This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH V3 1/3] Add AVX512 registers support to GDB.
- From: Pedro Alves <palves at redhat dot com>
- To: Michael Sturm <michael dot sturm at intel dot com>
- Cc: eliz at gnu dot org, mark dot kettenis at xs4all dot nl, walfred dot tedeschi at intel dot com, gdb-patches at sourceware dot org
- Date: Mon, 10 Mar 2014 15:02:31 +0000
- Subject: Re: [PATCH V3 1/3] Add AVX512 registers support to GDB.
- Authentication-results: sourceware.org; auth=none
- References: <1394011377-30798-1-git-send-email-michael dot sturm at intel dot com> <1394011377-30798-2-git-send-email-michael dot sturm at intel dot com>
Hi Michael,
Only a few nits below.
On 03/05/2014 09:22 AM, Michael Sturm wrote:
> +++ b/gdb/features/i386/32bit-avx512.xml
> @@ -0,0 +1,30 @@
> +<?xml version="1.0"?>
> +<!-- Copyright (C) 2013 Free Software Foundation, Inc.
Please remember to update these dates to '2013-2014' before pushing.
I spotted this at least in the xml and test files.
> @@ -29,7 +29,7 @@
> /* Register number for the "orig_eax" pseudo-register. If this
> pseudo-register contains a value >= 0 it is interpreted as the
> system call number that the kernel is supposed to restart. */
> -#define I386_LINUX_ORIG_EAX_REGNUM I386_MPX_NUM_REGS
> +#define I386_LINUX_ORIG_EAX_REGNUM I386_ZMM7H_REGNUM + 1
Wrap in parens to avoid nasty surprises:
#define I386_LINUX_ORIG_EAX_REGNUM (I386_ZMM7H_REGNUM + 1)
> --- a/gdb/i386-tdep.c
> +++ b/gdb/i386-tdep.c
> @@ -1,3 +1,4 @@
> +
> /* Intel 386 target-dependent stuff.
Spurious new line.
> + if (memcmp (raw, p, 16))
memcmp's result is not a boolean. Please write
if (memcmp (raw, p, 16) != 0)
(several places)
> +set test "print have_avx512 ()"
> +gdb_test_multiple "$test" $test {
> + -re ".. = 1\r\n$gdb_prompt $" {
> + pass "check whether processor supports AVX512"
> + }
> + -re ".. = 0\r\n$gdb_prompt $" {
> + unsupported "processor does not support AVX512; skipping AVX512 tests"
Hmm, I guess I'm missing something, but how does this result in
the rest of the tests being skipped? Won't they all fail as is?
Can you confirm whether this all passes cleanly on a machine that
does not support AVX512?
--
Pedro Alves