SHR status

[William Kenworthy]

On Mon, 2009-10-19 at 10:56 +0200, Michal Brzozowski wrote:
> 2009/10/19 W.Kenworthy <billk at iinet.net.au>
>  
>         debugfs exposes some knobs and innards of the kernel, the one
>         here
>         making the kernels default scheduler behave more like the BFS
>         scheduler
>         that hit the news a couple of months ago, creating a far more
>         responsive
>         device including when it rings.
> 
> It's not about responsiveness, there is simply a bug somewhere.
> Changing the scheduler might hide the bug, but that's rather bad isn't
> it? 
> 
> _______________________________________________


Yes, but it depends whether you want to receive calls or not - I prefer
to talk to my callers, not have to call them back :)

Its a known problem that has been there ever since they switched to FSO.
Its supposed to be gradually fixed as they move away from python towards
C for the underlying code, but I think scheduling is having a greater
effect at the moment as proven by this hack.

It doesnt totally fix it, as I was just using midori full screen (in
landscape via xrandr) and on exiting midori the caller hung up by the
time the display stabilised - and didnt leave a number :(

BillK