From: Felix Fietkau <nbd@openwrt.org>
To: ath9k-devel@lists.ath9k.org
Subject: [ath9k-devel] Enhancing ath9k for embedded systems
Date: Tue, 15 Dec 2009 17:17:30 +0100 [thread overview]
Message-ID: <4B27B69A.9000006@openwrt.org> (raw)
In-Reply-To: <15B47C78534E094AA9A41F58FB8D74F30182EF70DDB7@SC1EXMB-MBCL.global.atheros.com>
On 2009-12-15 2:18 PM, Ayee Goundan wrote:
> I apologize for the delay. Please use Sasi Subramaniam as your first
> point of contact for OpenWRT related questions, while cc'ing Senthil
> Balasubramanian. Sasi will coordinate appropriate help within the
> organization to provide as quick a response as possible.
Hi Sasi,
I've quoted my initial problem description below. Please let me know if
you have any questions about my details of my test setup.
Thanks,
- Felix
>> I'm running OpenWrt with our Linux 2.6.30 based kernel and ath9k from
>> compat-wireless on two identical devices based on AR7240 and AR9283
>> (Python+Merlin). The devices are 'Rocket M5' by Ubiquiti Networks.
>> One of them is configured to run as an AP, the other one is
>> connected to
>> it as STA.
>>
>> The main problem that I have been unable to track down so far is that
>> whenever I run traffic through those two devices, I get very bad
>> throughput (max. 60 mbit/s, mostly only 50 mbit/s).
>> Rate control stats show that the rate in use is mostly 216
>> MBit/s, with
>> a low packet error rate (fluctuating between 0 and 7%).
>> Packet captures
>> from an external device confirm this. Aggregation is
>> definitely working
>> properly, and the CPU load on both devices usually stays well
>> below 50%.
>>
>> I also ran similar tests with ath9k on completely different hardware,
>> with exactly the same issues. The other hardware combinations that I
>> tested were Intel IXP4xx based boards with AR5416, AR9160 and AR9220
>> based cards. The kernel that I used on those boards was 2.6.28.
>>
>> I can rule out issues with the Kernel or the Ethernet driver, since a
>> modified version of the Carrier LSDK driver is performing well in the
>> same setup (I included the carrier driver in the same firmware build
>> that I tested ath9k with).
>>
>> The weird part to this is that whenever I run the same set of
>> tests with
>> an older MacBook as STA against one of the ath9k devices as AP, I get
>> more than 95 MBit/s in both directions (limited by the Ethernet
>> connection speed). I also ran the same tests with a Ralink
>> RT2870 device
>> instead of the MacBook card, with equally good results.
>>
>> I've spent a considerable amount of time trying to track this
>> issue down
>> (including comparing register traces against a working
>> driver), with no
>> promising lead so far, so I'd appreciate any help from your side.
parent reply other threads:[~2009-12-15 16:17 UTC|newest]
Thread overview: expand[flat|nested] mbox.gz Atom feed
[parent not found: <15B47C78534E094AA9A41F58FB8D74F30182EF70DDB7@SC1EXMB-MBCL.global.atheros.com>]
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=4B27B69A.9000006@openwrt.org \
--to=nbd@openwrt.org \
--cc=ath9k-devel@lists.ath9k.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.