comments on late joyrides re: rt performance

[Ton van Overbeek]

Daniel Drake wrote:
> Hi Victor,
>
> Thanks a lot for the feedback.
>
> On Thu, 2008-08-07 at 21:45 +0100, victor wrote:
>   
>> One of the things I have been playing with is a little MIDI file
>> player
>> (one of the test/example activities in
>> devel.laptop.org/activities/csndsugui),
>> which uses a soundfont synthesis engine. On the older system, I
>> can play most files with only the occasional dropout, generally
>> quite solidly. But on the later (2232 and 2263), I get many more
>> dropouts
>>     
>
> We are already aware of an audio performance regression and believe it
> be related to 2 things:
>  - hal polls our battery a lot more than it used to, even when fully 
>    charged, and EC commands disable interrupts and cause sound stutters.
>    (I'm working on fixing hal, I sent patches upstream)
>  - the kernel upgrade from 2.6.22 to 2.6.25 also included some changes
>    (which we haven't identified yet) which cause these problems.
>    Downgrading to 2.6.22 restores the previous audio performance.
>
> Thanks for pointing out your test-case though! So far we have been
> reproducing these issues by using TamTamMini, TamTam is indeed
> affected. :(
>
> Here is the relevant ticket:
> http://dev.laptop.org/ticket/7603
> Contributions welcome! It is a tricky one.
>
> Thanks,
> Daniel
>   
Here an other observation which might have something to do with this.
In joyride-2266 (booted from SD card) I noticed that there is a lot of
logging of the traffic to the EC.
Just do dmesg and you see a burst of  about 18 olpc-ec: messages
approximately every 30 seconds.
Can this extra logging cause hic-ups ??
This might be a red herring, but you never know ....

Ton van Overbeek