From: Marc Gonzalez <marc.w.gonzalez@free.fr>
To: AML <linux-amlogic@lists.infradead.org>
Cc: Martin Blumenstingl <martin.blumenstingl@gmail.com>,
Christian Hewitt <christian@hewittfamily.org.uk>,
Linux Wireless <linux-wireless@vger.kernel.org>,
MMC <linux-mmc@vger.kernel.org>,
Neil Armstrong <neil.armstrong@linaro.org>,
Ulf Hansson <ulf.hansson@linaro.org>,
Pierre-Hugues Husson <phh@phh.me>
Subject: Re: [Performance regression] BCM4359/9 on S905X2
Date: Mon, 17 Apr 2023 16:22:36 +0200 [thread overview]
Message-ID: <ac5d9c2e-13c3-7f99-a20c-d857282c6451@free.fr> (raw)
In-Reply-To: <74feca6f-4177-5645-6614-bc5f3dbf0be1@free.fr>
On 06/04/2023 18:35, Marc Gonzalez wrote:
> # curl -o /dev/null http://192.168.1.254:8095/fixed/1G
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 1024M 100 1024M 0 0 9747k 0 0:01:47 0:01:47 --:--:-- 9544k
I was able to import the bcmdhd driver from:
https://android.googlesource.com/kernel/hikey-linaro/+log/refs/heads/android-amlogic-bmeson-6.1/drivers/net/wireless/bcmdhd
as pointed out by Neil.
I switched from brcmfmac to bcmdhd with:
-CONFIG_BRCMFMAC=y
-CONFIG_BRCMDBG=y
+CONFIG_BCMDHD=y
+CONFIG_BCMDHD_FW_PATH=""
+CONFIG_BCMDHD_NVRAM_PATH=""
+CONFIG_BCMDHD_SDIO_IRQ=y
and hard-coded the FW paths in dhd_set_path_params().
Here are the benchmark results on the bcm4359 (AP6398SR3 module)
# curl -o /dev/null http://192.168.1.254:8095/fixed/1G
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 1024M 100 1024M 0 0 17.3M 0 0:00:58 0:00:58 --:--:-- 17.4M
which is noticeably faster (80%) than brcmfmac results.
FTR, system was 18.5% busy for this benchmark:
81.5653% 195114 0:swapper
13.8287% 33080 57:irq/17-ffe03000
1.3967% 3341 173:curl
1.1350% 2715 63:dhd_rxf
1.0455% 2501 129:ksdioirqd/mmc2
1.0004% 2393 62:dhd_dpc
0.0167% 40 171:perf
0.0088% 21 19:ksoftirqd/1
I was not able to get the driver to work with bcm43752 (AP6398SR32 module).
It loops with:
[ 15.221622] wl_run_escan: LEGACY_SCAN sync ID: 0, bssidx: 0
[ 15.221661] CFG80211-ERROR) wl_run_escan :
[ 15.221664] Escan set error (-37)
[ 15.229067] dhd_ioctl_entry_local invalid parameter
[ 15.233902] CFG80211-ERROR) wl_run_escan :
[ 15.233904] error (-37), cnt=1
[ 15.241056] CFG80211-ERROR) wl_cfg80211_scan :
[ 15.241059] scan error (-11)
But I may have bungled adding the "new" chip, since the driver architecture
is... [censored].
I suppose the conclusion is:
1) bcmdhd is 80% faster than brcmfmac in one download scenario
2) CPU usage scales linearly with throughput (on this system, 1% ~ 7 Mbps)
Regards
_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic
next prev parent reply other threads:[~2023-04-17 14:23 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-21 10:40 Marc Gonzalez
2023-03-21 10:46 ` Jerome Brunet
2023-03-21 12:37 ` Performance lower than expected with " Marc Gonzalez
2023-03-21 13:54 ` Jerome Brunet
2023-03-21 16:48 ` Marc Gonzalez
[not found] ` <EEE9FD80-C106-4A9B-AA8D-5C151E540CFD@hewittfamily.org.uk>
2023-04-04 16:09 ` [Performance regression] " Marc Gonzalez
2023-04-04 21:06 ` Martin Blumenstingl
2023-04-06 16:35 ` Marc Gonzalez
2023-04-06 17:07 ` Christian Hewitt
2023-04-06 17:41 ` Neil Armstrong
2023-04-17 14:22 ` Marc Gonzalez [this message]
2023-03-21 11:45 ` Neil Armstrong
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=ac5d9c2e-13c3-7f99-a20c-d857282c6451@free.fr \
--to=marc.w.gonzalez@free.fr \
--cc=christian@hewittfamily.org.uk \
--cc=linux-amlogic@lists.infradead.org \
--cc=linux-mmc@vger.kernel.org \
--cc=linux-wireless@vger.kernel.org \
--cc=martin.blumenstingl@gmail.com \
--cc=neil.armstrong@linaro.org \
--cc=phh@phh.me \
--cc=ulf.hansson@linaro.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
all inboxes | Powered by JetHome®