x86-64: fix ZMM register state tracking
Commit Message
On 2018-09-18 09:37 AM, Jan Beulich wrote:
>>>> On 10.09.18 at 15:01, <simon.marchi@polymtl.ca> wrote:
>> On 2018-09-10 07:25, Jan Beulich wrote:
>>>>>> Simon Marchi <simon.marchi@ericsson.com> 09/08/18 1:13 AM >>>
>>>> Would it be possible to update or create a test to exercise that?
>>>> arch-specific tests are in testsuite/gdb.arch.
>>>
>>> I'm sure it would be possible, but while I was happy to invest the time
>>> to
>>> fix the actual bug (because it affects work I'm doing), I'm afraid I
>>> don't have
>>> the time to learn how gdb test cases are to be constructed (I'm
>>> familiar
>>> only with the binutils / gas side of things).
>>
>> I understand. If you can provide:
>>
>> - a minimal source file (C + assembly in this case, I suppose)
>> - GDB commands to reproduce the bug
>> - actual and expected output
>
> Attached. vzero.s is the source file used (no C file needed). gdb.log
> is a transcript of a session with a broken gdb (the one installed on
> the system), while gdb.txt is a transcript for the fixed one that I've
> built myself.
>
>> I can take care of turning it in a GDB test case.
>
> Thanks.
>
> Jan
>
>
Hi Jan,
Thanks for the instructions. There is already a test covering AVX512
instructions, so I figured I would add it there. However, I don't
have a processor that supports AVX512, so I'm unable to run the test.
Here's a patch, can you try to confirm that the test fails without the
fix and passes with the fix? I probably screwed up somewhere, but it
should be pretty close.
You can run the test with
$ make check TESTS="gdb.arch/i386-avx512.exp"
in the gdb/ build directory. The gdb/testsuite/gdb.log file is useful
to look at in case something fails.
Feel free to integrate this in your eventual v2 if there is one, and
feel free to add a comment in the test to say what we are testing here,
as I am not too sure how to describe it.
Thanks!
Simon
From a167d21f452f6f3f19eac6623fa872fde1162128 Mon Sep 17 00:00:00 2001
From: Simon Marchi <simon.marchi@ericsson.com>
Date: Mon, 24 Sep 2018 23:21:19 -0400
Subject: [PATCH] AVX512 test
---
gdb/testsuite/gdb.arch/i386-avx512.c | 5 +++++
gdb/testsuite/gdb.arch/i386-avx512.exp | 8 ++++++++
2 files changed, 13 insertions(+)
Comments
>>> On 25.09.18 at 05:28, <simon.marchi@ericsson.com> wrote:
> On 2018-09-18 09:37 AM, Jan Beulich wrote:
> Thanks for the instructions. There is already a test covering AVX512
> instructions, so I figured I would add it there. However, I don't
> have a processor that supports AVX512, so I'm unable to run the test.
>
> Here's a patch, can you try to confirm that the test fails without the
> fix and passes with the fix? I probably screwed up somewhere, but it
> should be pretty close.
There are two issues here: First of all, unrelated to this patch, the
construct around line 95 in i386-avx512.exp should look like
if [is_amd64_regs_target] {
set nr_regs 32
} else {
set nr_regs 8
}
Of course this also affects other tests in here, but without this correction
the loop you add does nothing at all.
And then that very loop and the i386-avx512.c addition are not in sync,
and I'm not sure which way you meant it to be: Either in the C file all 16
upper ZMM registers need to be set identically (not just ZMM16), or
there should be no loop.
Furthermore I think the C code addition and hence the test will need to
be x86-64-specific, as registers ZMM8 and higher are inaccessible in
32-bit mode.
So what I can confirm at this point is that with the fix in place there's
one less new failure from the test than with the fix no in place.
Jan
@@ -249,6 +249,11 @@ main (int argc, char **argv)
move back to array and check values. */
move_zmm_data_to_memory ();
asm ("nop"); /* sixth breakpoint here */
+
+ asm ("vpternlogd $0xff, %zmm0, %zmm0, %zmm0\n"
+ "vpternlogd $0xff, %zmm0, %zmm0, %zmm16\n"
+ "vzeroupper\n");
+ asm ("nop"); /* seventh breakpoint here */
}
return 0;
@@ -174,3 +174,11 @@ for { set r 0 } { $r < $nr_regs } { incr r } {
".. = \\{f = \\{[expr $r + 30], [expr $r.125 + 30], [expr $r.25 + 20], [expr $r.375 + 20], [expr $r.5 + 10], [expr $r.625 + 10], [expr $r.75 + 10], [expr $r.875 + 10]\\}\\}.*" \
"check contents of zmm_data\[$r\] after writing XMM regs"
}
+
+gdb_test "break [gdb_get_line_number "seventh breakpoint here"]" \
+ "Breakpoint .* at .*i386-avx512.c.*" \
+ "set seventh breakpoint in main"
+gdb_continue_to_breakpoint "continue to seventh breakpoint in main"
+for { set r 16 } { $r < $nr_regs } { incr r } {
+ gdb_test "print \$zmm${r}.v16_int32" "{-1 <repeats 16 times>}"
+}