flashdisk real-time behavior

[Brett Swimley]

Chris Johns wrote:
> Till Straumann wrote:
>   
>> With the new flashdisk support there is an opportunity
>> for me to reiterate my rant at libblock/bdbuf.
>>     
>
> Thanks for raising this issue. I was not aware of it.
>
>   
I have been working on "stress testing" the bdbuf routine using an 
sdcard driver.  I occasionally see some lockups.  I had one version that 
had applied your patches, Till, to eliminate some race conditions but 
abandoned them when I was tracking some other issues.  I will reapply 
them to my current version and retest.
>> Could someone familiar with the flashdisk comment, please?
>>     
>
> I think the situation be even longer as segments could be compacted then 
> erased and there could be a number of these.
>
>   
>> Anyways, IMHO libblock/bdbuf should be converted to using mutex protection
>> (I have started some work in that direction).
>>     
>
> Great. May I suggest opening a PR and attaching your patch as it is and I will 
> take a look and sort through it. I have a need to look at this code in detail 
> so this may all tie together.
>
>   
I agree with the idea of mutex protection.

I tried a mutex approach with the existing bdbuf function and had 
limited success.  Perhaps your fresh take on this would help.  Do you 
have any sort of version of bdbuf.c that I could drop in and try?

Brett Swimley
> _______________________________________________
> rtems-users mailing list
> rtems-users at rtems.com
> http://rtems.rtems.org/mailman/listinfo/rtems-users
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/users/attachments/20080205/360b1b4f/attachment.html>