From: "Schüll, Daniel" <daniel.schuell@fau.de>
To: Larry Finger <Larry.Finger@lwfinger.net>,
Phillip Potter <phil@philpotter.co.uk>
Cc: linux-wireless@vger.kernel.org
Subject: Re: [r8188eu] Worse performance with 5.15
Date: Sat, 29 Jan 2022 12:03:21 +0100 [thread overview]
Message-ID: <ba4e758af458bfbce51b8df79898f783@fau.de> (raw)
In-Reply-To: <0ef8d27e-d430-8016-e52a-fc2dbd87809a@lwfinger.net>
On 2022-01-29 04:55, Larry Finger wrote:
> On 1/28/22 16:13, Schüll, Daniel wrote:
>>
>> After upgrading to the 5.15 kernel I noticed a drop in throughput when
>> using my wifi dongle.
>> I'm using a TP-Link TL-WN725N-v2 (USB ID: 0bda:8179).
>>
>> On the 5.15 kernel I was only able to get around 30 MBit/s in download
>> (using google's speedtest).
>> After downgrading back to the 5.13 kernel I was back to 95 MBit/s I
>> used to have (close to the max from my ISP).
>>
>> A bit of research revealed, that you did a major overhaul of the
>> r8188eu driver in 5.15.
>> I then compiled the old rtl8188eu driver from the 5.13 sources for
>> 5.15, loaded it on 5.15 and was able to get 95 MBit/s again.
>> Finally I also tried the newest version of the new r8188eu driver
>> (Commit fa78315452) on 5.15, but still only got 30 MBits/s.
>> All test used the same location of AP and dongle and were done
>> multiple times, so influence from other wifi devices is negligible.
>> Seeing that the issue is fixable by using an other driver module on
>> the same kernel, it seems very likely,
>> that the issue is in the driver and not in the setup or other kernel
>> changes.
>>
>> Is the throughput a known regression?
>
> Daniel,
>
> This throughput regression is not known to me. I have not been very
> involved with the changes being made to r8188eu, but I do test it
> regularly to ensure that it can still make a connection. With my
> tests, I do not recall getting a speed over 40-45 Mbps connecting to a
> local LibreSpeed instance connected to my AP via a 1 GHz wired line.
> As I recall, I never have gotten anything near 95 Mbps.
>
> I generated a 5.13 kernel, and it did not get over 30 Mbps. I get the
> same download speed from Speedtest.net. As my up connection only runs
> at 12 Mbps, that is what I always get when using an external
> connection.
that's interesting. Can you confirm that you are using the 40 MHz
frequency range?
My AP is an AVM FritzBox!3390. In its menu I can see that the wifi
dongle is using the following configuration: 2.4 GHz, n, 40 MHz, WPA2, 1
x 1
With the one antenna of the rtl8188eu and 40 MHz the theoretical limit
is 150 Mbps.
I also retested with a local iperf server on the fritzbox.
This is where things get strange. With the new r8188eu driver iperf over
TCP gives me ~60 Mbps.
While google's speedtest still gives me only ~30 Mbps (I also tested
various other in-browser tests with same results).
With the old driver rtl8188eu again I get ~95 Mbps over iperf and
in-browser.
Regards,
Daniel
parent reply other threads:[~2022-01-29 11:11 UTC|newest]
Thread overview: expand[flat|nested] mbox.gz Atom feed
[parent not found: <0ef8d27e-d430-8016-e52a-fc2dbd87809a@lwfinger.net>]
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=ba4e758af458bfbce51b8df79898f783@fau.de \
--to=daniel.schuell@fau.de \
--cc=Larry.Finger@lwfinger.net \
--cc=linux-wireless@vger.kernel.org \
--cc=phil@philpotter.co.uk \
/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).