lima: occasional job timeout errors

Occasionally I see job timeout errors, unfortunately I do not have much details like apitrace at moment, but I will try to gather this information.

[  132.807125] hrtimer: interrupt took 529000 ns
[  158.560635] [drm:lima_sched_timedout_job [lima]] *ERROR* lima job timeout
[  159.336829] lima 1c40000.gpu: fail to save task state from plasmashell pid 1850: error task list is full
[  159.336861] lima 1c40000.gpu: gp task error int_state=0 status=0
[  159.343220] lima 1c40000.gpu: gp soft reset time out
[  159.348913] lima 1c40000.gpu: gp soft reset time out
[  159.369856] lima 1c40000.gpu: gp soft reset time out
[  159.387419] lima 1c40000.gpu: gp soft reset time out
[  159.433901] lima 1c40000.gpu: gp soft reset time out
[  159.443988] lima 1c40000.gpu: gp soft reset time out

I think this is regression from either kernel patches for the lima runtime powermanagement or totally different thing but at least I could not reproduce this issue few months ago. I will try to reproduce this and provide more information in coming days.