Zachary Amsden wrote:
> The calls in general, I think should be C-like in their properties -
> having EAX, EDX, ECX for local use, and clobbering flags and memory
> implicitly by nature of being a (non-pure) function call.
Yes, that's the plan. We'll declare all the pv_ops functions as
"fastcall" to mean regparm(3), so they'll all be expected to use/clobber
eax, edx and ecx. For critical operations which are likely to be always
inlined in a real implementation, we can define the operation as only
clobbering a subset of those registers, leaving the burden of
saving/restoring the other registers to the less optimal sequences which
do actually make calls.
Andi, is there some way we can get the CFI stuff right for those
pushes/pops?
> I would highly recommend using ECX or EDX. The choice of ESI being
> unused is really arbitrary, and it makes it incredibly difficult to
> try to express passing a VCPU data reference in ESI (that is what it
> is used for, right?) to a C function. In several places, entry.S
> calls out to C code and expects EAX, EDX, and ECX to come back
> clobbered, so there are wide windows there in which the registers are
> not live, whereas the code could quite reasonably expect ESI to be
> preserved across those boundaries.
I think you're right; it shouldn't be too hard to do a bit of register
renaming in entry.S to make things work out nicely.
J
