst_srng = &ar->ab->hal.srng_list[dp->rxdma_mon_dst_ring.ring_id]; >> + spin_lock_bh(&mon_dst_srng->lock); > > Why initialise mon_dst_srng differently? The commit message mentions > nothing about this change. Because need to fetch spin lock inside 'struct hal_srng'. If still use 'void *mon_dst_srng', need to perform a variable type cast. Bur 'struct hal_srng' will make this line too long: 'struct hal_srng *mon_dst_srng = &ar->ab->hal.srng_list[dp->rxdma_mon_dst_ring.ring_id];' So I separated the definition and initialization. ath11k_dp_rx_reap_mon_status_ring()/ath11k_dp_process_rx and others ring process function is the same. >
On 12/13/2024 2:17 PM, Kang Yang wrote: > > > On 12/12/2024 10:53 PM, Kalle Valo wrote: >> Kang Yang <quic_kangyang@quicinc.com> writes: >> >>> ath11k_hal_srng_* should be used with srng->lock to protect srng data. >>> >>> For ath11k_dp_rx_mon_dest_process() and ath11k_dp_full_mon_process_rx(), >>> they use ath11k_hal_srng_* for many times but never call srng->lock. >>> >>> So when running (full) monitor mode, warning will occur: >>> RIP: 0010:ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] >>> Call Trace: >>> ? ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] >>> ath11k_dp_rx_process_mon_status+0xc45/0x1190 [ath11k] >>> ? idr_alloc_u32+0x97/0xd0 >>> ath11k_dp_rx_process_mon_rings+0x32a/0x550 [ath11k] >>> ath11k_dp_service_srng+0x289/0x5a0 [ath11k] >>> ath11k_pcic_ext_grp_napi_poll+0x30/0xd0 [ath11k] >>> __napi_poll+0x30/0x1f0 >>> net_rx_action+0x198/0x320 >>> __do_softirq+0xdd/0x319 >>> >>> So add srng->lock for them to avoid such warnings. >>> >>> Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125- >>> QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30 >>> Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1 >>> >>> Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax >>> devices") >>> Signed-off-by: Kang Yang <quic_kangyang@quicinc.com> >> >> [...] >> >>> @@ -5607,7 +5609,7 @@ static int ath11k_dp_full_mon_process_rx(struct >>> ath11k_base *ab, int mac_id, >>> struct hal_sw_mon_ring_entries *sw_mon_entries; >>> struct ath11k_pdev_mon_stats *rx_mon_stats; >>> struct sk_buff *head_msdu, *tail_msdu; >>> - void *mon_dst_srng = &ar->ab->hal.srng_list[dp- >>> >rxdma_mon_dst_ring.ring_id]; >>> + struct hal_srng *mon_dst_srng; >>> void *ring_entry; >>> u32 rx_bufs_used = 0, mpdu_rx_bufs_used; >>> int quota = 0, ret; >>> @@ -5623,6 +5625,9 @@ static int ath11k_dp_full_mon_process_rx(struct >>> ath11k_base *ab, int mac_id, >>> goto reap_status_ring; >>> } >>> + mon_dst_srng = &ar->ab->hal.srng_list[dp- >>> >rxdma_mon_dst_ring.ring_id]; >>> + spin_lock_bh(&mon_dst_srng->lock); >> >> Why initialise mon_dst_srng differently? The commit message mentions >> nothing about this change. > > > Because need to fetch spin lock inside 'struct hal_srng'. If still use > 'void *mon_dst_srng', need to perform a variable type cast. > > Bur 'struct hal_srng' will make this line too long: > 'struct hal_srng *mon_dst_srng = &ar->ab->hal.srng_list[dp- > >rxdma_mon_dst_ring.ring_id];' > So I separated the definition and initialization. > > ath11k_dp_rx_reap_mon_status_ring()/ath11k_dp_process_rx and others ring > process function is the same. > Do i need to send a new version for this? > > >> > >
Kang Yang <quic_kangyang@quicinc.com> writes: >>>> @@ -5623,6 +5625,9 @@ static int >>>> ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, >>>> goto reap_status_ring; >>>> } >>>> + mon_dst_srng = &ar->ab->hal.srng_list[dp- >>>> >rxdma_mon_dst_ring.ring_id]; >>>> + spin_lock_bh(&mon_dst_srng->lock); >>> >>> Why initialise mon_dst_srng differently? The commit message mentions >>> nothing about this change. >> Because need to fetch spin lock inside 'struct hal_srng'. If still >> use 'void *mon_dst_srng', need to perform a variable type cast. >> Bur 'struct hal_srng' will make this line too long: >> 'struct hal_srng *mon_dst_srng = &ar->ab->hal.srng_list[dp- >> >rxdma_mon_dst_ring.ring_id];' >> So I separated the definition and initialization. >> ath11k_dp_rx_reap_mon_status_ring()/ath11k_dp_process_rx and others >> ring process function is the same. >> > > Do i need to send a new version for this? Yeah, please mention in the commit message why you mon_dst_srng initialisation.
diff --git a/drivers/net/wireless/ath/ath11k/dp_rx.c b/drivers/net/wireless/ath/ath11k/dp_rx.c index 072963d3396c..5fd681b5df2e 100644 --- a/drivers/net/wireless/ath/ath11k/dp_rx.c +++ b/drivers/net/wireless/ath/ath11k/dp_rx.c @@ -5146,7 +5146,7 @@ static void ath11k_dp_rx_mon_dest_process(struct ath11k *ar, int mac_id, struct ath11k_mon_data *pmon = (struct ath11k_mon_data *)&dp->mon_data; const struct ath11k_hw_hal_params *hal_params; void *ring_entry; - void *mon_dst_srng; + struct hal_srng *mon_dst_srng; u32 ppdu_id; u32 rx_bufs_used; u32 ring_id; @@ -5163,6 +5163,7 @@ static void ath11k_dp_rx_mon_dest_process(struct ath11k *ar, int mac_id, spin_lock_bh(&pmon->mon_lock); + spin_lock_bh(&mon_dst_srng->lock); ath11k_hal_srng_access_begin(ar->ab, mon_dst_srng); ppdu_id = pmon->mon_ppdu_info.ppdu_id; @@ -5221,6 +5222,7 @@ static void ath11k_dp_rx_mon_dest_process(struct ath11k *ar, int mac_id, mon_dst_srng); } ath11k_hal_srng_access_end(ar->ab, mon_dst_srng); + spin_unlock_bh(&mon_dst_srng->lock); spin_unlock_bh(&pmon->mon_lock); @@ -5607,7 +5609,7 @@ static int ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, struct hal_sw_mon_ring_entries *sw_mon_entries; struct ath11k_pdev_mon_stats *rx_mon_stats; struct sk_buff *head_msdu, *tail_msdu; - void *mon_dst_srng = &ar->ab->hal.srng_list[dp->rxdma_mon_dst_ring.ring_id]; + struct hal_srng *mon_dst_srng; void *ring_entry; u32 rx_bufs_used = 0, mpdu_rx_bufs_used; int quota = 0, ret; @@ -5623,6 +5625,9 @@ static int ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, goto reap_status_ring; } + mon_dst_srng = &ar->ab->hal.srng_list[dp->rxdma_mon_dst_ring.ring_id]; + spin_lock_bh(&mon_dst_srng->lock); + ath11k_hal_srng_access_begin(ar->ab, mon_dst_srng); while ((ring_entry = ath11k_hal_srng_dst_peek(ar->ab, mon_dst_srng))) { head_msdu = NULL; @@ -5666,6 +5671,7 @@ static int ath11k_dp_full_mon_process_rx(struct ath11k_base *ab, int mac_id, } ath11k_hal_srng_access_end(ar->ab, mon_dst_srng); + spin_unlock_bh(&mon_dst_srng->lock); spin_unlock_bh(&pmon->mon_lock); if (rx_bufs_used) {
ath11k_hal_srng_* should be used with srng->lock to protect srng data. For ath11k_dp_rx_mon_dest_process() and ath11k_dp_full_mon_process_rx(), they use ath11k_hal_srng_* for many times but never call srng->lock. So when running (full) monitor mode, warning will occur: RIP: 0010:ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] Call Trace: ? ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k] ath11k_dp_rx_process_mon_status+0xc45/0x1190 [ath11k] ? idr_alloc_u32+0x97/0xd0 ath11k_dp_rx_process_mon_rings+0x32a/0x550 [ath11k] ath11k_dp_service_srng+0x289/0x5a0 [ath11k] ath11k_pcic_ext_grp_napi_poll+0x30/0xd0 [ath11k] __napi_poll+0x30/0x1f0 net_rx_action+0x198/0x320 __do_softirq+0xdd/0x319 So add srng->lock for them to avoid such warnings. Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30 Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1 Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices") Signed-off-by: Kang Yang <quic_kangyang@quicinc.com> --- drivers/net/wireless/ath/ath11k/dp_rx.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-)