[LITMUS^RT] Budget Consumed By Jobs In Ready Queue
Jonathan Herman
hermanjl at cs.unc.edu
Mon Sep 10 16:55:43 CEST 2012
I can't reproduce this bug using 2012.2. If this bug reappears and we
can reproduce it, we'll integrate the new patch. For now, let's hold
off.
On Thu, Sep 6, 2012 at 5:01 PM, Jeremy Erickson <jerickso at cs.unc.edu> wrote:
> The patch does fix the bug. I should point out that Jonathan's solution
> also does fix the bug as well - that's the solution you explain as more
> complicated (requiring changes in multiple plugins), correct?
>
> My testing method was to use a task system/scheduler combination (with some
> modifications I made to G-EDF) that consistently triggers the bug without
> either of the patches, and running it numerous times. It's the same system
> that led me to discover the bug in the first place. I do know that vanilla
> GSN-EDF (pre-patches) also triggers the bug, but it does so much less
> reliably.
>
> -Jeremy Erickson
>
>
> On Thu, Sep 6, 2012 at 4:30 AM, Björn Brandenburg <bbb at mpi-sws.org> wrote:
>>
>>
>> On Sep 5, 2012, at 11:27 PM, Jeremy Erickson wrote:
>>
>> > I verified that the bug was occurring in 2012.2. That release should
>> > have the patches, shouldn't it?
>>
>> Yes, thanks for testing this thoroughly. It's a real bug then.
>>
>> I've pushed a proposed fix to wip-budget-bug on Github.
>>
>>
>> https://github.com/LITMUS-RT/litmus-rt/commit/4b22bff2bac8c7b4cd50e5aef4fec75bbd96b234
>>
>> The patch compiles, but I haven't tested it. Jeremy, can you please
>> thoroughly test the patch and report whether it fixes your problem?
>>
>> Thanks,
>> Björn
>>
>>
>> _______________________________________________
>> litmus-dev mailing list
>> litmus-dev at lists.litmus-rt.org
>> https://lists.litmus-rt.org/listinfo/litmus-dev
>
>
>
> _______________________________________________
> litmus-dev mailing list
> litmus-dev at lists.litmus-rt.org
> https://lists.litmus-rt.org/listinfo/litmus-dev
>
--
Jonathan Herman
Department of Computer Science at UNC Chapel Hill
More information about the litmus-dev
mailing list