linux-kernel - Re: [patch V2 50/58] x86/apic: Provide common init infrastructure

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 1 Aug 2023 09:08:17 +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 50/58] x86/apic: Provide common init infrastructure

On 01.08.23 08:49, Juergen Gross wrote:
> On 01.08.23 08:41, Thomas Gleixner wrote:
>> On Mon, Jul 31 2023 at 17:48, Thomas Gleixner wrote:
>>
>>> On Mon, Jul 31 2023 at 15:10, Juergen Gross wrote:
>>>> On 31.07.23 15:01, Thomas Gleixner wrote:
>>>>>>     apic_driver(xen_pv_apic);
>>>>>
>>>>> I wonder whether this explicit install is actually needed at all.
>>>>> Shouldn't the driver be installed via the APIC probing mechanism
>>>>> automagically?
>>>>
>>>> Only in case x86_init.irq.intr_mode_init is set appropriately. Today it is
>>>> a nop for Xen PV, but that can be changed. I'll have a look.
>>>
>>> You could simply set that callback to default_setup_apic_routing() and
>>> be done with it.
>>
>> Doesn't work because XEN overrides it already. So sure, lets just go
> 
> It is overriding it with x86_init_noop().
> 
>> with the solution you proposed. One more ugly or less in XEN/PV does not
>> really matter much :)
>>
>> Let me grab this and put it into the right position in the queue.
> 
> Wait a few minutes, please. I'm just about to test your suggestion.

Using the following diff on top of your patch is working fine:

diff --git a/arch/x86/xen/apic.c b/arch/x86/xen/apic.c
index 804a26b7c85e..d7743ba0212d 100644
--- a/arch/x86/xen/apic.c
+++ b/arch/x86/xen/apic.c
@@ -166,11 +166,6 @@ static void __init xen_apic_check(void)
  void __init xen_init_apic(void)
  {
         x86_apic_ops.io_apic_read = xen_io_apic_read;
-       /* On PV guests the APIC CPUID bit is disabled so none of the
-        * routines end up executing. */
-       if (!xen_initial_domain())
-               apic_install_driver(&xen_pv_apic);
-
         x86_platform.apic_post_init = xen_apic_check;
  }
  apic_driver(xen_pv_apic);
diff --git a/arch/x86/xen/enlighten_pv.c b/arch/x86/xen/enlighten_pv.c
index 93b658248d01..164e5be23a45 100644
--- a/arch/x86/xen/enlighten_pv.c
+++ b/arch/x86/xen/enlighten_pv.c
@@ -1326,7 +1326,7 @@ asmlinkage __visible void __init xen_start_kernel(struct 
start_info *si)

         x86_init.resources.memory_setup = xen_memory_setup;
         x86_init.irqs.intr_mode_select  = x86_init_noop;
-       x86_init.irqs.intr_mode_init    = x86_init_noop;
+       x86_init.irqs.intr_mode_init    = x86_64_probe_apic;
         x86_init.oem.arch_setup = xen_arch_setup;
         x86_init.oem.banner = xen_banner;
         x86_init.hyper.init_platform = xen_pv_init_platform;


Juergen

Download attachment "OpenPGP_0xB0DE9DD628BF132F.asc" of type "application/pgp-keys" (3099 bytes)

Download attachment "OpenPGP_signature" of type "application/pgp-signature" (496 bytes)