[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Z748ZaNc9IclOUPi@gmail.com>
Date: Tue, 25 Feb 2025 22:55:49 +0100
From: Ingo Molnar <mingo@...nel.org>
To: Jinyu Tang <tjytimi@....com>
Cc: Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>,
x86@...nel.org, "H . Peter Anvin" <hpa@...or.com>,
Will Deacon <will@...nel.org>, Anup Patel <apatel@...tanamicro.com>,
Julien Thierry <julien.thierry.kdev@...il.com>, kvm@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH RFC] x86: avoid one cpu limit for kvm
* Jinyu Tang <tjytimi@....com> wrote:
> I run kernel by kvmtool but only one cpu can start in guset now,
> because kvm use virt-ioapic and kvmtool set noapic cmdline for
> x86 to disable ioapic route in kernel, and the latest cpu topo
> code below makes cpu limitted just to one.
>
> For x86 kvm, noapic cmdline is reasonable, virt-ioapic don't
> need to init hardware ioapic, so change it for x86 kvm guest to
> avoid one num limit.
>
> Signed-off-by: Jinyu Tang <tjytimi@....com>
> ---
> arch/x86/kernel/cpu/topology.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/topology.c b/arch/x86/kernel/cpu/topology.c
> index 621a151ccf7d..a73847b1a841 100644
> --- a/arch/x86/kernel/cpu/topology.c
> +++ b/arch/x86/kernel/cpu/topology.c
> @@ -429,7 +429,9 @@ void __init topology_apply_cmdline_limits_early(void)
> unsigned int possible = nr_cpu_ids;
>
> /* 'maxcpus=0' 'nosmp' 'nolapic' 'disableapic' 'noapic' */
> - if (!setup_max_cpus || ioapic_is_disabled || apic_is_disabled)
> + if (!setup_max_cpus ||
> + (ioapic_is_disabled && (x86_hyper_type != X86_HYPER_KVM)) ||
> + apic_is_disabled)
> possible = 1;
>
> /* 'possible_cpus=N' */
> @@ -443,8 +445,10 @@ void __init topology_apply_cmdline_limits_early(void)
>
> static __init bool restrict_to_up(void)
> {
> - if (!smp_found_config || ioapic_is_disabled)
> + if (!smp_found_config ||
> + (ioapic_is_disabled && (x86_hyper_type != X86_HYPER_KVM)))
> return true;
> +
> /*
> * XEN PV is special as it does not advertise the local APIC
> * properly, but provides a fake topology for it so that the
Just to confirm that this bug got fixed, current kvmtool boots fine
with multiple CPUs, correct?
Thanks,
Ingo
Powered by blists - more mailing lists