On Wed, Apr 10, 2019 at 02:29:26PM -0700, Dave Hansen wrote:
> On 4/10/19 2:25 PM, Sebastian Andrzej Siewior wrote:
> >> static inline void write_pkru(u32 pkru)
> >> {
> >> if (boot_cpu_has(X86_FEATURE_OSPKE))
> >> - __write_pkru(pkru);
> >> + wrpkru(pkru);
> > I think if this is a simple
> >
> > 's@__write_pkru_ins@wrpkru@g'
> > 's@__read_pkru_ins@rdpkru@g'
> >
> > then it should work just fine and match what Dave asked for.
>
> I'm fine with it either way, fwiw.
Final version:
---
From: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
Date: Wed, 3 Apr 2019 18:41:41 +0200
Subject: [PATCH] x86/pkru: Provide *pkru() helpers
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Dave Hansen asked for __read_pkru() and __write_pkru() to be
symmetrical.
As part of the series __write_pkru() will read back the value and only
write it if it is different.
In order to make both functions symmetrical, move the function
containing only the opcode asm into a function called like the
instruction itself.
__write_pkru() will just invoke wrpkru() but in a follow-up patch will
also read back the value.
[ bp: Convert asm opcode wrapper names to rd/wrpkru(). ]
Suggested-by: Dave Hansen <dave.hansen@xxxxxxxxx>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
Signed-off-by: Borislav Petkov <bp@xxxxxxx>
Reviewed-by: Dave Hansen <dave.hansen@xxxxxxxxx>
Reviewed-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Andi Kleen <ak@xxxxxxxxxxxxxxx>
Cc: Andy Lutomirski <luto@xxxxxxxxxx>
Cc: "H. Peter Anvin" <hpa@xxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: "Jason A. Donenfeld" <Jason@xxxxxxxxx>
Cc: Joerg Roedel <jroedel@xxxxxxx>
Cc: Juergen Gross <jgross@xxxxxxxx>
Cc: "Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx>
Cc: kvm ML <kvm@xxxxxxxxxxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxx>
Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
Cc: "Radim Krčmář" <rkrcmar@xxxxxxxxxx>
Cc: Rik van Riel <riel@xxxxxxxxxxx>
Cc: x86-ml <x86@xxxxxxxxxx>
Link: https://lkml.kernel.org/r/20190403164156.19645-13-bigeasy@xxxxxxxxxxxxx
---
arch/x86/include/asm/pgtable.h | 2 +-
arch/x86/include/asm/special_insns.h | 12 +++++++++---
arch/x86/kvm/vmx/vmx.c | 2 +-
3 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 2779ace16d23..e8875ca75623 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -127,7 +127,7 @@ static inline int pte_dirty(pte_t pte)
static inline u32 read_pkru(void)
{
if (boot_cpu_has(X86_FEATURE_OSPKE))
- return __read_pkru();
+ return rdpkru();
return 0;
}
diff --git a/arch/x86/include/asm/special_insns.h b/arch/x86/include/asm/special_insns.h
index 43c029cdc3fe..34897e2b52c9 100644
--- a/arch/x86/include/asm/special_insns.h
+++ b/arch/x86/include/asm/special_insns.h
@@ -92,7 +92,7 @@ static inline void native_write_cr8(unsigned long val)
#endif
#ifdef CONFIG_X86_INTEL_MEMORY_PROTECTION_KEYS
-static inline u32 __read_pkru(void)
+static inline u32 rdpkru(void)
{
u32 ecx = 0;
u32 edx, pkru;
@@ -107,7 +107,7 @@ static inline u32 __read_pkru(void)
return pkru;
}
-static inline void __write_pkru(u32 pkru)
+static inline void wrpkru(u32 pkru)
{
u32 ecx = 0, edx = 0;
@@ -118,8 +118,14 @@ static inline void __write_pkru(u32 pkru)
asm volatile(".byte 0x0f,0x01,0xef\n\t"
: : "a" (pkru), "c"(ecx), "d"(edx));
}
+
+static inline void __write_pkru(u32 pkru)
+{
+ wrpkru(pkru);
+}
+
#else
-static inline u32 __read_pkru(void)
+static inline u32 rdpkru(void)
{
return 0;
}
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index ab432a930ae8..dd1e1eea4f0b 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -6501,7 +6501,7 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
*/
if (static_cpu_has(X86_FEATURE_PKU) &&
kvm_read_cr4_bits(vcpu, X86_CR4_PKE)) {
- vcpu->arch.pkru = __read_pkru();
+ vcpu->arch.pkru = rdpkru();
if (vcpu->arch.pkru != vmx->host_pkru)
__write_pkru(vmx->host_pkru);
}
--
2.21.0
--
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.
