[tip:mm/pkeys] x86/fpu: Add placeholder for 'Processor Trace' XSAVE state

From: tip-bot for Dave Hansen
Date: Thu Feb 18 2016 - 15:17:41 EST


Commit-ID: 1f96b1efbad4bb753e7fd265753f6cac1cdc5648
Gitweb: http://git.kernel.org/tip/1f96b1efbad4bb753e7fd265753f6cac1cdc5648
Author: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
AuthorDate: Fri, 12 Feb 2016 13:01:58 -0800
Committer: Ingo Molnar <mingo@xxxxxxxxxx>
CommitDate: Tue, 16 Feb 2016 10:11:13 +0100

x86/fpu: Add placeholder for 'Processor Trace' XSAVE state

There is an XSAVE state component for Intel Processor Trace (PT).
But, we do not currently use it.

We add a placeholder in the code for it so it is not a mystery and
also so we do not need an explicit enum initialization for Protection
Keys in a moment.

Why don't we use it?

We might end up using this at _some_ point in the future. But,
this is a "system" state which requires using the currently
unsupported XSAVES feature. Unlike all the other XSAVE states,
PT state is also not directly tied to a thread. You might
context-switch between threads, but not want to change any of the
PT state. Or, you might switch between threads, and *do* want to
change PT state, all depending on what is being traced.

We currently just manually set some MSRs to do this PT context
switching, and it is unclear whether replacing our direct MSR use
with XSAVE will be a net win or loss, both in code complexity and
performance.

Signed-off-by: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
Reviewed-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: Andy Lutomirski <luto@xxxxxxxxxxxxxx>
Cc: Borislav Petkov <bp@xxxxxxxxx>
Cc: Brian Gerst <brgerst@xxxxxxxxx>
Cc: Dave Hansen <dave@xxxxxxxx>
Cc: Denys Vlasenko <dvlasenk@xxxxxxxxxx>
Cc: H. Peter Anvin <hpa@xxxxxxxxx>
Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Rik van Riel <riel@xxxxxxxxxx>
Cc: fenghua.yu@xxxxxxxxx
Cc: linux-mm@xxxxxxxxx
Cc: yu-cheng.yu@xxxxxxxxx
Link: http://lkml.kernel.org/r/20160212210158.5E4BCAE2@xxxxxxxxxxxxxxxxxx
Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
---
arch/x86/include/asm/fpu/types.h | 1 +
arch/x86/kernel/fpu/xstate.c | 10 ++++++++--
2 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/arch/x86/include/asm/fpu/types.h b/arch/x86/include/asm/fpu/types.h
index 1c6f6ac..aad3181 100644
--- a/arch/x86/include/asm/fpu/types.h
+++ b/arch/x86/include/asm/fpu/types.h
@@ -108,6 +108,7 @@ enum xfeature {
XFEATURE_OPMASK,
XFEATURE_ZMM_Hi256,
XFEATURE_Hi16_ZMM,
+ XFEATURE_PT_UNIMPLEMENTED_SO_FAR,

XFEATURE_MAX,
};
diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
index d425cda5..c2e2349 100644
--- a/arch/x86/kernel/fpu/xstate.c
+++ b/arch/x86/kernel/fpu/xstate.c
@@ -13,6 +13,11 @@

#include <asm/tlbflush.h>

+/*
+ * Although we spell it out in here, the Processor Trace
+ * xfeature is completely unused. We use other mechanisms
+ * to save/restore PT state in Linux.
+ */
static const char *xfeature_names[] =
{
"x87 floating point registers" ,
@@ -23,7 +28,7 @@ static const char *xfeature_names[] =
"AVX-512 opmask" ,
"AVX-512 Hi256" ,
"AVX-512 ZMM_Hi256" ,
- "unknown xstate feature" ,
+ "Processor Trace (unused)" ,
};

/*
@@ -470,7 +475,8 @@ static void check_xstate_against_struct(int nr)
* numbers.
*/
if ((nr < XFEATURE_YMM) ||
- (nr >= XFEATURE_MAX)) {
+ (nr >= XFEATURE_MAX) ||
+ (nr == XFEATURE_PT_UNIMPLEMENTED_SO_FAR)) {
WARN_ONCE(1, "no structure for xstate: %d\n", nr);
XSTATE_WARN_ON(1);
}