From: Jerry Van Baren <gerald.vanbaren_at_smiths-aerospace.com>
Date: Wed, 23 Feb 2005 15:24:05 -0500

Jens Arnold wrote:
> Hello Jerry

[snip]

>>An update: In the software I attempt to close the
>>powermgmt.csv file when USB is plugged in, but I'm not sure it
>>is successful. I ran a recharge test overnight and ended up
>>with a zero length file :-(.
>
> This close() attempt cannot work the way it is implemented,
> because the usb_inserted() check is asynchronous to the other
> threads. Chances are >99% that the USB connection is already
> accepted and disk control handed over to the USB bridge when
> this code snippet runs.
>
> If you want to implement it the proper way, you would have to
> give the power thread an event queue. This would then require
> the power thread to accept the USB connection, and it has the
> chance to close the file before doing so. Of course introducing
> an event queue then complicates the timing, because the thread
> cannot simply sleep a determined amount of time.
>
> Regards, Jens

Hi Jens,

Thanks for the tip. The logging is only for test purposes, it isn't
worth the complications of a queue, etc. I was just hoping that a
simple close would get the job done. My "work-around" is satisfactory
(at least to me), given that the logging "feature" is something that
only a very few people will be using.

gvb
_______________________________________________
http://cool.haxx.se/mailman/listinfo/rockbox
Received on 2005-02-23