mbox series

[v3,0/5] wifi:ath11k/ath12k: refactor tx_arvif retrieval

Message ID 20250203214448.1978156-1-aloka.dixit@oss.qualcomm.com
Headers show
Series wifi:ath11k/ath12k: refactor tx_arvif retrieval | expand

Message

Aloka Dixit Feb. 3, 2025, 9:44 p.m. UTC 
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

Comments

Aloka Dixit Feb. 4, 2025, 6:28 p.m. UTC | #1 
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.