From: Seraj Mohammed <seraj.mohammed@sondrel.com>
To: Peter Ujfalusi <peter.ujfalusi@ti.com>,
Seraj Alijan <seraj.alijan@sondrel.com>,
vkoul@kernel.org
Cc: dmaengine@vger.kernel.org, dan.j.williams@intel.com,
james.hartley@sondrel.com, sifan.naeem@sondrel.com,
ed.blake@sondrel.com
Subject: [V2,5/5] dmaengine: Documentation: Add documentation for multi chan testing
Date: Fri, 31 Aug 2018 16:48:16 +0100 [thread overview]
Message-ID: <c5cc69da-859d-167d-828f-78acf38cb3d8@sondrel.com> (raw)
Hi,
On 31/08/18 15:22, Peter Ujfalusi wrote:
> Hi,
>
> On 2018-08-31 16:41, Seraj Mohammed wrote:
>>> Can I still use:
>>> echo "" > /sys/module/dmatest/parameters/channel
>>>
>>> to stop dmatest to run on the specific channel(s) or do I need to reboot
>>> the machine to do so?
>>
>> is this behavior documented somewhere? or did you discover this from your
>> experience with using the module?
>
> If the channels is "" then it is not set. We can change the channels
> currently and there is no difference between selecting a channel or
> selecting none.
>
>> The short answer to your question is no, with this change you cannot stop
>> a test by echoing an empty string to the channel parameter, you will either
>> have to run the test, or reboot the machine.
>>
>> Setting the channel parameter to an empty string will request all
>> channels up to "max_channels" parameter value.
>
> Yes, and it is very useful. What would be even more useful is if via the
> channel I could select the starting channel from where the max_channels
> will start...
Since this patch already allows you to specifically select which
channels to run on, i don't see the benefit in allocating channels using
ranges.
>> This is because with this
>> change, editing the channel parameter will invoke the process of allocating
>> a thread for the specified channel and storing that thread in a thread
>> list, after that, setting run to 1 will simply iterate through the thread
>> list and start all pending threads.
>
> We have automated regression tests via dmatest's memcpy, it does changes
> parameters between runs to gather the data for different transfer sizes,
> runs in selected channels and runs also with different max_channels.
>
> It is a regression if we need to reboot between every test to change
> parameters, channels, etc.
Once you have successfully completed a test, you don't need to reboot to
start a new one with different parameters.
> - Péter
>
> Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
> Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
>
next reply other threads:[~2018-08-31 15:48 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-08-31 15:48 Seraj Mohammed [this message]
-- strict thread matches above, loose matches on Subject: below --
2018-09-04 19:15 [V2,5/5] dmaengine: Documentation: Add documentation for multi chan testing Peter Ujfalusi
2018-09-03 14:49 Seraj Mohammed
2018-09-03 8:19 Peter Ujfalusi
2018-08-31 15:42 Seraj Mohammed
2018-08-31 15:01 Seraj Mohammed
2018-08-31 14:22 Peter Ujfalusi
2018-08-31 14:02 Peter Ujfalusi
2018-08-31 13:41 Seraj Mohammed
2018-08-31 11:29 Peter Ujfalusi
2018-08-31 11:17 Peter Ujfalusi
2018-08-31 11:03 Seraj Alijan
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=c5cc69da-859d-167d-828f-78acf38cb3d8@sondrel.com \
--to=seraj.mohammed@sondrel.com \
--cc=dan.j.williams@intel.com \
--cc=dmaengine@vger.kernel.org \
--cc=ed.blake@sondrel.com \
--cc=james.hartley@sondrel.com \
--cc=peter.ujfalusi@ti.com \
--cc=seraj.alijan@sondrel.com \
--cc=sifan.naeem@sondrel.com \
--cc=vkoul@kernel.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