Re: [PATCH v4 for 4.23] Add GIC SGI boot/self tests in Xen

["Grall, Julien" <julien@xxxxxxx>]

Hi Ayan,

On 17/09/2026 14:13, Halder, Ayan Kumar wrote:
> On 17/09/2026 08:57, Grall, Julien wrote:
> > On 16/09/2026 12:35, Ayan Kumar Halder wrote:
> > > +static void __init expect_sgi(const cpumask_t *mask,
> > > +                              const unsigned int *before, const char 
> > > *what)
> > > +{
> > > +    s_time_t deadline = NOW() + MILLISECS(100);
> > > +    unsigned int cpu;
> > > +
> > > +    for_each_cpu ( cpu, mask )
> > > +    {
> > > +        while ( sgi_count(cpu) == before[cpu] )
> > > +        {
> > > +            if ( NOW() > deadline )
> > > +                panic("GIC selftest: %s: CPU%u did not receive 
> > > GIC_SGI_TEST\n",
> > > +                      what, cpu);
> > > +            cpu_relax();
> > > +        }
> > > +    }
> > > +
> > > +    printk("GIC selftest: CPU%u: %s: OK\n", smp_processor_id(), what);
> > > +}
> > > +
> > > +/*
> > > + * "All but self" is only meaningful once every CPU can take an SGI, 
> > > so it is
> > > + * run by whichever CPU observes that it is the last one to get here.
> > > + */
> > > +static int __init gic_sgi_selftest(void)
> > > +{
> > > +    static atomic_t __initdata seen = ATOMIC_INIT(0);
> > > +    unsigned int before[NR_CPUS] = { };
> > Sorry I didn't spot this earlier. NR_CPUS can be quite large (up to 
> > 16K). So this will blow up the stack.
We discussed this offline at Xen Summit. But I will answer here as well.

> As this is a test, we have hardcoded NR_CPUS = 4 in automation/scripts/ 
> qemu-boot-selftest-arm64.sh.
I agree with your current test harness we only test 4 CPUs. However, 
this is not enforced by the Kconfig in Xen. So anyone could modify the 
script (or write their own) with a much higher number of NR_CPUS 
(possibly to match the number of pCPUs on their hardware).
Depending on the value, they could face a stack overflow.

> If there are 16K CPUs, our test only checks for 4 cpus only.
>
> However if there is still a concern ....
I would preferred if this is solved. Assuming this is fixed:

Reviewed-by: Julien Grall <julien@xxxxxxx>

Cheers,

--
Julien Grall