File xsa435-0-34.patch of Package xen.35285

Overview Repositories Revisions Requests Users Attributes Meta

File xsa435-0-34.patch of Package xen.35285

From ef1987fcb0fdfaa7ee148024037cb5fa335a7b2d Mon Sep 17 00:00:00 2001
From: Andrew Cooper <andrew.cooper3@citrix.com>
Date: Fri, 12 May 2023 13:52:39 +0100
Subject: x86/boot: Rework dom0 feature configuration

Right now, dom0's feature configuration is split between between the common
path and a dom0-specific one.  This mostly is by accident, and causes some
very subtle bugs.

First, start by clearly defining init_dom0_cpuid_policy() to be the domain
that Xen builds automatically.  The late hwdom case is still constructed in a
mostly normal way, with the control domain having full discretion over the CPU
policy.

Identifying this highlights a latent bug - the two halves of the MSR_ARCH_CAPS
bodge are asymmetric with respect to the hardware domain.  This means that
shim, or a control-only dom0 sees the MSR_ARCH_CAPS CPUID bit but none of the
MSR content.  This in turn declares the hardware to be retpoline-safe by
failing to advertise the {R,}RSBA bits appropriately.  Restrict this logic to
the hardware domain, although the special case will cease to exist shortly.

For the CPUID Faulting adjustment, the comment in ctxt_switch_levelling()
isn't actually relevant.  Provide a better explanation.

Move the recalculate_cpuid_policy() call outside of the dom0-cpuid= case.
This is no change for now, but will become necessary shortly.

Finally, place the second half of the MSR_ARCH_CAPS bodge after the
recalculate_cpuid_policy() call.  This is necessary to avoid transiently
breaking the hardware domain's view while the handling is cleaned up.  This
special case will cease to exist shortly.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/cpu-policy.c
+++ b/xen/arch/x86/cpu-policy.c
@@ -549,29 +549,6 @@ int init_domain_cpu_policy(struct domain
     if ( !p )
         return -ENOMEM;
 
-    /* See comment in ctxt_switch_levelling() */
-    if ( !opt_dom0_cpuid_faulting && is_control_domain(d) && is_pv_domain(d) )
-        p->platform_info.cpuid_faulting = false;
-
-    /*
-     * Expose the "hardware speculation behaviour" bits of ARCH_CAPS to dom0,
-     * so dom0 can turn off workarounds as appropriate.  Temporary, until the
-     * domain policy logic gains a better understanding of MSRs.
-     */
-    if ( is_hardware_domain(d) && cpu_has_arch_caps )
-    {
-        uint64_t val;
-
-        rdmsrl(MSR_ARCH_CAPABILITIES, val);
-
-        p->arch_caps.raw = val &
-            (ARCH_CAPS_RDCL_NO | ARCH_CAPS_IBRS_ALL | ARCH_CAPS_RSBA |
-             ARCH_CAPS_SSB_NO | ARCH_CAPS_MDS_NO | ARCH_CAPS_IF_PSCHANGE_MC_NO |
-             ARCH_CAPS_TAA_NO | ARCH_CAPS_SBDR_SSDP_NO | ARCH_CAPS_FBSDP_NO |
-             ARCH_CAPS_PSDP_NO | ARCH_CAPS_FB_CLEAR | ARCH_CAPS_RRSBA |
-             ARCH_CAPS_BHI_NO | ARCH_CAPS_PBRSB_NO);
-    }
-
     d->arch.cpu_policy = p;
 
     recalculate_cpuid_policy(d);
@@ -729,11 +706,15 @@ void recalculate_cpuid_policy(struct dom
         p->extd.raw[0x19] = EMPTY_LEAF;
 }
 
+/*
+ * Adjust the CPU policy for dom0.  Really, this is "the domain Xen builds
+ * automatically on boot", and might not have the domid 0 (e.g. pvshim).
+ */
 void __init init_dom0_cpuid_policy(struct domain *d)
 {
     struct cpu_policy *p = d->arch.cpuid;
 
-    /* dom0 can't migrate.  Give it ITSC if available. */
+    /* Dom0 doesn't migrate relative to Xen.  Give it ITSC if available. */
     if ( cpu_has_itsc )
         p->extd.itsc = true;
 
@@ -742,7 +723,7 @@ void __init init_dom0_cpuid_policy(struc
      * so dom0 can turn off workarounds as appropriate.  Temporary, until the
      * domain policy logic gains a better understanding of MSRs.
      */
-    if ( cpu_has_arch_caps )
+    if ( is_hardware_domain(d) && cpu_has_arch_caps )
         p->feat.arch_caps = true;
 
     /* Apply dom0-cpuid= command line settings, if provided. */
@@ -760,8 +741,32 @@ void __init init_dom0_cpuid_policy(struc
         }
 
         x86_cpu_featureset_to_policy(fs, p);
+    }
+
+    /*
+     * PV Control domains used to require unfiltered CPUID.  This was fixed in
+     * Xen 4.13, but there is an cmdline knob to restore the prior behaviour.
+     *
+     * If the domain is getting unfiltered CPUID, don't let the guest kernel
+     * play with CPUID faulting either, as Xen's CPUID path won't cope.
+     */
+    if ( !opt_dom0_cpuid_faulting && is_control_domain(d) && is_pv_domain(d) )
+        p->platform_info.cpuid_faulting = false;
+
+    recalculate_cpuid_policy(d);
+
+    if ( is_hardware_domain(d) && cpu_has_arch_caps )
+    {
+        uint64_t val;
 
-        recalculate_cpuid_policy(d);
+        rdmsrl(MSR_ARCH_CAPABILITIES, val);
+
+        p->arch_caps.raw = val &
+            (ARCH_CAPS_RDCL_NO | ARCH_CAPS_IBRS_ALL | ARCH_CAPS_RSBA |
+             ARCH_CAPS_SSB_NO | ARCH_CAPS_MDS_NO | ARCH_CAPS_IF_PSCHANGE_MC_NO |
+             ARCH_CAPS_TAA_NO | ARCH_CAPS_SBDR_SSDP_NO | ARCH_CAPS_FBSDP_NO |
+             ARCH_CAPS_PSDP_NO | ARCH_CAPS_FB_CLEAR | ARCH_CAPS_RRSBA |
+             ARCH_CAPS_BHI_NO | ARCH_CAPS_PBRSB_NO);
     }
 }

Places

File xsa435-0-34.patch of Package xen.35285

Places