At the first (and maybe second) glance the call to setup_clear_cpu_cap()
(which is __init) from psr_cpu_init() (which isn't) looks wrong. The
earlier cpu_has_pqe makes it safe, though. Nevertheless we can do better,
by simply moving the check ahead of the BSP invocation of the function.

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/psr.c
+++ b/xen/arch/x86/psr.c
@@ -1583,12 +1583,6 @@ static void psr_cpu_init(void)
     if ( !psr_alloc_feat_enabled() || !cpu_has_pqe )
         goto assoc_init;
 
-    if ( boot_cpu_data.cpuid_level < PSR_CPUID_LEVEL_CAT )
-    {
-        setup_clear_cpu_cap(X86_FEATURE_PQE);
-        goto assoc_init;
-    }
-
     socket = cpu_to_socket(cpu);
     info = socket_info + socket;
     if ( info->feat_init )
@@ -1708,6 +1702,9 @@ static int __init cf_check psr_presmp_in
     if ( psr_cpu_prepare() )
         psr_free();
 
+    if ( boot_cpu_data.cpuid_level < PSR_CPUID_LEVEL_CAT )
+        setup_clear_cpu_cap(X86_FEATURE_PQE);
+
     psr_cpu_init();
     if ( psr_cmt_enabled() || psr_alloc_feat_enabled() )
         register_cpu_notifier(&cpu_nfb);

On 27/08/2025 9:16 am, Jan Beulich wrote:
> At the first (and maybe second) glance the call to setup_clear_cpu_cap()
> (which is __init) from psr_cpu_init() (which isn't) looks wrong. The
> earlier cpu_has_pqe makes it safe, though. Nevertheless we can do better,
> by simply moving the check ahead of the BSP invocation of the function.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>

This is still terribly convoluted code.