mac80211: HT RX reorder buffer cleanup and timeout workaround
cover mac80211: HT RX reorder buffer cleanup and timeout workaround
There are some cases where frames in RX reorder buffer can get stuck for extended time due to various errors (implementation bugs, frames getting lost completely, etc.). The workaround added here seems to make us recover much more quickly in some of these cases. This does not cover all cases, but is a useful initial step in avoiding some HT aggregation issues.
Signature · 1 line
Jouni Malinen PGP id EFC895FA
PATCH 1/3 mac80211: Use a shared function to release frames from RX reorder buf
No need to duplicate the same code in two places (and that would be three after the followup patch).
--- net/mac80211/rx.c | 70 +++++++++++++++++++++++------------------------------- 1 file changed, 31 insertions(+), 39 deletions(-)
net/mac80211/rx.c +31 -39
@@ -2284,6 +2284,34 @@ static inline u16 seq_sub(u16 sq1, u16 s } +static void ieee80211_release_reorder_frame (struct ieee80211_hw *hw, +struct tid_ampdu_rx *tid_agg_rx, +int index) +{ +struct ieee80211_supported_band *sband; +struct ieee80211_rate *rate; +struct ieee80211_rx_status status; + +if (!tid_agg_rx->reorder_buf[index]) + goto no_frame; + +/* release the reordered frames to stack */ +memcpy (&status, tid_agg_rx->reorder_buf[index]->cb,sizeof (status)); + sband = hw->wiphy->bands[status.band]; +if (status.flag &RX_FLAG_HT ) + rate = sband->bitrates;/* TODO: HT rates */ +else + rate = &sband->bitrates[status.rate_idx]; +__ieee80211_rx_handle_packet (hw, tid_agg_rx->reorder_buf[index], + &status, rate); + tid_agg_rx->stored_mpdu_num--; + tid_agg_rx->reorder_buf[index] =NULL ; + +no_frame: + tid_agg_rx->head_seq_num =seq_inc (tid_agg_rx->head_seq_num); +} + +/* * As it function blongs to Rx path it must be called with * the proper rcu_read_lock protection for its flow.
@@ -2295,12 +2323,8 @@ static u8 ieee80211_sta_manage_reorder_b u16 mpdu_seq_num, int bar_req) { -struct ieee80211_local *local =hw_to_local (hw); -struct ieee80211_rx_status status;u16 head_seq_num, buf_size;int index; -struct ieee80211_supported_band *sband; -struct ieee80211_rate *rate; buf_size = tid_agg_rx->buf_size; head_seq_num = tid_agg_rx->head_seq_num;
@@ -2325,28 +2349,8 @@ static u8 ieee80211_sta_manage_reorder_b index =seq_sub (tid_agg_rx->head_seq_num, tid_agg_rx->ssn) % tid_agg_rx->buf_size; - -if (tid_agg_rx->reorder_buf[index]) { -/* release the reordered frames to stack */ -memcpy (&status, - tid_agg_rx->reorder_buf[index]->cb, -sizeof (status)); - sband = local->hw.wiphy->bands[status.band]; -if (status.flag &RX_FLAG_HT ) { -/* TODO: HT rates */ - rate = sband->bitrates; - }else { - rate = &sband->bitrates - [status.rate_idx]; - } -__ieee80211_rx_handle_packet (hw, - tid_agg_rx->reorder_buf[index], - &status, rate); - tid_agg_rx->stored_mpdu_num--; - tid_agg_rx->reorder_buf[index] =NULL ; - } - tid_agg_rx->head_seq_num = -seq_inc (tid_agg_rx->head_seq_num); +ieee80211_release_reorder_frame (hw, tid_agg_rx, + index); }if (bar_req)return 1 ;
@@ -2380,19 +2384,7 @@ static u8 ieee80211_sta_manage_reorder_b index =seq_sub (tid_agg_rx->head_seq_num, tid_agg_rx->ssn) % tid_agg_rx->buf_size;while (tid_agg_rx->reorder_buf[index]) { -/* release the reordered frame back to stack */ -memcpy (&status, tid_agg_rx->reorder_buf[index]->cb, -sizeof (status)); - sband = local->hw.wiphy->bands[status.band]; -if (status.flag &RX_FLAG_HT ) - rate = sband->bitrates;/* TODO: HT rates */ -else - rate = &sband->bitrates[status.rate_idx]; -__ieee80211_rx_handle_packet (hw, tid_agg_rx->reorder_buf[index], - &status, rate); - tid_agg_rx->stored_mpdu_num--; - tid_agg_rx->reorder_buf[index] =NULL ; - tid_agg_rx->head_seq_num =seq_inc (tid_agg_rx->head_seq_num); +ieee80211_release_reorder_frame (hw, tid_agg_rx, index); index =seq_sub (tid_agg_rx->head_seq_num, tid_agg_rx->ssn) % tid_agg_rx->buf_size; }
Signature · 2 lines
-- Jouni Malinen PGP id EFC895FA
PATCH 2/3 mac80211: Add a timeout for frames in the RX reorder buffer
In reply to Jouni MalinenThis patch allows skbs to be released from the RX reorder buffer in case they have been there for an unexpectedly long time without us having received the missing frames before them. Previously, these frames were only released when the reorder window moved and that could take very long time unless new frames were received constantly (e.g., TCP connections could be killed more or less indefinitely).
This situation should not happen very frequently, but it looks like there are some scenarious that trigger it for some reason. As such, this should be considered mostly a workaround to speed up recovery from unexpected siutation that could result in connections hanging for long periods of time.
The changes here will only check for timeout situation when adding new RX frames to the reorder buffer. It does not handle all possible cases, but seems to help for most cases that could result from common network usage (e.g., TCP retrying at least couple of times). For more completely coverage, a timer could be used to periodically check whether there are any frames remaining in the reorder buffer if no new frames are received.
--- net/mac80211/agg-rx.c | 8 +++++++- net/mac80211/rx.c | 47 ++++++++++++++++++++++++++++++++++++++++++++++- net/mac80211/sta_info.h | 2 ++ 3 files changed, 55 insertions(+), 2 deletions(-)
net/mac80211/agg-rx.c +7 -1
@@ -68,6 +68,7 @@ void __ieee80211_stop_rx_ba_session(struspin_lock_bh (&sta->lock);/* free resources */ kfree (sta->ampdu_mlme.tid_rx[tid]->reorder_buf); +kfree (sta->ampdu_mlme.tid_rx[tid]->reorder_time);if (!sta->ampdu_mlme.tid_rx[tid]->shutdown) {kfree (sta->ampdu_mlme.tid_rx[tid]);
@@ -268,13 +269,18 @@ void ieee80211_process_addba_request(str/* prepare reordering buffer */ tid_agg_rx->reorder_buf =kcalloc (buf_size,sizeof (struct sk_buff *),GFP_ATOMIC ); -if (!tid_agg_rx->reorder_buf) { + tid_agg_rx->reorder_time = +kcalloc (buf_size,sizeof (unsigned long ),GFP_ATOMIC ); +if (!tid_agg_rx->reorder_buf || !tid_agg_rx->reorder_time) {#ifdef CONFIG_MAC80211_HT_DEBUG if (net_ratelimit ())printk (KERN_ERR "can not allocate reordering buffer " "to tid %d\n" , tid);#endif +kfree (tid_agg_rx->reorder_buf); +kfree (tid_agg_rx->reorder_time);kfree (sta->ampdu_mlme.tid_rx[tid]); + sta->ampdu_mlme.tid_rx[tid] =NULL ; goto end; }
net/mac80211/sta_info.h +2 -0
@@ -88,6 +88,7 @@ struct tid_ampdu_tx {* struct tid_ampdu_rx - TID aggregation information (Rx). * * @reorder_buf: buffer to reorder incoming aggregated MPDUs +* @reorder_time: jiffies when skb was added * @session_timer: check if peer keeps Tx-ing on the TID (by timeout value) * @head_seq_num: head sequence number in reordering buffer. * @stored_mpdu_num: number of MPDUs in reordering buffer
@@ -99,6 +100,7 @@ struct tid_ampdu_tx {*/ struct tid_ampdu_rx {struct sk_buff **reorder_buf; +unsigned long *reorder_time;struct timer_list session_timer;u16 head_seq_num;u16 stored_mpdu_num;
net/mac80211/rx.c +46 -1
65 lines not on this page · raw message
Signature · 2 lines
-- Jouni Malinen PGP id EFC895FA
PATCH 3/3 mac80211: Comment the order of HT RX reorder handler vs. RX handlers
In reply to Jouni MalinenWe are currently processing block ack reordering as a separate task before all other RX handlers. In theory, this is wrong since this step should be done only after duplicate removal (see Figure 6-1 in IEEE 802.11n). However, moving this needs some work and the current situation is not too bad. Add a comment here so that this small detail does not get forgotten and who knows, maybe someone has some extra time to take a look at cleaning this up.
--- net/mac80211/rx.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)
net/mac80211/rx.c +12 -0
@@ -2551,6 +2551,18 @@ void __ieee80211_rx(struct ieee80211_hwreturn ; } +/* +* In theory, the block ack reordering should happen after duplicate +* removal (ieee80211_rx_h_check(), which is an RX handler). As such, +* the call to ieee80211_rx_reorder_ampdu() should really be moved to +* happen as a new RX handler between ieee80211_rx_h_check and +* ieee80211_rx_h_decrypt. This cleanup may eventually happen, but for +* the time being, the call can be here since RX reorder buf processing +* will implicitly skip duplicates. We could, in theory at least, +* process frames that ieee80211_rx_h_passive_scan would drop (e.g., +* frames from other than operational channel), but that should not +* happen in normal networks. +*/ if (!ieee80211_rx_reorder_ampdu (local, skb, status))__ieee80211_rx_handle_packet (hw, skb, status, rate);
Signature · 2 lines
-- Jouni Malinen PGP id EFC895FA
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
In reply to Jouni MalinenHi Jouni,
Quoted 7 lines
This patch allows skbs to be released from the RX reorder buffer in case they have been there for an unexpectedly long time without us having received the missing frames before them. Previously, these frames were only released when the reorder window moved and that could take very long time unless new frames were received constantly (e.g., TCP connections could be killed more or less indefinitely).
Quoted 5 lines
This situation should not happen very frequently, but it looks like there are some scenarious that trigger it for some reason. As such, this should be considered mostly a workaround to speed up recovery from unexpected siutation that could result in connections hanging for long periods of time.
I can confirm that this used to be a regular situation between my X200 and a D-Link access point. I was originally thinking this was a driver issue, but your patch makes it possible to use wireless-testing tree again. I am not kidding here, before your patch it was impossible to use it at all. For me it was a real frequent situation. It is still present, but now it handles it more gracefully:
Open BA session requested for 00:1c:f0:xx:xx:xx tid 0
iwlagn 0000:03:00.0: iwl_tx_agg_start on ra = 00:1c:f0:xx:xx:xx tid = 0 activated addBA response timer on tid 0
switched off addBA timer for tid 0
Aggregation is on for tid 0
Rx A-MPDU request on tid 0 result 0
unexpected AddBA Req from 00:1c:f0:xx:xx:xx on tid 0
phy0: release an RX reorder frame due to timeout on earlier frames
phy0: release an RX reorder frame due to timeout on earlier frames
My VPN connection doesn't always like it, but that could also be a total different issue with the provider.
Regards
Marcel
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
Hi Johannes,
Quoted 20 lines
Quoted 7 lines
This patch allows skbs to be released from the RX reorder buffer in case they have been there for an unexpectedly long time without us having received the missing frames before them. Previously, these frames were only released when the reorder window moved and that could take very long time unless new frames were received constantly (e.g., TCP connections could be killed more or less indefinitely).
Quoted 5 lines
This situation should not happen very frequently, but it looks like there are some scenarious that trigger it for some reason. As such, this should be considered mostly a workaround to speed up recovery from unexpected siutation that could result in connections hanging for long periods of time.
I can confirm that this used to be a regular situation between my X200 and a D-Link access point. I was originally thinking this was a driver issue, but your patch makes it possible to use wireless-testing tree again. I am not kidding here, before your patch it was impossible to use it at all. For me it was a real frequent situation. It is still present, but now it handles it more gracefully:
Quoted 9 lines
Open BA session requested for 00:1c:f0:xx:xx:xx tid 0
iwlagn 0000:03:00.0:iwl_tx_agg_starton ra = 00:1c:f0:xx:xx:xx tid = 0 activated addBA response timer on tid 0
switched off addBA timer for tid 0
Aggregation is on for tid 0
Rx A-MPDU request on tid 0 result 0
unexpected AddBA Req from 00:1c:f0:xx:xx:xx on tid 0
phy0: release an RX reorder frame due to timeout on earlier frames
phy0: release an RX reorder frame due to timeout on earlier frames
so I finally got the debug output for you. Took only over a day :)
phy0: AddBA: ssn=195, dialog_token=1 tid=0 timeout=0ba_policy=1
addba: d0 00 40 01 00 16 eb 05 46 5c 00 1c f0 62 88 5b
addba: 00 1c f0 62 88 5b 50 d7 03 00 01 02 10 00 00 30
addba: 0c
Rx A-MPDU request on tid 0 result 0
Open BA session requested for 00:1c:f0:xx:xx:xx tid 0
iwlagn 0000:03:00.0: iwl_tx_agg_start on ra = 00:1c:f0:xx:xx:xx tid = 0
activated addBA response timer on tid 0
switched off addBA timer for tid 0
Aggregation is on for tid 0
phy0: AddBA: ssn=371, dialog_token=2 tid=0 timeout=0ba_policy=1
addba: d0 00 40 01 00 16 eb 05 46 5c 00 1c f0 62 88 5b
addba: 00 1c f0 62 88 5b 90 f0 03 00 02 02 10 00 00 30
addba: 17
phy0: RX reorder buf: head_seq=371 ssn=195 buf_size=64 stored_mpdu_num=0 timeout=0 index=48; - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - unexpected AddBA Req from 00:1c:f0:xx:xx:xx on tid 0
phy0: release an RX reorder frame due to timeout on earlier frames
phy0: RX reorder buf: head_seq=2001 ssn=195 buf_size=64 stored_mpdu_num=3 timeout=0 index=14; - - - - - - - - - - - - - - - - 236 229 0 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
phy0: release an RX reorder frame due to timeout on earlier frames
phy0: RX reorder buf: head_seq=2004 ssn=195 buf_size=64 stored_mpdu_num=2 timeout=0 index=17; - - - - - - - - - - - - - - - - - 229 0 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
phy0: release an RX reorder frame due to timeout on earlier frames
phy0: RX reorder buf: head_seq=2197 ssn=195 buf_size=64 stored_mpdu_num=3 timeout=0 index=18; - - - - - - - - - - - - - - - - - - - 124 124 0 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
phy0: release an RX reorder frame due to timeout on earlier frames
phy0: RX reorder buf: head_seq=2199 ssn=195 buf_size=64 stored_mpdu_num=2 timeout=0 index=20; - - - - - - - - - - - - - - - - - - - - 124 0 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Regards
Marcel
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
Quoted 2 lines · Marcel Holtmann
Quoted 2 lines
I can confirm that this used to be a regular situation between my X200 and a D-Link access point.
Which D-Link model is that AP? I think I should try to get one of those added to me test bed.. ;-)
Quoted 1 line
so I finally got the debug output for you. Took only over a day :)
Thanks! Would you happen to have timing information available for these (e.g., from klogd)? It looks like the AP is sending out an ADDBA Request to update some parameters, but we currently ignore that request. However, at least in this particular case, our RX reorder bug head_seq matches with the ssn from the ADDBA Request, so I'm not sure whether ignoring the ADDBA contents is really causing harm here (anyway, we should really process these updates, too).
The timeouts on RX reorder frames look similar to what I have seen in my tests and the workaround was indeed trying to address that type of issue, so it is nice to hear that it helped in this case, too.
Signature · 1 line
Jouni Malinen PGP id EFC895FA
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
Hi Jouni,
Quoted 5 lines
Quoted 2 lines
Quoted 2 lines
I can confirm that this used to be a regular situation between my X200 and a D-Link access point.
Which D-Link model is that AP? I think I should try to get one of those added to me test bed.. ;-)
from the WPS information element:
* Version: 1.0
* Manufacturer: D-Link Systems
* Model: DIR-615
* Device name: Wireless N Router
* Config methods: Label
I bought it over a year ago in Canada. It is actually my second of these since I fried the first one :)
Quoted 10 lines
Quoted 1 line
so I finally got the debug output for you. Took only over a day :)
Thanks! Would you happen to have timing information available for these (e.g., from klogd)? It looks like the AP is sending out an ADDBA Request to update some parameters, but we currently ignore that request. However, at least in this particular case, our RX reorder bug
head_seqmatches with the ssn from the ADDBA Request, so I'm not sure whether ignoring the ADDBA contents is really causing harm here (anyway, we should really process these updates, too).
Quoted 3 lines
The timeouts on RX reorder frames look similar to what I have seen in my tests and the workaround was indeed trying to address that type of issue, so it is nice to hear that it helped in this case, too.
I don't have the timing details since my Fedora for some reason doesn't log them to the filesystem. I have to fix that.
Also re-producing this is not as simple as you think. I have no idea when this happens. There is no way to actually trigger it from my side.
Regards
Marcel
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
In reply to Jouni MalinenHi Jouni,
Quoted 10 lines
Quoted 1 line
so I finally got the debug output for you. Took only over a day :)
Thanks! Would you happen to have timing information available for these (e.g., from klogd)? It looks like the AP is sending out an ADDBA Request to update some parameters, but we currently ignore that request. However, at least in this particular case, our RX reorder bug
head_seqmatches with the ssn from the ADDBA Request, so I'm not sure whether ignoring the ADDBA contents is really causing harm here (anyway, we should really process these updates, too).
Quoted 3 lines
The timeouts on RX reorder frames look similar to what I have seen in my tests and the workaround was indeed trying to address that type of issue, so it is nice to hear that it helped in this case, too.
so I enabled timing information for dmesg and got another report. It is not exactly the same and I don't know how and why it happened.
[ 249.579036] phy0: AddBA: ssn=50, dialog_token=1 tid=0 timeout=0ba_policy=1 [ 249.579046] addba: d0 00 40 01 00 16 eb 05 46 5c 00 1c f0 62 88 5b [ 249.579052] addba: 00 1c f0 62 88 5b 30 8f 03 00 01 02 10 00 00 20 [ 249.579057] addba: 03 [ 249.579073] Rx A-MPDU request on tid 0 result 0 [ 250.572022] phy0: AddBA: ssn=178, dialog_token=2 tid=0 timeout=0ba_policy=1 [ 250.572032] addba: d0 00 40 01 00 16 eb 05 46 5c 00 1c f0 62 88 5b [ 250.572038] addba: 00 1c f0 62 88 5b e0 8f 03 00 02 02 10 00 00 20 [ 250.572042] addba: 0b [ 250.572048] phy0: RX reorder buf: head_seq=178 ssn=50 buf_size=64 stored_mpdu_num=0 timeout=0 index=0; - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [ 250.572108] unexpected AddBA Req from 00:1c:f0:xx:xx:xx on tid 0 [ 266.263651] Open BA session requested for 00:1c:f0:xx:xx:xx tid 0 [ 266.263659] iwlagn 0000:03:00.0: iwl_tx_agg_start on ra = 00:1c:f0:xx:xx:xx tid = 0 [ 266.263726] activated addBA response timer on tid 0 [ 266.265534] wrong addBA response token, tid 0 [ 267.263467] addBA response timer expired on tid 0 [ 267.263516] Stopping Tx BA session for 00:1c:f0:xx:xx:xx tid 0 [ 326.083081] Open BA session requested for 00:1c:f0:xx:xx:xx tid 0 [ 326.083097] iwlagn 0000:03:00.0: iwl_tx_agg_start on ra = 00:1c:f0:xx:xx:xx tid = 0 [ 326.083197] activated addBA response timer on tid 0 [ 326.088517] wrong addBA response token, tid 0 [ 327.080206] addBA response timer expired on tid 0 [ 327.080262] Stopping Tx BA session for 00:1c:f0:xx:xx:xx tid 0
This kernel is running latest wireless-testing.git with debug patch and the disassoc change (for testing) from Johannes.
Regards
Marcel
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
In reply to Marcel HoltmannQuoted 10 lines · Marcel Holtmann
from the WPS information element:
* Version: 1.0
* Manufacturer: D-Link Systems
* Model: DIR-615
* Device name: Wireless N Router
* Config methods: LabelI bought it over a year ago in Canada. It is actually my second of these since I fried the first one :)
Thanks! I do actually have a DIR-615. However, it may be completely different WLAN chipset depending on the hardware revision (I love the vendors who do this..). If you have a chance, could you please take a look at which hardware revision your AP is? It is likely in the label on the AP (something like "H/W Ver.: D1 F/W Ver.: 4.00).
Signature · 1 line
Jouni Malinen PGP id EFC895FA
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
Hi Jouni,
Quoted 8 lines
Quoted 8 lines
from the WPS information element:
* Version: 1.0
* Manufacturer: D-Link Systems
* Model: DIR-615
* Device name: Wireless N Router
* Config methods: Label
Quoted 8 lines
Quoted 2 lines
I bought it over a year ago in Canada. It is actually my second of these since I fried the first one :)
Thanks! I do actually have a DIR-615. However, it may be completely different WLAN chipset depending on the hardware revision (I love the vendors who do this..). If you have a chance, could you please take a look at which hardware revision your AP is? It is likely in the label on the AP (something like "H/W Ver.: D1 F/W Ver.: 4.00).
it is Hardware Version: B2 and Firmware Version: 2.23.
Regards
Marcel
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
In reply to Marcel HoltmannHi Jouni,
Quoted 41 lines
Quoted 10 lines
Quoted 1 line
so I finally got the debug output for you. Took only over a day :)
Thanks! Would you happen to have timing information available for these (e.g., from klogd)? It looks like the AP is sending out an ADDBA Request to update some parameters, but we currently ignore that request. However, at least in this particular case, our RX reorder bug
head_seqmatches with the ssn from the ADDBA Request, so I'm not sure whether ignoring the ADDBA contents is really causing harm here (anyway, we should really process these updates, too).Quoted 3 lines
The timeouts on RX reorder frames look similar to what I have seen in my tests and the workaround was indeed trying to address that type of issue, so it is nice to hear that it helped in this case, too.
so I enabled timing information for dmesg and got another report. It is not exactly the same and I don't know how and why it happened.
[ 249.579036] phy0: AddBA: ssn=50, dialog_token=1 tid=0 timeout=0ba_policy=1 [ 249.579046] addba: d0 00 40 01 00 16 eb 05 46 5c 00 1c f0 62 88 5b [ 249.579052] addba: 00 1c f0 62 88 5b 30 8f 03 00 01 02 10 00 00 20 [ 249.579057] addba: 03 [ 249.579073] Rx A-MPDU request on tid 0 result 0 [ 250.572022] phy0: AddBA: ssn=178, dialog_token=2 tid=0 timeout=0ba_policy=1 [ 250.572032] addba: d0 00 40 01 00 16 eb 05 46 5c 00 1c f0 62 88 5b [ 250.572038] addba: 00 1c f0 62 88 5b e0 8f 03 00 02 02 10 00 00 20 [ 250.572042] addba: 0b [ 250.572048] phy0: RX reorder buf: head_seq=178 ssn=50 buf_size=64 stored_mpdu_num=0 timeout=0 index=0; - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [ 250.572108] unexpected AddBA Req from 00:1c:f0:xx:xx:xx on tid 0 [ 266.263651] Open BA session requested for 00:1c:f0:xx:xx:xx tid 0 [ 266.263659] iwlagn 0000:03:00.0: iwl_tx_agg_start on ra = 00:1c:f0:xx:xx:xx tid = 0 [ 266.263726] activated addBA response timer on tid 0 [ 266.265534] wrong addBA response token, tid 0 [ 267.263467] addBA response timer expired on tid 0 [ 267.263516] Stopping Tx BA session for 00:1c:f0:xx:xx:xx tid 0 [ 326.083081] Open BA session requested for 00:1c:f0:xx:xx:xx tid 0 [ 326.083097] iwlagn 0000:03:00.0: iwl_tx_agg_start on ra = 00:1c:f0:xx:xx:xx tid = 0 [ 326.083197] activated addBA response timer on tid 0 [ 326.088517] wrong addBA response token, tid 0 [ 327.080206] addBA response timer expired on tid 0 [ 327.080262] Stopping Tx BA session for 00:1c:f0:xx:xx:xx tid 0
Quoted 2 lines
This kernel is running latest wireless-testing.git with debug patch and the disassoc change (for testing) from Johannes.
I had problems with my DIR-615 and Intel 5350 card. No stable connection at all. No idea why it breaks down. Worked fine for the last few days and today it breaks after a few bytes. Maybe my environment is not clean enough. So I switched to use a Time Capsule in A-band only mode. Seems more stable with exactly the same kernel. The reorder details from the debug patch are still showing up:
[20236.165674] Open BA session requested for 00:1f:f3:xx:x:xx tid 0 [20236.165688] iwlagn 0000:03:00.0: iwl_tx_agg_start on ra = 00:1f:f3:xx:xx:xx tid = 0 [20236.165769] activated addBA response timer on tid 0 [20236.166197] switched off addBA timer for tid 0 [20236.166202] Aggregation is on for tid 0 [21108.747984] phy0: AddBA: ssn=347, dialog_token=1 tid=0 timeout=0ba_policy=1 [21108.747994] addba: d0 00 3c 00 00 16 eb 05 46 5c 00 1f f3 c3 a3 21 [21108.748000] addba: 00 1f f3 c3 a3 21 20 7e 03 00 01 02 10 00 00 b0 [21108.748005] addba: 15 [21108.748021] Rx A-MPDU request on tid 0 result 0 [21108.996253] phy0: AddBA: ssn=609, dialog_token=2 tid=0 timeout=0ba_policy=1 [21108.996264] addba: d0 00 3c 00 00 16 eb 05 46 5c 00 1f f3 c3 a3 21 [21108.996270] addba: 00 1f f3 c3 a3 21 50 7e 03 00 02 02 10 00 00 10 [21108.996274] addba: 26 [21108.996281] phy0: RX reorder buf: head_seq=609 ssn=347 buf_size=64 stored_mpdu_num=0 timeout=0 index=6; - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [21108.996342] unexpected AddBA Req from 00:1f:f3:xx:xx:xx on tid 0 [21114.164106] phy0: AddBA: ssn=686, dialog_token=3 tid=1 timeout=0ba_policy=1 [21114.164110] addba: d0 00 3c 00 00 16 eb 05 46 5c 00 1f f3 c3 a3 21 [21114.164112] addba: 00 1f f3 c3 a3 21 f0 81 03 00 03 06 10 00 00 e0 [21114.164114] addba: 2a [21114.164122] Rx A-MPDU request on tid 1 result 0 [21122.003948] delba from 00:1f:f3:c3:a3:21 (initiator) tid 1 reason code 1 [21122.003958] Rx BA session stop requested for 00:1f:f3:xx:xx:xx tid 1
Regards
Marcel
Re: [PATCH 2/3] mac80211: Add a timeout for frames in the RX reorder buffer
In reply to Marcel HoltmannQuoted 13 lines · Marcel Holtmann
Hi Jouni,
Quoted 11 lines
Quoted 10 lines
from the WPS information element:
* Version: 1.0
* Manufacturer: D-Link Systems
* Model: DIR-615
* Device name: Wireless N Router
* Config methods: LabelI bought it over a year ago in Canada. It is actually my second of these since I fried the first one :)
Quoted 7 lines
Quoted 5 lines
Thanks! I do actually have a DIR-615. However, it may be completely different WLAN chipset depending on the hardware revision (I love the vendors who do this..). If you have a chance, could you please take a look at which hardware revision your AP is? It is likely in the label on the AP (something like "H/W Ver.: D1 F/W Ver.: 4.00).
it is Hardware Version: B2 and Firmware Version: 2.23.
B and later uses a Ubicom integrated chipset; A series was Marvel TopDog.
Dan