00:38 <hanetzer> so question. generally speaking, does `b .` increment lr/pc/sp?

00:59 <jn> lr and sp should be untouched by jumping to the same location; pc will effectively stay the same

01:14 <hanetzer> noted.

06:51 <milkylainen> Anyone bothered to dig into TF-A TBB?

09:22 <ajb-linaro> maz: does the VMM ever give any guidance to KVM on how to schedule vCPU threads on host cores or is it all left up to the kernel scheduler?

15:01 <maz> ajb-linaro: the vmm has the *freedom* to do so (it can set the thread affinity). KVM isn't involved in that.

15:01 <maz> ajb-linaro: there were some kvmtool patches to do so when dealing with heterogeneous PMUs.

15:02 <ajb-linaro> maz: but that is just using the usual thread affinity interfaces right? taskset and the like?

15:03 <ajb-linaro> maz: will the scheduler usually spreed vCPUs out if they are all fully loaded?

15:05 <maz> ajb-linaro: if you don't tell the scheduler anything else (using something like sched_setaffinity(), for example), then the threads are free to roam anywhere in the system.

15:05 <maz> and vcpus are just random threads as far as the scheduler is concerned.