[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4bf0020b-dff9-091f-dcb4-45f1ee864e2a@suse.com>
Date: Mon, 31 Jul 2023 20:19:25 +0200
From: Juergen Gross <jgross@...e.com>
To: Thomas Gleixner <tglx@...utronix.de>,
LKML <linux-kernel@...r.kernel.org>
Cc: x86@...nel.org, Andrew Cooper <andrew.cooper3@...rix.com>,
Tom Lendacky <thomas.lendacky@....com>,
Paolo Bonzini <pbonzini@...hat.com>,
Wei Liu <wei.liu@...nel.org>,
Arjan van de Ven <arjan@...ux.intel.com>,
Michael Kelley <mikelley@...rosoft.com>,
Peter Keresztes Schmidt <peter@...esztesschmidt.de>,
"Peter Zijlstra (Intel)" <peterz@...radead.org>
Subject: Re: [patch V2 16/58] x86/apic: Sanitize num_processors handling
On 31.07.23 17:57, Thomas Gleixner wrote:
> On Mon, Jul 31 2023 at 14:50, Thomas Gleixner wrote:
>> Why can't XENPV pretend that it has a smp configuration detected,
>> i.e. setting smp_found_config as any other special get_smp_config()
>> implementation does?
>
> The below should do the trick, no?
Something like that, yes.
I'm just hunting another regression in the series. With patch 23 of the
topology series applied the APs of a Xen PV guests won't be onlined. I
guess this is due to missing topology data initialization somewhere in
the Xen related code.
I'll check your suggestion after finding the reason for the regression.
Juergen
>
>
> --- a/arch/x86/xen/smp_pv.c
> +++ b/arch/x86/xen/smp_pv.c
> @@ -182,7 +182,8 @@ static void __init _get_smp_config(unsig
> if (subtract)
> set_nr_cpu_ids(nr_cpu_ids - subtract);
> #endif
> -
> + /* Pretend to be a proper enumerated system */
> + smp_found_config = 1;
> }
>
> static void __init xen_pv_smp_prepare_boot_cpu(void)
Download attachment "OpenPGP_0xB0DE9DD628BF132F.asc" of type "application/pgp-keys" (3099 bytes)
Download attachment "OpenPGP_signature" of type "application/pgp-signature" (496 bytes)
Powered by blists - more mailing lists