Re: [PATCH 1/2] x86: unify pt_regs accessors ptrace.h

Re: [PATCH 1/2] x86: unify pt_regs accessors ptrace.h

[H. Peter Anvin]

Harvey Harrison wrote:
> Unify the definiton of:
> v8086_mode
> user_mode
> user_mode_vm
> stack_pointer
> instruction_pointer
> frame_pointer
> 
> in ptrace.h to make it clear where the differences are between
> 32 and 64 bit.  Changes macros to static inlines as well.
> 

Can you put a comment on stack_pointer() that it applies to a 
kernel-mode pt_regs only?

	-hpa

[PATCH 1/2] x86: unify pt_regs accessors ptrace.h

[Harvey Harrison]

Unify the definiton of:
v8086_mode
user_mode
user_mode_vm
stack_pointer
instruction_pointer
frame_pointer

in ptrace.h to make it clear where the differences are between
32 and 64 bit.  Changes macros to static inlines as well.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
---
 include/asm-x86/ptrace.h |   84 +++++++++++++++++++++++++++++----------------
 1 files changed, 54 insertions(+), 30 deletions(-)

diff --git a/include/asm-x86/ptrace.h b/include/asm-x86/ptrace.h
index ee4b595..0ad9a2b 100644
--- a/include/asm-x86/ptrace.h
+++ b/include/asm-x86/ptrace.h
@@ -76,30 +76,6 @@ convert_ip_to_linear(struct task_struct *child, struct pt_regs *regs);
 
 extern void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code);
 
-/*
- * user_mode_vm(regs) determines whether a register set came from user mode.
- * This is true if V8086 mode was enabled OR if the register set was from
- * protected mode with RPL-3 CS value.  This tricky test checks that with
- * one comparison.  Many places in the kernel can bypass this full check
- * if they have already ruled out V8086 mode, so user_mode(regs) can be used.
- */
-static inline int user_mode(struct pt_regs *regs)
-{
-	return (regs->cs & SEGMENT_RPL_MASK) == USER_RPL;
-}
-static inline int user_mode_vm(struct pt_regs *regs)
-{
-	return ((regs->cs & SEGMENT_RPL_MASK) |
-		(regs->flags & VM_MASK)) >= USER_RPL;
-}
-static inline int v8086_mode(struct pt_regs *regs)
-{
-	return (regs->flags & VM_MASK);
-}
-
-#define instruction_pointer(regs) ((regs)->ip)
-#define frame_pointer(regs) ((regs)->bp)
-#define stack_pointer(regs) ((unsigned long)(regs))
 #define regs_return_value(regs) ((regs)->ax)
 
 extern unsigned long profile_pc(struct pt_regs *regs);
@@ -167,12 +143,6 @@ struct pt_regs {
 /* top of stack page */
 };
 
-#define user_mode(regs) (!!((regs)->cs & 3))
-#define user_mode_vm(regs) user_mode(regs)
-#define v8086_mode(regs) 0	/* No V86 mode support in long mode */
-#define instruction_pointer(regs) ((regs)->ip)
-#define frame_pointer(regs) ((regs)->bp)
-#define stack_pointer(regs) ((regs)->sp)
 #define regs_return_value(regs) ((regs)->ax)
 
 extern unsigned long profile_pc(struct pt_regs *regs);
@@ -189,6 +159,60 @@ convert_ip_to_linear(struct task_struct *child, struct pt_regs *regs);
 #ifdef __KERNEL__
 
 /*
+ * user_mode_vm(regs) determines whether a register set came from user mode.
+ * This is true if V8086 mode was enabled OR if the register set was from
+ * protected mode with RPL-3 CS value.  This tricky test checks that with
+ * one comparison.  Many places in the kernel can bypass this full check
+ * if they have already ruled out V8086 mode, so user_mode(regs) can be used.
+ */
+static inline int user_mode(struct pt_regs *regs)
+{
+#ifdef CONFIG_X86_32
+	return (regs->cs & SEGMENT_RPL_MASK) == USER_RPL;
+#else
+	return !!(regs->cs & 3);
+#endif
+}
+
+static inline int user_mode_vm(struct pt_regs *regs)
+{
+#ifdef CONFIG_X86_32
+	return ((regs->cs & SEGMENT_RPL_MASK) |
+		(regs->flags & VM_MASK)) >= USER_RPL;
+#else
+	return user_mode(regs);
+#endif
+}
+
+static inline int v8086_mode(struct pt_regs *regs)
+{
+#ifdef CONFIG_X86_32
+	return (regs->flags & VM_MASK);
+#else
+	return 0;	/* No V86 mode support in long mode */
+#endif
+}
+
+static inline unsigned long stack_pointer(struct pt_regs *regs)
+{
+#ifdef CONFIG_X86_32
+	return (unsigned long)regs;
+#else
+	return regs->sp;
+#endif
+}
+
+static inline unsigned long instruction_pointer(struct pt_regs *regs)
+{
+	return regs->ip;
+}
+
+static inline unsigned long frame_pointer(struct pt_regs *regs)
+{
+	return regs->bp;
+}
+
+/*
  * These are defined as per linux/ptrace.h, which see.
  */
 #define arch_has_single_step()	(1)
-- 
1.5.4.rc2.1164.g6451

Re: [PATCH 1/2] x86: unify pt_regs accessors ptrace.h

[H. Peter Anvin]

Harvey Harrison wrote:
> On Thu, 2008-01-17 at 20:56 -0500, H. Peter Anvin wrote:
>> Harvey Harrison wrote:
>>> Unify the definiton of:
>>> v8086_mode
>>> user_mode
>>> user_mode_vm
>>> stack_pointer
>>> instruction_pointer
>>> frame_pointer
>>>
>>> in ptrace.h to make it clear where the differences are between
>>> 32 and 64 bit.  Changes macros to static inlines as well.
>>>
>> Can you put a comment on stack_pointer() that it applies to a 
>> kernel-mode pt_regs only?
> 
> How about a follow-on patch that picks a better name and adds the
> comment?  I'd like to leave this as a movement-only patch.  Once we
> decide on a name I'll do the rename/comment.  It seems like there are
> very few users currently (oprofile/backtrace.c only?).
> 
> Then I'll also change the kprobes code to use this instead of their
> stack_addr helper.
> 

IMO, go ahead and add a comment, but yes, renaming should be a separate 
patch.

	-hpa

Re: [PATCH 1/2] x86: unify pt_regs accessors ptrace.h

[Harvey Harrison]

On Thu, 2008-01-17 at 20:56 -0500, H. Peter Anvin wrote:
> Harvey Harrison wrote:
> > Unify the definiton of:
> > v8086_mode
> > user_mode
> > user_mode_vm
> > stack_pointer
> > instruction_pointer
> > frame_pointer
> > 
> > in ptrace.h to make it clear where the differences are between
> > 32 and 64 bit.  Changes macros to static inlines as well.
> > 
> 
> Can you put a comment on stack_pointer() that it applies to a 
> kernel-mode pt_regs only?

How about a follow-on patch that picks a better name and adds the
comment?  I'd like to leave this as a movement-only patch.  Once we
decide on a name I'll do the rename/comment.  It seems like there are
very few users currently (oprofile/backtrace.c only?).

Then I'll also change the kprobes code to use this instead of their
stack_addr helper.

Harvey