>> maybe I should report to openwrt-devel list, but I think it is the problem
>> of ath9k, caused by using the module minstrel_ht.
>
> I have the same issue using ath9k/minstrel_ht. I think the problem is due to
> the way the "streams" variable is calculated in minstrel_ht_get_group_idx().
> I printed some info when the warning occurs:
>
> minstrel_ht_get_group_idx[120]: streams 12 rate->flags & flags 32
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[0].streams 1
> minstrel_mcs_groups[0].flags 0
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[1].streams 2
> minstrel_mcs_groups[1].flags 0
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[2].streams 1
> minstrel_mcs_groups[2].flags 128
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[3].streams 2
> minstrel_mcs_groups[3].flags 128
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[4].streams 1
> minstrel_mcs_groups[4].flags 32
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[5].streams 2
> minstrel_mcs_groups[5].flags 32
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[6].streams 1
> minstrel_mcs_groups[6].flags 160
> minstrel_ht_get_group_idx[125]: minstrel_mcs_groups[7].streams 2
> minstrel_mcs_groups[7].flags 160
> ------------[ cut here ]------------
> WARNING: at
> /home/lorenzo/src/openwrt/buildroot/build_dir/linux-ar71xx/compat-wireless-2010-03-03/net/mac80211/rc80211_minstrel_ht.c:127
> 0x81f)
> Modules linked in: ath9k ath9k_common ath9k_hw ath mac80211 cfg80211
> compat_firmware_class compat deflate
> cbc
> Call Trace:[<80069d48>] 0x80069d48
> [<80069d48>] 0x80069d48
> [<81fa32f4>] 0x81fa32f4
> [<80082d58>] 0x80082d58
> [<81fa32f4>] 0x81fa32f4
> [<81fa32f4>] 0x81fa32f4
> [<800b9f90>] 0x800b9f90
> [<8006de40>] 0x8006de40
> [<81f81cbc>] 0x81f81cbc
> [<81f280b4>] 0x81f280b4
> [<81f2a26c>] 0x81f2a26c
> [<800ba25c>] 0x800ba25c
> [<81fc11f0>] 0x81fc11f0
> [<81f2aa3c>] 0x81f2aa3c
> [<81f273c8>] 0x81f273c8
> [<81f262c4>] 0x81f262c4
> [<8020b214>] 0x8020b214
> [<81f23570>] 0x81f23570
> [<800891d0>] 0x800891d0
> [<800897e4>] 0x800897e4
> [<80089e54>] 0x80089e54
> [<8006082c>] 0x8006082c
> [<800897e4>] 0x800897e4
> [<804179e8>] 0x804179e8
> [<81f23570>] 0x81f23570
> [<80060820>] 0x80060820
> [<80074024>] 0x80074024
> [<80060820>] 0x80060820
> [<8052bb0c>] 0x8052bb0c
> [<804179e8>] 0x804179e8
> [<80417a08>] 0x80417a08
> [<8050f6e8>] 0x8050f6e8
> [<80417a08>] 0x80417a08
> [<8056c0fc>] 0x8056c0fc
> [<8055be90>] 0x8055be90
> [<80445428>] 0x80445428
> [<805705f4>] 0x805705f4
> [<8006f730>] 0x8006f730
> [<805704c8>] 0x805704c8
> [<8006f720>] 0x8006f720
>
> ---[ end trace 454674867bb3dfb4 ]---
>
>> the reason I try minstrel_ht is I could only get around 50Mbps throughput
>> rate by using ath9k for my AR9280 card. some developer said the low
>> throughput is a known problem for ath9k on IPX4xx platform, it was kind of
>> optimizing work. I still cannot see any light on this problem since last
>> year.
>>
>> I saw the crashing on several trunk since OpenWrt r20294, till r20851, yes
>> it was caused by iperf test while it was just nice by ping test.
>> topology: PC1 - STA - AP - PC2
>>
>> iperf/ping was form PC1 to PC2, the crash happened on both AP and STA,
>>
>> see attachment for the log, Thank You !!
>>
>>
>>
>> _______________________________________________
>> ath9k-devel mailing list
>> ath9k-devel@lists.ath9k.org
>> https://lists.ath9k.org/mailman/listinfo/ath9k-devel
>>
>>
>
> Regards
>
> Lorenzo
>
Could you please try this??
Signed-off-by: Lorenzo Bianconi <lorenzo.biancon...@gmail.com>
---
--- a/net/mac80211/rc80211_minstrel_ht.c
+++ b/net/mac80211/rc80211_minstrel_ht.c
@@ -318,9 +318,9 @@
if (!mg->supported)
continue;
- if (++mg->index > MCS_GROUP_RATES) {
+ if (++mg->index >= MCS_GROUP_RATES) {
mg->index = 0;
- if (++mg->column > ARRAY_SIZE(sample_table))
+ if (++mg->column >= ARRAY_SIZE(sample_table))
mg->column = 0;
}
break;
--
Regards
Lorenzo
_______________________________________________
ath9k-devel mailing list
ath9k-devel@lists.ath9k.org
https://lists.ath9k.org/mailman/listinfo/ath9k-devel
