Re: [PATCH 5/8] x86: remove dependencies on CONFIG_M... CPU options

[David Laight <david.laight.linux@gmail.com>]

On Fri, 22 May 2026 21:00:00 +0200
Jürgen Groß <jgross@suse.com> wrote:

> On 22.05.26 20:54, Arnd Bergmann wrote:
> > On Fri, May 22, 2026, at 17:46, Juergen Gross wrote:  
> >> On 22.05.26 16:19, Arnd Bergmann wrote:  
> >>>
> >>> --- a/arch/x86/xen/Kconfig
> >>> +++ b/arch/x86/xen/Kconfig
> >>> @@ -10,8 +10,7 @@ config XEN
> >>>    	select X86_HV_CALLBACK_VECTOR
> >>>    	select HIBERNATE_CALLBACKS
> >>>    	depends on X86_64 || (X86_32 && X86_PAE)
> >>> -	depends on X86_64 || (X86_GENERIC || MPENTIUM4 || MATOM)
> >>> -	depends on X86_LOCAL_APIC && X86_TSC
> >>> +	depends on X86_LOCAL_APIC  
> >>
> >> This is problematic. See commit 93cd05976498.
> >>
> >> Maybe use "depends on X86_LOCAL_APIC && X86_L1_CACHE_SHIFT >= 6" instead?  
> > 
> > After patch 6/8, this is always true, so I suppose it would
> > if that came first, right?  
> 
> Yes, agreed.
> 
> > 
> > On a related note, are there still remaining Xen users on new
> > 32-bit guest kernels? I understand that one could still run
> > modern kernels on 32-bit Xen-4.2 (EOL in 2015), and that 64-bit
> > Xen still supports older 32-bit guest operating systems for
> > compatibility, but I'm not sure if there is any reason to
> > legitimately run a 32-bit linux-7.x kernel on a 64-bit Xen
> > instead of running a better supported 64-bit guest kernel.  
> 
>  From the logical point of view I totally agree. OTOH users are humans,
> so they might not agree with that. ;-)

You might just want to test a 32bit kernel ?

-- David

> 
> 
> Juergen