Re: [Gumstix-users] Question about mmc and preemptability

[Craig H. <cr...@hu...>]

I'm actually not 100% sure that statement is still correct.  What it 
meant was that if you compile MMC support into the kernel, you could 
not also compile in "pre-emptible kernel threads" support.  That is, 
system calls would not be pre-emptible if MMC support was in the 
kernel.  Now that the MMC driver is actually folded into mainline and 
not just an added patch, this problem may very well have been resolved. 
  If you need pre-emptible system calls, then please do just try it, and 
see if it works.  If it does, then update the wiki :)

Thanks,

C

On Nov 19, 2004, at 6:25 AM, Dave Hylands wrote:

> Hi,
>
> In the wiki, 
> http://www.gumstix.org/tikiwiki/tiki-index.php?page_ref_id=14
> it says,:
>
>     using the mmc makes the kernel non-pre-emptible.
>
> Does this mean that it's only non-pre-emptible while the mmc is being
> accessed (I'm hoping that this is the case), or that it's
> non-preemptible if mmc is mounted?