[LITMUS^RT] (Un)loading Modules

[Björn Brandenburg]

On Sep 23, 2013, at 10:00 PM, Kevin Burns <kevinpb at vt.edu> wrote:

> Taking a closer look reveals that the problem lies within bheap_delete. As there is no deactivate function for gsn_edf this happens in the reactivation. I am wondering if there is a way to get around this problem by limited the use of static numbers like NR_CPUS. I am however open to suggestions!
> 
> Thanks,
> -Kevin
> 
> 
> On Mon, Sep 23, 2013 at 2:53 PM, Kevin Burns <kevinpb at vt.edu> wrote:
> Hello,
> 
> I am trying to run a test with a procedure like this:
> 
> 1) setsched GSN-EDF
> 2) run test
> 3) disable core on my machine
> 4) rerun test
> 
> However, it would seem that litmus cannot handle this change of hardware configuration without reloading the schedulers. So, my question is, is there a way to reload GSN-EDF from userspace, without rebootign the system?
> 
> Let me know if this question is unclear.
> 
> Thanks,
> -Kevin

Hi Kevin,

LITMUS^RT has currently no support for CPU hot-plugging. As you've correctly identified, a lot of the current code assumes that the number of online CPUs is static. I don't think anyone is working on fixing this at the moment, but I'd be happy to merge patches that rectify this assumption if they are not too intrusive.

Thanks,
Björn

PS: In the future, please subscribe to the list before posting; due to spam issues, I need to manually process each post by a non-member. Thanks.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.litmus-rt.org/pipermail/litmus-dev/attachments/20130924/e0b419f1/attachment.html>