On Fri, Jul 16, 2021 at 10:54:08AM +0200, Cornelia Huck wrote:
> On Wed, Jul 14 2021, Pierre Morel <pmo...@linux.ibm.com> wrote:
>
> > We need a s390x dedicated SMP parsing to handle s390x specificities.
> >
> > In this patch we only handle threads, cores and sockets for
> > s390x:
> > - do not support threads, we always have 1 single thread per core
> > - the sockets are filled one after the other with the cores
> >
> > Both these handlings are different from the standard smp_parse
> > functionement and reflect the CPU topology in the simple case
> > where all CPU belong to the same book.
> >
> > Topology levels above sockets, i.e. books, drawers, are not
> > considered at this stage and will be introduced in a later patch.
> >
> > Signed-off-by: Pierre Morel <pmo...@linux.ibm.com>
> > ---
> > hw/s390x/s390-virtio-ccw.c | 42 ++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 42 insertions(+)
> >
> > diff --git a/hw/s390x/s390-virtio-ccw.c b/hw/s390x/s390-virtio-ccw.c
> > index e4b18aef49..899d3a4137 100644
> > --- a/hw/s390x/s390-virtio-ccw.c
> > +++ b/hw/s390x/s390-virtio-ccw.c
> > @@ -582,6 +582,47 @@ static ram_addr_t s390_fixup_ram_size(ram_addr_t sz)
> > return newsz;
> > }
> >
> > +/*
> > + * In S390CCW machine we do not support threads for now,
> > + * only sockets and cores.
> > + */
> > +static void s390_smp_parse(MachineState *ms, QemuOpts *opts)
>
> It seems you based this on an older version of the code? The current
> signature of this function since 1e63fe685804 ("machine: pass QAPI
> struct to mc->smp_parse") is
>
> void (*smp_parse)(MachineState *ms, SMPConfiguration *config, Error **errp);
>
> That affects your parsing, and also lets you get rid of the ugly exit(1)
> statements.
>
> > +{
> > + unsigned cpus = qemu_opt_get_number(opts, "cpus", 1);
> > + unsigned sockets = qemu_opt_get_number(opts, "sockets", 1);
> > + unsigned cores = qemu_opt_get_number(opts, "cores", 1);
> > +
> > + if (opts) {
> > + if (cpus == 0 || sockets == 0 || cores == 0) {
>
> This behaviour looks different from what we do for other targets: if you
> specify the value as 0, a value is calculated from the other values;
> here, you error out. It's probably not a good idea to differ.
I increasingly worry that we're making a mistake by going down the
route of having custom smp_parse implementations per target, as this
is showing signs of inconsistent behaviour and error reportings. I
think the differences / restrictions have granularity at a different
level that is being tested in many cases too.
Whether threads != 1 is valid will likely vary depending on what
CPU model is chosen, rather than what architecture is chosen.
The same is true for dies != 1. We're not really checking this
closely even in x86 - for example I can request nonsense such
as a 25 year old i486 CPU model with hyperthreading and multiple
dies
qemu-system-x86_64 -cpu 486 -smp 16,cores=4,dies=2,threads=2
In this patch, there is no error reporting if the user specifies
dies != 1 or threads != 1 - it just silently ignores the request
which is not good.
Some machine types may have constraints on CPU sockets.
This can of course all be handled by custom smp_parse impls, but
this is ultimately going to lead to alot of duplicated and
inconsistent logic I fear.
I wonder if we would be better off having machine class callback
that can report topology constraints for the current configuration,
along lines of
smp_constraints(MachineState *ms,
int *max_sockets,
int *max_dies,
int *max_cores,
int *max_threads)
And then have only a single smp_parse impl that takes into account
these constraints, to report errors / fill in missing fields / etc ?
Regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
