Re: [Xen-devel] [PATCH] xen: credit2: improve debug dump output.

[George Dunlap <george.dunlap@xxxxxxxxxx>]

On 18/01/17 01:10, Dario Faggioli wrote:
> Scheduling information debug dump for Credit2 is hard
> to read as it contains the same information repeated
> multiple time in different ways.
> 
> In fact, in Credit2, CPUs are grouped in runqueus. Before
> this change, for each CPU, we were printing the while
> content of the runqueue, as shown below:
> 
>  CPU[00]  sibling=000003, core=0000ff
>     run: [32767.0] flags=0 cpu=0 credit=-1073741824 [w=0] load=0 (~0%)
>       1: [0.0] flags=0 cpu=2 credit=3860932 [w=256] load=262144 (~100%)
>       2: [0.1] flags=0 cpu=2 credit=3859906 [w=256] load=262144 (~100%)
>  CPU[01]  sibling=000003, core=0000ff
>     run: [32767.1] flags=0 cpu=1 credit=-1073741824 [w=0] load=0 (~0%)
>       1: [0.0] flags=0 cpu=2 credit=2859840 [w=256] load=262144 (~100%)
>       2: [0.3] flags=0 cpu=2 credit=-17466062 [w=256] load=262144 (~100%)
>  CPU[02]  sibling=00000c, core=0000ff
>     run: [0.0] flags=2 cpu=2 credit=1858628 [w=256] load=262144 (~100%)
>       1: [0.3] flags=0 cpu=2 credit=-17466062 [w=256] load=262144 (~100%)
>       2: [0.1] flags=0 cpu=2 credit=-23957055 [w=256] load=262144 (~100%)
>  CPU[03]  sibling=00000c, core=0000ff
>     run: [32767.3] flags=0 cpu=3 credit=-1073741824 [w=0] load=0 (~0%)
>       1: [0.1] flags=0 cpu=2 credit=-3957055 [w=256] load=262144 (~100%)
>       2: [0.0] flags=0 cpu=2 credit=-6216254 [w=256] load=262144 (~100%)
>  CPU[04]  sibling=000030, core=0000ff
>     run: [32767.4] flags=0 cpu=4 credit=-1073741824 [w=0] load=0 (~0%)
>       1: [0.1] flags=0 cpu=2 credit=3782667 [w=256] load=262144 (~100%)
>       2: [0.3] flags=0 cpu=2 credit=-16287483 [w=256] load=262144 (~100%)
> 
> As it can be seen, all the CPUs print the whole content
> of the runqueue they belong to, at the time of their
> sampling, and this is cumbersome and hard to interpret!
> 
> In new output format we print, for each CPU, only the vCPU
> that is running there (if that's not the idle vCPU, in which
> case, nothing is printed), while the runqueues content
> is printed only once, in a dedicated section.
> 
> An example:
> 
>  CPUs info:
>  CPU[02]  runq=0, sibling=00000c, core=0000ff
>     run: [0.3] flags=2 cpu=2 credit=8054391 [w=256] load=262144 (~100%)
>  CPU[14]  runq=1, sibling=00c000, core=00ff00
>     run: [0.4] flags=2 cpu=14 credit=8771420 [w=256] load=262144 (~100%)
>  ... ... ... ... ... ... ... ... ...
>  Runqueue info:
>  runqueue 0:
>       0: [0.1] flags=0 cpu=2 credit=7869771 [w=256] load=262144 (~100%)
>       1: [0.0] flags=0 cpu=2 credit=7709649 [w=256] load=262144 (~100%)
>  runqueue 1:
>       0: [0.5] flags=0 cpu=14 credit=-1188 [w=256] load=262144 (~100%)
> 
> Note that there still is risk of inconsistency between
> what is printed in the 'Runqueue info:' and in 'CPUs info:'
> sections. That is unavoidable, as the relevant locks are
> released and re-acquired, around each single operation.
> 
> At least, the inconsistency is less severe than before.
> 
> Signed-off-by: Dario Faggioli <dario.faggioli@xxxxxxxxxx>

Reviewed-by: George Dunlap <george.dunlap@xxxxxxxxxx>


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel