Re: [PATCH v6 1/6] arm64: Add HAVE_REGS_AND_STACK_ACCESS_API feature

[Catalin Marinas <catalin.marinas@arm.com>]

On Wed, May 20, 2015 at 11:29:24PM -0400, David Long wrote:
> On 05/20/15 09:39, Catalin Marinas wrote:
> >On Mon, Apr 20, 2015 at 04:19:42PM -0400, David Long wrote:
> >>diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> >>index 6913643..58c0223 100644
> >>--- a/arch/arm64/include/uapi/asm/ptrace.h
> >>+++ b/arch/arm64/include/uapi/asm/ptrace.h
> >>@@ -61,6 +61,42 @@
> >>
> >>  #ifndef __ASSEMBLY__
> >>
> >>+#define ARM_pstate	pstate
> >>+#define ARM_pc		pc
> >>+#define ARM_sp		sp
> >>+#define ARM_lr		regs[30]
> >>+#define ARM_fp		regs[29]
> >>+#define ARM_x28		regs[28]
> >>+#define ARM_x27		regs[27]
> >>+#define ARM_x26		regs[26]
> >>+#define ARM_x25		regs[25]
> >>+#define ARM_x24		regs[24]
> >>+#define ARM_x23		regs[23]
> >>+#define ARM_x22		regs[22]
> >>+#define ARM_x21		regs[21]
> >>+#define ARM_x20		regs[20]
> >>+#define ARM_x19		regs[19]
> >>+#define ARM_x18		regs[18]
> >>+#define ARM_ip1		regs[17]
> >>+#define ARM_ip0		regs[16]
> >>+#define ARM_x15		regs[15]
> >>+#define ARM_x14		regs[14]
> >>+#define ARM_x13		regs[13]
> >>+#define ARM_x12		regs[12]
> >>+#define ARM_x11		regs[11]
> >>+#define ARM_x10		regs[10]
> >>+#define ARM_x9		regs[9]
> >>+#define ARM_x8		regs[8]
> >>+#define ARM_x7		regs[7]
> >>+#define ARM_x6		regs[6]
> >>+#define ARM_x5		regs[5]
> >>+#define ARM_x4		regs[4]
> >>+#define ARM_x3		regs[3]
> >>+#define ARM_x2		regs[2]
> >>+#define ARM_x1		regs[1]
> >>+#define ARM_x0		regs[0]
> >>+#define ARM_ORIG_x0	orig_x0
> >
> >I replied some time ago on this part. I don't see the point these
> >macros.
> 
> I replied belatedly on April 20 saying what I did matches (more or less) how
> it's done on various other platforms, including arm and powerpc.
> It looks like this comes from the pt_regs structure defining the
> registers as an array instead of a list of structure fields. It looks
> to me like that design choice is pretty widely depended upon now and
> would be quite disruptive to change.  It also seems to me a relatively
> clean way to do it on systems with a uniform register set.

I see why we need to cope with the regs[] array but why do we need these
definitions in a uapi file?

> >>+
> >>  /*
> >>   * User structures for general purpose, floating point and debug registers.
> >>   */
> >>diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
> >>index d882b83..a889f79 100644
> >>--- a/arch/arm64/kernel/ptrace.c
> >>+++ b/arch/arm64/kernel/ptrace.c
> >>@@ -48,6 +48,122 @@
> >>  #define CREATE_TRACE_POINTS
> >>  #include <trace/events/syscalls.h>
> >>
> >>+struct pt_regs_offset {
> >>+	const char *name;
> >>+	int offset;
> >>+};
> >>+
> >>+#define REG_OFFSET_NAME(r) \
> >>+	{.name = #r, .offset = offsetof(struct pt_regs, ARM_##r)}
> 
> >Can you not just use "offsetof(struct pt_regs, r)" here? That would be
> >the same as x86, powerpc.
> 
> The registers (except for pc, pstate, and sp) are not separate structure
> fields, they are slots in a single array. To reference them the symbolic
> name has to be converted to an index (integer register number) somehow.

Can we not keep them local to this file, say __reg_x0 etc. (something to
make it clear they are for internal use)?

-- 
Catalin