iwl3945: increase plcp check threshold
PATCH 1/1 iwl3945: increase plcp check threshold
iwl3945: increase plcp check threshold
Patch fixes:
https://bugzilla.redhat.com/show_bug.cgi?id=654599
Many users report very low speed problem on 3945 devices, this patch fixes problem, but only for some of them.
For unknown reason, sometimes after hw scanning, device is not able to receive frames at high rate. Since plcp health check may request hw scan to "reset radio", performance problem start to be observable after update kernel to .35, where plcp check was introduced.
Bug reporter confirmed that increase plcp check threshold fixed problem for him.
---
Patch is oneliner, but does not apply on older kernel versions, I will provide -stable patches when fix will land in mainline
drivers/net/wireless/iwlwifi/iwl-3945.c +1 -1
@@ -2775,7 +2775,7 @@ static struct iwl_base_params iwl3945_base_params = { .use_isr_legacy = true, .led_compensation = 64, .broken_powersave = true, - .plcp_delta_threshold = IWL_MAX_PLCP_ERR_LONG_THRESHOLD_DEF, + .plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_MAX, .wd_timeout = IWL_DEF_WD_TIMEOUT, .max_event_log_size = 512, .tx_power_by_driver = true,
Stanislaw,
Quoted 28 lines · Stanislaw Gruszka
iwl3945: increase plcp check threshold
Patch fixes:
https://bugzilla.redhat.com/show_bug.cgi?id=654599Many users report very low speed problem on 3945 devices, this patch fixes problem, but only for some of them.
For unknown reason, sometimes after hw scanning, device is not able to receive frames at high rate. Since plcp health check may request hw scan to "reset radio", performance problem start to be observable after update kernel to .35, where plcp check was introduced.
Bug reporter confirmed that increase plcp check threshold fixed problem for him.
Reported-and-tested-by SilvioTO <[email protected]>Cc [email protected] # 2.6.35+Signed-off-by Stanislaw Gruszka <[email protected]>---
Patch is oneliner, but does not apply on older kernel versions, I will provide -stable patches when fix will land in mainlinedrivers/net/wireless/iwlwifi/iwl-3945.c +0
-0
drivers/net/wireless/iwlwifi/iwl-3945.c static struct iwl_base_params iwl3945_base_params = { · line 2781
Quoted 8 lines
drivers/net/wireless/iwlwifi/iwl-3945.c +1
-1@@ -2775,7 +2775,7 @@ static struct iwl_base_params iwl3945_base_params = { .use_isr_legacy = true, .led_compensation = 64, .broken_powersave = true, - .plcp_delta_threshold = IWL_MAX_PLCP_ERR_LONG_THRESHOLD_DEF, + .plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_MAX, .wd_timeout = IWL_DEF_WD_TIMEOUT, .max_event_log_size = 512, .tx_power_by_driver = true,
Do you see any help for 3945 with plcp checking at all?
Wey
Quoted 24 lines · wwguy
Quoted 23 lines · Stanislaw Gruszka
iwl3945: increase plcp check threshold
Patch fixes:
https://bugzilla.redhat.com/show_bug.cgi?id=654599Many users report very low speed problem on 3945 devices, this patch fixes problem, but only for some of them.
For unknown reason, sometimes after hw scanning, device is not able to receive frames at high rate. Since plcp health check may request hw scan to "reset radio", performance problem start to be observable after update kernel to .35, where plcp check was introduced.
Bug reporter confirmed that increase plcp check threshold fixed problem for him.
Reported-and-tested-by SilvioTO <[email protected]>Cc [email protected] # 2.6.35+Signed-off-by Stanislaw Gruszka <[email protected]>---
Patch is oneliner, but does not apply on older kernel versions, I will provide -stable patches when fix will land in mainline
Quoted 15 lines
drivers/net/wireless/iwlwifi/iwl-3945.c static struct iwl_base_params iwl3945_base_params = { · line 2781
Quoted 13 lines
drivers/net/wireless/iwlwifi/iwl-3945.c +1
-1@@ -2775,7 +2775,7 @@ static struct iwl_base_params iwl3945_base_params = { .use_isr_legacy = true, .led_compensation = 64, .broken_powersave = true, - .plcp_delta_threshold = IWL_MAX_PLCP_ERR_LONG_THRESHOLD_DEF, + .plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_MAX, .wd_timeout = IWL_DEF_WD_TIMEOUT, .max_event_log_size = 512, .tx_power_by_driver = true,Do you see any help for 3945 with plcp checking at all?
Not really, I see only demage :-)
Ok, if we argree to remove plcp check from 3945 at all, I will repost with slightly different patch for fix problem on stable.
Stanislaw
PATCH v2 1/1 iwl3945: remove plcp check
Patch fixes:
https://bugzilla.redhat.com/show_bug.cgi?id=654599
Many users report very low speed problem on 3945 devices, this patch fixes problem, but only for some of them.
For unknown reason, sometimes after hw scanning, device is not able to receive frames at high rate. Since plcp health check may request hw scan to "reset radio", performance problem start to be observable after update kernel to .35, where plcp check was introduced.
Bug reporter confirmed that removing plcp check fixed problem for him.
---
v1 -> v2
Instead of tuning threshold remove the check completely
This is minimal fix, without any additional code removal (what introduce one compile warning). I will remove dead code after Intel push they current driver split work, to do not create merge problems now.
Patch also does not apply cleanly on 2.6.35, I will provide additional -stable patch when this one hit mainline.
drivers/net/wireless/iwlwifi/iwl-3945.c +0 -1
@@ -2734,7 +2734,6 @@ static struct iwl_lib_ops iwl3945_lib = { .isr_ops = { .isr = iwl_isr_legacy, }, - .check_plcp_health = iwl3945_good_plcp_health, .debugfs_ops = { .rx_stats_read = iwl3945_ucode_rx_stats_read,
Re: [PATCH v2] iwl3945: remove plcp check
Quoted 11 lines · Stanislaw Gruszka
Patch fixes:
https://bugzilla.redhat.com/show_bug.cgi?id=654599Many users report very low speed problem on 3945 devices, this patch fixes problem, but only for some of them.
For unknown reason, sometimes after hw scanning, device is not able to receive frames at high rate. Since plcp health check may request hw scan to "reset radio", performance problem start to be observable after update kernel to .35, where plcp check was introduced.
Quoted 5 lines
Bug reporter confirmed that removing plcp check fixed problem for him.
Reported-and-tested-by SilvioTO <[email protected]>Cc [email protected] # 2.6.35+Signed-off-by Stanislaw Gruszka <[email protected]>
Quoted 4 lines
---
v1 -> v2
Instead of tuning threshold remove the check completely
looks good for just minimal changes
Wey