Il 28/02/2013 14:32, Anthony Liguori ha scritto:
>> >> qemu -numa
>> >> node=1,cpus=0,cpus=1,cpus=2,cpus=3,cpus=8,cpus=9,cpus=10,cpus=11
> >
> > Let me pick up the baby you just threw out with the bathwater for you:
> >
> > qemu -numa node,nodeid=1,cpus=0-3,cpus=8-11
If you're okay with making '-' be special syntax, why are you not okay
with ':' being special syntax?
For example because another kind of string list could have a colon
inside, and that would in turn need escaping (in fact that's already the
case for guestfwd); repeating the key is a syntax that is easily
reusable (and indeed is already in use). Instead, the '-' is parsed
within the NUMA code, and is completely opaque to QemuOpts. The NUMA
code knows that the '-' will never need escaping, because it only deals
with positive integers.
A perhaps better question would have been "if you're okay with making
',' be special syntax, why not ':'". And the answer is that indeed
','
already brings some problems, but likely they outweight the advantages
of having say only a "-set" option. But adding a second escaped
character is already much more debatable in my opinion.
What I assume your proposing is making cpus be a string list and
then
parsing within the NUMA code. Why not do it all in QemuOpts core code?
What would the QemuOpts parsing code do? Do you have in mind bitmasks
as a first-class QemuOpts type? If so, that would be an argument in
favor of ':', but against the prototype patch you posted yesterday.
Paolo