From: Ben Greear <greearb@candelatech.com>
To: Adrian Chadd <adrian@freebsd.org>
Cc: Sujith Manoharan <sujith@msujith.org>,
"linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>
Subject: Re: 3.7.6+: ath9k: tx logic locks up after taking attenuation very high.
Date: Thu, 21 Feb 2013 16:11:13 -0800 [thread overview]
Message-ID: <5126B7A1.6030301@candelatech.com> (raw)
In-Reply-To: <CAJ-Vmom3v1+==u0RS8d6m+P-=DR8A755--t2_gQR+6Szf5N5VQ@mail.gmail.com>
On 02/21/2013 03:56 PM, Adrian Chadd wrote:
> Right. As you increase attenuation, two things happen:
>
> * You're going to see lots more filtered frames pop up; which may
> involve pause/unpause of the queue;
> * You're going to see BAR frames go out, which will involve
> pausing/unpausing the queue.
>
> So you may have found some very nice way to reproduce pause/unpause
> related problems that are brought on by lots of TX failures leading to
> BARs and filtered frames.
>
> (Having your BAR frames be filtered .. god I hope not.)
>
> So see whether there's some odd scenario where a queue is unpaused and
> things aren't drained from the pending queue?
The hang I see is where the stats sit like this (below) and never change (for the first queue)
To me, the interesting part is that the BE queue is stopped while
axq-depth and axq-ampdu depth are zero. I don't fully understand
this code..but I sort of expected the depths to be non-zero when
packets were being processed by the hardware..maybe that's totally
incorrect...
It seems to hang forever once it gets stuck in this case unless
I reset the chip.
I'm not sure if it's related, but the tx rates do not drop as low
as I believe they should. These systems are using ath9k rate control,
and I plan to change that to minstrel_ht once I can get the current
config to at least recover when attenuation gets better....
[root@lf0301-demo1 ~]# cat /debug/ieee80211/wiphy0/ath9k/xmit ;sleep 10; cat /debug/ieee80211/wiphy0/ath9k/xmit ;
Num-Tx-Queues: 10 tx-queues-setup: 0x10f poll-work-seen: 1823
BE BK VI VO
MPDUs Queued: 500 0 0 4102
MPDUs Completed: 473 0 0 4102
MPDUs XRetried: 27 0 0 0
Aggregates: 200 0 0 0
AMPDUs Queued HW: 97039 0 0 0
AMPDUs Queued SW: 11299 0 0 0
AMPDUs Completed: 107982 0 0 0
AMPDUs Retried: 410 0 0 0
AMPDUs XRetried: 116 0 0 0
TXERR Filtered: 224 0 0 0
FIFO Underrun: 0 0 0 0
TXOP Exceeded: 0 0 0 0
TXTIMER Expiry: 0 0 0 0
DESC CFG Error: 0 0 0 0
DATA Underrun: 0 0 0 0
DELIM Underrun: 0 0 0 0
TX-Pkts-All: 108598 0 0 4102
TX-Bytes-All: 164436206 0 0 304838
hw-put-tx-buf: 107863 0 0 4096
hw-tx-start: 0 0 0 0
hw-tx-proc-desc: 0 0 0 0
TX-Failed: 0 0 0 0
txq-memory-address: f2b724e8 f2b725a4 f2b7242c f2b72370
axq-qnum: 2 3 1 0
axq-depth: 0 0 0 0
axq-ampdu_depth: 0 0 0 0
axq-stopped 1 0 0 0
tx-in-progress 0 0 0 0
pending-frames 123 0 0 0
txq_headidx: 7 0 0 0
txq_tailidx: 7 0 0 0
axq_q empty: 1 1 1 1
axq_acq empty: 1 1 1 1
txq_fifo[0] empty: 1 1 1 1
txq_fifo[1] empty: 1 1 1 1
txq_fifo[2] empty: 1 1 1 1
txq_fifo[3] empty: 1 1 1 1
txq_fifo[4] empty: 1 1 1 1
txq_fifo[5] empty: 1 1 1 1
txq_fifo[6] empty: 1 1 1 1
txq_fifo[7] empty: 1 1 1 1
Num-Tx-Queues: 10 tx-queues-setup: 0x10f poll-work-seen: 1833
BE BK VI VO
MPDUs Queued: 500 0 0 4140
MPDUs Completed: 473 0 0 4140
MPDUs XRetried: 27 0 0 0
Aggregates: 200 0 0 0
AMPDUs Queued HW: 97039 0 0 0
AMPDUs Queued SW: 11299 0 0 0
AMPDUs Completed: 107982 0 0 0
AMPDUs Retried: 410 0 0 0
AMPDUs XRetried: 116 0 0 0
TXERR Filtered: 224 0 0 0
FIFO Underrun: 0 0 0 0
TXOP Exceeded: 0 0 0 0
TXTIMER Expiry: 0 0 0 0
DESC CFG Error: 0 0 0 0
DATA Underrun: 0 0 0 0
DELIM Underrun: 0 0 0 0
TX-Pkts-All: 108598 0 0 4140
TX-Bytes-All: 164436206 0 0 306788
hw-put-tx-buf: 107863 0 0 4134
hw-tx-start: 0 0 0 0
hw-tx-proc-desc: 0 0 0 0
TX-Failed: 0 0 0 0
txq-memory-address: f2b724e8 f2b725a4 f2b7242c f2b72370
axq-qnum: 2 3 1 0
axq-depth: 0 0 0 0
axq-ampdu_depth: 0 0 0 0
axq-stopped 1 0 0 0
tx-in-progress 0 0 0 0
pending-frames 123 0 0 0
txq_headidx: 7 0 0 6
txq_tailidx: 7 0 0 6
axq_q empty: 1 1 1 1
axq_acq empty: 1 1 1 1
txq_fifo[0] empty: 1 1 1 1
txq_fifo[1] empty: 1 1 1 1
txq_fifo[2] empty: 1 1 1 1
txq_fifo[3] empty: 1 1 1 1
txq_fifo[4] empty: 1 1 1 1
txq_fifo[5] empty: 1 1 1 1
txq_fifo[6] empty: 1 1 1 1
txq_fifo[7] empty: 1 1 1 1
Thanks,
Ben
--
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc http://www.candelatech.com
next prev parent reply other threads:[~2013-02-22 0:11 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-02-11 18:18 3.7.6+: ath9k: tx logic locks up after taking attenuation very high Ben Greear
2013-02-12 0:11 ` Sujith Manoharan
2013-02-13 16:41 ` Ben Greear
2013-02-19 2:14 ` Sujith Manoharan
2013-02-19 2:36 ` Sujith Manoharan
2013-02-19 5:18 ` Ben Greear
2013-02-20 22:51 ` Ben Greear
2013-02-21 1:06 ` Ben Greear
2013-02-21 1:38 ` Sujith Manoharan
2013-02-21 5:41 ` Ben Greear
2013-02-21 17:34 ` Ben Greear
2013-02-21 23:11 ` Ben Greear
2013-02-21 23:49 ` Adrian Chadd
2013-02-21 23:53 ` Ben Greear
2013-02-21 23:56 ` Adrian Chadd
2013-02-22 0:11 ` Ben Greear [this message]
2013-02-22 4:23 ` Sujith Manoharan
2013-02-21 1:36 ` Sujith Manoharan
2013-02-21 2:02 ` Ben Greear
2013-02-21 2:04 ` Sujith Manoharan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5126B7A1.6030301@candelatech.com \
--to=greearb@candelatech.com \
--cc=adrian@freebsd.org \
--cc=linux-wireless@vger.kernel.org \
--cc=sujith@msujith.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).