On 11/21/24 12:24, Ben Greear wrote:
On 11/21/24 12:11 PM, Korenblit, Miriam Rachel wrote:
-----Original Message-----
From: greearb@xxxxxxxxxxxxxxx <greearb@xxxxxxxxxxxxxxx>
Sent: Wednesday, 28 August 2024 23:26
To: linux-wireless@xxxxxxxxxxxxxxx
Cc: Ben Greear <greearb@xxxxxxxxxxxxxxx>
Subject: [PATCH 1/2] wifi: iwlwifi: Fix eMLSR band comparison.
From: Ben Greear <greearb@xxxxxxxxxxxxxxx>
Do not make assumptions about what band 'a' or 'b' are on.
And add new reason code for when eMLSR is disabled due to band.
Signed-off-by: Ben Greear <greearb@xxxxxxxxxxxxxxx>
---
drivers/net/wireless/intel/iwlwifi/mvm/link.c | 13 ++++++++++---
drivers/net/wireless/intel/iwlwifi/mvm/mvm.h | 2 ++
2 files changed, 12 insertions(+), 3 deletions(-)
diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/link.c
b/drivers/net/wireless/intel/iwlwifi/mvm/link.c
index bb3de07bc6be..f3fb37fec8a8 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/link.c
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/link.c
@@ -16,6 +16,7 @@
HOW(EXIT_LOW_RSSI) \
HOW(EXIT_COEX) \
HOW(EXIT_BANDWIDTH) \
+ HOW(EXIT_BAND) \
HOW(EXIT_CSA) \
HOW(EXIT_LINK_USAGE)
@@ -750,11 +751,17 @@ bool iwl_mvm_mld_valid_link_pair(struct
ieee80211_vif *vif,
iwl_mvm_esr_disallowed_with_link(mvm, vif, b, false))
return false;
- if (a->chandef->width != b->chandef->width ||
- !(a->chandef->chan->band == NL80211_BAND_6GHZ &&
- b->chandef->chan->band == NL80211_BAND_5GHZ))
+ if (a->chandef->width != b->chandef->width)
ret |= IWL_MVM_ESR_EXIT_BANDWIDTH;
+ /* Only supports 5g and 6g bands at the moment */
+ if (((a->chandef->chan->band != NL80211_BAND_6GHZ) &&
+ (a->chandef->chan->band != NL80211_BAND_5GHZ)) ||
+ ((b->chandef->chan->band != NL80211_BAND_6GHZ) &&
+ (b->chandef->chan->band != NL80211_BAND_5GHZ)) ||
+ (a->chandef->chan->band == b->chandef->chan->band))
+ ret |= IWL_MVM_ESR_EXIT_BAND;
+
if (ret) {
IWL_DEBUG_INFO(mvm,
"Links %d and %d are not a valid pair for EMLSR, a-
chwidth: %d b: %d band-a: %d band-b: %d\n", diff --git
a/drivers/net/wireless/intel/iwlwifi/mvm/mvm.h
b/drivers/net/wireless/intel/iwlwifi/mvm/mvm.h
index ac4e135ed91b..22bec9ca46bb 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/mvm.h
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/mvm.h
@@ -368,6 +368,7 @@ struct iwl_mvm_vif_link_info {
* preventing the enablement of EMLSR
* @IWL_MVM_ESR_EXIT_CSA: CSA happened, so exit EMLSR
* @IWL_MVM_ESR_EXIT_LINK_USAGE: Exit EMLSR due to low tpt on secondary
link
+ * @IWL_MVM_ESR_EXIT_BAND: Exit EMLSR due to incompatible Bands
*/
enum iwl_mvm_esr_state {
IWL_MVM_ESR_BLOCKED_PREVENTION = 0x1,
@@ -382,6 +383,7 @@ enum iwl_mvm_esr_state {
IWL_MVM_ESR_EXIT_BANDWIDTH = 0x80000,
IWL_MVM_ESR_EXIT_CSA = 0x100000,
IWL_MVM_ESR_EXIT_LINK_USAGE = 0x200000,
+ IWL_MVM_ESR_EXIT_BAND = 0x400000,
};
#define IWL_MVM_BLOCK_ESR_REASONS 0xffff
--
2.42.0
Hi Ben.
It is actually required that a (the better link) will be on 6 GHz and b on 5 GHz
Regarding the new exit reason, it is not really needed as we can easily differentiate between the cases (from other logs)
Hello Miri,
I tested this patch, and it fixed problems for me when I ran a test that created
interfering traffic on 5ghz and then later on 6Ghz. I expected eMLSR mode to stay
active no matter where the interfering traffic existed. With this patch, and a few
others I posted, the be200 then works fairly well.
6Ghz is not always better, for instance in case where it is congested with
external traffic.
Can you please let me know *why* you think the better link must always be 6ghz in this case?
Hello Miriam,
I wanted to check to see if you still consider this patch invalid? If so, I'll adjust
it to work better as out-of-tree patch and add it to my pile.
If you think core logic is fine but the patch needs some tweaks, please let me know your
suggestions.
Thanks,
Ben
