Re: [PATCH 5.10 187/575] ath10k: high latency fixes for beacon buffer

Hi Pavel,

On 16/11/2021 08:59, Pavel Machek wrote:

There's GFP_KERNEL vs. GFP_ATOMIC confusion here:

...

@@ -5466,10 +5470,17 @@ static int ath10k_add_interface(struct ieee80211_hw *hw, if (vif->type == NL80211_IFTYPE_ADHOC || vif->type == NL80211_IFTYPE_MESH_POINT || vif->type == NL80211_IFTYPE_AP) {

• arvif->beacon_buf = dma_alloc_coherent(ar->dev,
  

• 				       IEEE80211_MAX_FRAME_LEN,
  

• 				       &arvif->beacon_paddr,
  

• 				       GFP_ATOMIC);
  

• if (ar->bus_param.dev_type == ATH10K_DEV_TYPE_HL) {
  

• 	arvif->beacon_buf = kmalloc(IEEE80211_MAX_FRAME_LEN,
  

• 				    GFP_KERNEL);
  

• 	arvif->beacon_paddr = (dma_addr_t)arvif->beacon_buf;
  

• } else {
  

• 	arvif->beacon_buf =
  

• 		dma_alloc_coherent(ar->dev,
  

• 				   IEEE80211_MAX_FRAME_LEN,
  

• 				   &arvif->beacon_paddr,
  

• 				   GFP_ATOMIC);
  

• }
  

  if (!arvif->beacon_buf) { ret = -ENOMEM; ath10k_warn(ar, "failed to allocate beacon buffer: %d\n",

I'd expect both allocations to use same GFP_ flags.

Good catch.

Let me prepare a patch, test it and submit it soon.

Thanks,

Fabio Estevam