| Message ID | 20250203214448.1978156-1-aloka.dixit@oss.qualcomm.com |
|---|---|
| Headers | show |
| Series | wifi:ath11k/ath12k: refactor tx_arvif retrieval | expand |
On 2/3/2025 8:41 PM, Aditya Kumar Singh wrote: > On 2/4/25 03:14, Aloka Dixit wrote: >> Function ath11k_mac_setup_bcn_tmpl() retrieves tx_arvif only for >> a sanity check and then calls ath11k_mac_setup_bcn_tmpl_mbssid() >> or ath11k_mac_setup_bcn_tmpl_ema() both of which again retrieve >> the same pointer. Instead store the pointer and pass it to the >> latter two functions. >> > > Same, Is this tested? Perhaps you forgot to add "Tested-on:" tag? > >> Signed-off-by: Aloka Dixit <aloka.dixit@oss.qualcomm.com> >> --- > > No, only ath12k patches are tested hence no tag here for ath11k. This patch doesn't change handling functionally and I confirmed that all places using 'tx_arvif' first do NULL check because the refactored function can return NULL. Functions ath11k_mac_setup_bcn_tmpl_mbssid() and ath11k_mac_setup_bcn_tmpl_ema() always receive a non-NULL value now because the caller ath11k_mac_setup_bcn_tmpl() sets 'tx_arvif = arvif' whenever applicable.
On 2/4/25 23:58, Aloka Dixit wrote: > On 2/3/2025 8:41 PM, Aditya Kumar Singh wrote: >> On 2/4/25 03:14, Aloka Dixit wrote: >>> Function ath11k_mac_setup_bcn_tmpl() retrieves tx_arvif only for >>> a sanity check and then calls ath11k_mac_setup_bcn_tmpl_mbssid() >>> or ath11k_mac_setup_bcn_tmpl_ema() both of which again retrieve >>> the same pointer. Instead store the pointer and pass it to the >>> latter two functions. >>> >> >> Same, Is this tested? Perhaps you forgot to add "Tested-on:" tag? >> >>> Signed-off-by: Aloka Dixit <aloka.dixit@oss.qualcomm.com> >>> --- >> >> > > No, only ath12k patches are tested hence no tag here for ath11k. > This patch doesn't change handling functionally and I confirmed that all > places using 'tx_arvif' first do NULL check because the refactored > function can return NULL. Functions ath11k_mac_setup_bcn_tmpl_mbssid() > and ath11k_mac_setup_bcn_tmpl_ema() always receive a non-NULL value now > because the caller ath11k_mac_setup_bcn_tmpl() sets 'tx_arvif = arvif' > whenever applicable. Okay so in that case "No functionality change. Compile tested only." line should be there in commit text?
On 2/4/2025 7:58 PM, Aditya Kumar Singh wrote: > On 2/4/25 23:58, Aloka Dixit wrote: >> On 2/3/2025 8:41 PM, Aditya Kumar Singh wrote: >>> On 2/4/25 03:14, Aloka Dixit wrote: >>>> Function ath11k_mac_setup_bcn_tmpl() retrieves tx_arvif only for >>>> a sanity check and then calls ath11k_mac_setup_bcn_tmpl_mbssid() >>>> or ath11k_mac_setup_bcn_tmpl_ema() both of which again retrieve >>>> the same pointer. Instead store the pointer and pass it to the >>>> latter two functions. >>>> >>> >>> Same, Is this tested? Perhaps you forgot to add "Tested-on:" tag? >>> >>>> Signed-off-by: Aloka Dixit <aloka.dixit@oss.qualcomm.com> >>>> --- >>> >>> >> >> No, only ath12k patches are tested hence no tag here for ath11k. >> This patch doesn't change handling functionally and I confirmed that >> all places using 'tx_arvif' first do NULL check because the refactored >> function can return NULL. Functions ath11k_mac_setup_bcn_tmpl_mbssid() >> and ath11k_mac_setup_bcn_tmpl_ema() always receive a non-NULL value >> now because the caller ath11k_mac_setup_bcn_tmpl() sets 'tx_arvif = >> arvif' whenever applicable. > > Okay so in that case "No functionality change. Compile tested only." > line should be there in commit text? > Sure, let me know any other comments, will send all together in a new version.
On 2/3/2025 1:44 PM, Aloka Dixit wrote: > MLO MBSSID series for cfg80211/mac80211, which is in review here > https://patchwork.kernel.org/project/linux-wireless/list/?series=902914&state=%2A&archive=both, > changes the way transmitted profile data is stored for each link. > As drivers currently retrieve 'arvif' for transmitted profile at > multiple places, MLO MBSSID changes will becomes unnecessarily > bloated with RCU operations for every retrieval. Refactor 'tx_arvif' > to avoid this. > > Additionally, fix an issue in ath11k_mac_update_vif_chan() where > tx_arvif is not reset to NULL inside 'for' loop during each > iteration. > > v3: address comment from v2 and removed NULL check for > vif->mbssid_tx_vif from patch #1. > > v2: fixed kernel test robot error in patch 1/5 for tx_vif. > https://lore.kernel.org/oe-kbuild-all/202501160626.Jbb3GHnk-lkp@intel.com > > Aloka Dixit (5): > wifi: ath11k: refactor transmitted arvif retrieval > wifi: ath11k: pass tx arvif for MBSSID and EMA beacon generation > wifi: ath12k: refactor transmitted arvif retrieval > wifi: ath12k: pass tx arvif for MBSSID and EMA beacon generation > wifi: ath12k: pass BSSID index as input for EMA > > drivers/net/wireless/ath/ath11k/mac.c | 75 ++++++++++++------------- > drivers/net/wireless/ath/ath12k/mac.c | 81 +++++++++++++-------------- > 2 files changed, 76 insertions(+), 80 deletions(-) > > > base-commit: 1ed13fdf7d8284742c8ede36dd553387839e5163 can you rebase? this has a conflict with: wifi: ath12k: fix handling of CSA offsets in beacon template command