[xen master] x86: Always have CR4.PKE set in HVM context

[patchbot@xxxxxxx]

commit 95aaafb5e2a09cad2ab42223e4aa55f88833ef3c
Author:     Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
AuthorDate: Thu Apr 29 14:28:43 2021 +0100
Commit:     Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
CommitDate: Fri Apr 30 17:07:29 2021 +0100

    x86: Always have CR4.PKE set in HVM context
    
    The sole user of read_pkru() is the emulated pagewalk, and guarded behind
    guest_pku_enabled() which restricts the path to HVM (hap, even) context 
only.
    
    The commentary in read_pkru() concerning _PAGE_GNTTAB overlapping with
    _PAGE_PKEY_BITS is only applicable to PV guests.
    
    The context switch path, via write_ptbase() unconditionally writes CR4 on 
any
    context switch.
    
    Therefore, we can guarantee to separate CR4.PKE between PV and HVM context 
at
    no extra cost.  Set PKE in mmu_cr4_features on boot, so it becomes set in 
HVM
    context, and clear it in pv_make_cr4().
    
    Rename read_pkru() to rdpkru() now that it is a simple wrapper around the
    instruction.  This saves two CR4 writes on every pagewalk, which typically
    occur more than one per emulation.
    
    Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
    Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>
---
 xen/arch/x86/mm/guest_walk.c    |  2 +-
 xen/arch/x86/pv/domain.c        | 16 +++++++++++++++-
 xen/arch/x86/setup.c            |  3 +++
 xen/include/asm-x86/processor.h | 10 +---------
 4 files changed, 20 insertions(+), 11 deletions(-)

diff --git a/xen/arch/x86/mm/guest_walk.c b/xen/arch/x86/mm/guest_walk.c
index 1c601314f3..30d83cf1e0 100644
--- a/xen/arch/x86/mm/guest_walk.c
+++ b/xen/arch/x86/mm/guest_walk.c
@@ -416,7 +416,7 @@ guest_walk_tables(const struct vcpu *v, struct p2m_domain 
*p2m,
          guest_pku_enabled(v) )
     {
         unsigned int pkey = guest_l1e_get_pkey(gw->l1e);
-        unsigned int pkru = read_pkru();
+        unsigned int pkru = rdpkru();
 
         if ( read_pkru_ad(pkru, pkey) ||
              ((walk & PFEC_write_access) && read_pkru_wd(pkru, pkey) &&
diff --git a/xen/arch/x86/pv/domain.c b/xen/arch/x86/pv/domain.c
index f1cb92585e..6ad533183b 100644
--- a/xen/arch/x86/pv/domain.c
+++ b/xen/arch/x86/pv/domain.c
@@ -182,7 +182,21 @@ unsigned long pv_make_cr4(const struct vcpu *v)
 {
     const struct domain *d = v->domain;
     unsigned long cr4 = mmu_cr4_features &
-        ~(X86_CR4_PCIDE | X86_CR4_PGE | X86_CR4_TSD);
+        ~(X86_CR4_PCIDE | X86_CR4_PGE | X86_CR4_TSD | X86_CR4_PKE);
+
+    /*
+     * We want CR4.PKE set in HVM context when available, but don't support it
+     * in PV context at all.
+     *
+     * _PAGE_PKEY_BITS where previously software available PTE bits.  In
+     * principle, we could let an aware PV guest enable PKE.
+     *
+     * However, Xen uses _PAGE_GNTTAB in debug builds which overlaps with
+     * _PAGE_PKEY_BITS, and the ownership of (and eligibility to move)
+     * software PTE bits is not considered in the PV ABI at all.  For now,
+     * punt the problem to whichever unluckly person finds a compelling
+     * usecase for PKRU in PV guests.
+     */
 
     /*
      * PCIDE or PGE depends on the PCID/XPTI settings, but must not both be
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index f2dff2ae6a..8105dc36bb 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -1790,6 +1790,9 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     if ( boot_cpu_has(X86_FEATURE_FSGSBASE) )
         set_in_cr4(X86_CR4_FSGSBASE);
 
+    if ( boot_cpu_has(X86_FEATURE_PKU) )
+        set_in_cr4(X86_CR4_PKE);
+
     if ( opt_invpcid && cpu_has_invpcid )
         use_invpcid = true;
 
diff --git a/xen/include/asm-x86/processor.h b/xen/include/asm-x86/processor.h
index d5f467d245..d8d0dc8034 100644
--- a/xen/include/asm-x86/processor.h
+++ b/xen/include/asm-x86/processor.h
@@ -367,20 +367,12 @@ static always_inline void set_in_cr4 (unsigned long mask)
     write_cr4(read_cr4() | mask);
 }
 
-static inline unsigned int read_pkru(void)
+static inline unsigned int rdpkru(void)
 {
     unsigned int pkru;
-    unsigned long cr4 = read_cr4();
 
-    /*
-     * _PAGE_PKEY_BITS have a conflict with _PAGE_GNTTAB used by PV guests,
-     * so that X86_CR4_PKE  is disabled on hypervisor. To use RDPKRU, CR4.PKE
-     * gets temporarily enabled.
-     */
-    write_cr4(cr4 | X86_CR4_PKE);
     asm volatile (".byte 0x0f,0x01,0xee"
         : "=a" (pkru) : "c" (0) : "dx");
-    write_cr4(cr4);
 
     return pkru;
 }
--
generated by git-patchbot for /home/xen/git/xen.git#master