[LITMUS^RT] new patches and new branch

[Björn Brandenburg]

On Jun 26, 2013, at 5:54 PM, Björn Brandenburg <bbb at mpi-sws.org> wrote:

> On Jun 26, 2013, at 5:43 PM, Glenn Elliott <gelliott at cs.unc.edu> wrote:
>> 
>> On Jun 26, 2013, at 7:08 AM, Björn Brandenburg <bbb at mpi-sws.org> wrote:
>> 
>>> On Jun 25, 2013, at 4:52 PM, Glenn Elliott <gelliott at cs.unc.edu> wrote:
>>> 
>>>> did you test Litmus with "threadirqs" set in the kernel boot parameters? Litmus tasks blocked for an I/O interrupt could suffer an unbounded priority inversion though...
> 
> We ran the kernel "as is", without specifying additional options. threadirqs would probably have reduced the outliers considerably, but not as much as PREEMPT_RT.

Hi Glenn,

a quick followup to your question. Felipe ran additional experiments with threadirqs enabled, and it turned out that it hardly helps at all. Threading ISRs is somewhat pointless if the ISRs still disable preemptions for prolonged times. You can find the details starting on slide 99 in the presentation, which I've just uploaded.

	https://www.mpi-sws.org/~bbb/papers/talks/ospert13.pdf

- Björn