[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [patch 00/37] cpu/hotplug, x86: Reworked parallel CPU bringup
- To: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
- From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
- Date: Mon, 17 Apr 2023 12:30:50 +0200
- Cc: LKML <linux-kernel@xxxxxxxxxxxxxxx>, x86@xxxxxxxxxx, David Woodhouse <dwmw@xxxxxxxxxxxxx>, Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Brian Gerst <brgerst@xxxxxxxxx>, Arjan van de Veen <arjan@xxxxxxxxxxxxxxx>, Paolo Bonzini <pbonzini@xxxxxxxxxx>, Paul McKenney <paulmck@xxxxxxxxxx>, Tom Lendacky <thomas.lendacky@xxxxxxx>, Sean Christopherson <seanjc@xxxxxxxxxx>, Oleksandr Natalenko <oleksandr@xxxxxxxxxxxxxx>, Paul Menzel <pmenzel@xxxxxxxxxxxxx>, "Guilherme G. Piccoli" <gpiccoli@xxxxxxxxxx>, Piotr Gorski <lucjan.lucjanov@xxxxxxxxx>, David Woodhouse <dwmw@xxxxxxxxxxxx>, Usama Arif <usama.arif@xxxxxxxxxxxxx>, Juergen Gross <jgross@xxxxxxxx>, Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx, Russell King <linux@xxxxxxxxxxxxxxx>, Arnd Bergmann <arnd@xxxxxxxx>, linux-arm-kernel@xxxxxxxxxxxxxxxxxxx, Catalin Marinas <catalin.marinas@xxxxxxx>, Will Deacon <will@xxxxxxxxxx>, Guo Ren <guoren@xxxxxxxxxx>, linux-csky@xxxxxxxxxxxxxxx, Thomas Bogendoerfer <tsbogend@xxxxxxxxxxxxxxxx>, linux-mips@xxxxxxxxxxxxxxx, "James E.J. Bottomley" <James.Bottomley@xxxxxxxxxxxxxxxxxxxxx>, Helge Deller <deller@xxxxxx>, linux-parisc@xxxxxxxxxxxxxxx, Paul Walmsley <paul.walmsley@xxxxxxxxxx>, Palmer Dabbelt <palmer@xxxxxxxxxxx>, linux-riscv@xxxxxxxxxxxxxxxxxxx, Mark Rutland <mark.rutland@xxxxxxx>, Sabin Rapan <sabrapan@xxxxxxxxxx>
- Delivery-date: Mon, 17 Apr 2023 10:32:30 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
On Sat, Apr 15, 2023 at 01:44:13AM +0200, Thomas Gleixner wrote:
> Background
> ----------
>
> The reason why people are interested in parallel bringup is to shorten
> the (kexec) reboot time of cloud servers to reduce the downtime of the
> VM tenants. There are obviously other interesting use cases for this
> like VM startup time, embedded devices...
...
> There are two issue there:
>
> a) The death by MCE broadcast problem
>
> Quite some (contemporary) x86 CPU generations are affected by
> this:
>
> - MCE can be broadcasted to all CPUs and not only issued locally
> to the CPU which triggered it.
>
> - Any CPU which has CR4.MCE == 0, even if it sits in a wait
> for INIT/SIPI state, will cause an immediate shutdown of the
> machine if a broadcasted MCE is delivered.
When doing kexec, CR4.MCE should already have been set to 1 by the prior
kernel, no?
|
