Re: [PATCH v2 2/2] x86/sev-es: Only set x86_virt_bits to correct value

[Dave Hansen]

On 10/2/23 13:04, Nathan Chancellor wrote:
> On Mon, Sep 11, 2023 at 05:27:03PM -0700, Adam Dunlap wrote:
>> Instead of setting x86_virt_bits to a possibly-correct value and then
>> correcting it later, do all the necessary checks before setting it.
>>
>> At this point, the #VC handler references boot_cpu_data.x86_virt_bits,
>> and in the previous version, it would be triggered by the cpuids between
>> the point at which it is set to 48 and when it is set to the correct
>> value.
>>
>> Suggested-by: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
>> Signed-off-by: Adam Dunlap <acdunlap@xxxxxxxxxx>
> Our continuous integration started seeing panics when booting ARCH=i386
> without KVM after this change landed in -tip as commit fbf6449f84bf
> ("x86/sev-es: Set x86_virt_bits to the correct value straight away,
> instead of a two-phase approach"):

I can't reproduce this, but I'm running a gcc-built kernel and I haven't
tried very hard to replicate your qemu setup.

I did notice, though, that the patch in question forgot to move one
assignment.  Could you see if the attached patch fixes things for you?diff -puN arch/x86/kernel/cpu/common.c~cache-bits-debug arch/x86/kernel/cpu/common.c
--- a/arch/x86/kernel/cpu/common.c~cache-bits-debug	2023-10-02 14:20:48.273415423 -0700
+++ b/arch/x86/kernel/cpu/common.c	2023-10-02 14:24:52.710129051 -0700
@@ -1142,6 +1142,7 @@ void get_cpu_address_sizes(struct cpuinf
 		}
 	}
 	c->x86_cache_bits = c->x86_phys_bits;
+	c->x86_cache_alignment = c->x86_clflush_size;
 }
 
 static void identify_cpu_without_cpuid(struct cpuinfo_x86 *c)
@@ -1595,8 +1596,6 @@ static void __init cpu_parse_early_param
  */
 static void __init early_identify_cpu(struct cpuinfo_x86 *c)
 {
-	c->x86_cache_alignment = c->x86_clflush_size;
-
 	memset(&c->x86_capability, 0, sizeof(c->x86_capability));
 	c->extended_cpuid_level = 0;
 
_