Re: [Qemu-devel] Meaning of "-smp threads" on mips_malta

[Aleksandar Markovic]

On Monday, January 14, 2019, Eduardo Habkost <address@hidden> wrote:

> Hi,
>
> I'm trying to refactor the SMP topology code in QEMU


>
Eduardo, I truly appreciate your interest in details of Malta
implementations, but before I answer your questions, could you please
outline the motivation and the current concept of your envisioned
refactoring of SMP, so that I have some picture of the whole context?

Sincerely, Aleksandar



>
> and I found
> some suspicious code on mips_malta.c:
>
> static void malta_mips_config(MIPSCPU *cpu)
> {
>     CPUMIPSState *env = &cpu->env;
>     CPUState *cs = CPU(cpu);
>
>     env->mvp->CP0_MVPConf0 |= ((smp_cpus - 1) << CP0MVPC0_PVPE) |
>                          ((smp_cpus * cs->nr_threads - 1) << CP0MVPC0_PTC);
> }
>
>
> The (smp_cpus * cs->nr_threads) expression here doesn't make
> sense to me (because smp_cpus is already supposed to be a
> multiple of smp_threads), and seems to indicate that the code has
> some unusual assumptions about the semantics of the -smp option.
>
> So, I'd like to know: do all the examples below make sense for
> Malta?
>
>  -smp 1
>  -smp 2
>  -smp 2,threads=1
>  -smp 2,threads=2
>  -smp 1,threads=2 [*]
>  -smp 2,threads=3 [*]
>
> The generic -smp parsing code considers the last 2 entries
> above[*] to be invalid.  If they make sense for Malta, we need to
> find a way to fix that.  Replacing "-smp threads=..." with a
> "-cpu" or "-machine" option seems like the best alternative.
>
> --
> Eduardo
>
>