public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Carl Love <cel@us.ibm.com>
To: Joel Brobecker <brobecker@adacore.com>, cel@us.ibm.com
Cc: Pedro Alves <pedro@palves.net>,
	will schmidt <will_schmidt@vnet.ibm.com>,
	gdb-patches@sourceware.org,
	Ulrich Weigand <Ulrich.Weigand@de.ibm.com>,
	Tulio Magno <tuliom@linux.ibm.com>,
	Rogerio Alves <rogealve@br.ibm.com>
Subject: Re: [PATCH 2/2 Version 5] Add recording support for the ISA 3.1 Powerpc instructions
Date: Mon, 18 Apr 2022 12:21:39 -0700	[thread overview]
Message-ID: <9c137bdf94b78fa5fb3f43c6b63df29a5f91751f.camel@us.ibm.com> (raw)
In-Reply-To: <YlxE7QAwXnwjKSWs@adacore.com>

Joel, Pedro, Will, GDB maintainers:

On Sun, 2022-04-17 at 09:48 -0700, Joel Brobecker wrote:
> 
> As mentioned by Pedro: PowerPC

<snip>

> > This patch adds Powerpc specific tests to verify recording of
> > various
> 
> And here also.

Yup, missed them.  I went thru these patches and the rest of the
patches I am working on to fix in in all of them. Have some additional
posted patches that need to have this fixed as well.


Fixed the additional space issues you mentioned.

<snip>
> 
> > +# Check initial and new of vs0 are different.
> > +set test_vs0_init_new "check vs0 initial versus vs0 new"
> > +if {[string compare $vs0_initial $vs0_new] == 0} {
> > +    fail $test_vs0_init_new
> > +} else {
> > +    pass $test_vs0_init_new
> > +}
> 
> I looked at gdb_assert following Pedro's suggestion, and I agree
> it would really simplify your testcase. Any reason why you didn't
> follow his advice?

I looked at his suggestion, but it didn't seem to me to make much
difference as you would still need the if statement.  I looked at it
again and realized that it really does make is a lot simpler then I
initially thought since you don't need the if statement.  The
gdb_assert takes care of doing the test pass/fail.   So, yea it is a
good change.  I changed the string compares in both the ISA 2.07 and
ISA 3.1 test cases.

<snip>

> > +# Execute in reverse to before test 1
> > +gdb_test "set exec-direction reverse" "" "reverse to start of test
> > 1"
> > +#gdb_test_no_output "set exec-direction reverse"
> 
> I just noticied this, because we normally avoid commented-out code
> without a good reason behind it (which we also would typically
> provide as a comment next to the code).
> 
> But looking at it more closely, I think the commented-out
> version seems to be the correct one, not the one you are using.
> In particular, there is an implicit ".*" at the start of the
> expected output's matching in gdb_test, which means using ""
> as the expected output is the same as matching anything.

Don't remember if I was playing around with using the
gdb_test_no_output or not.  I can't remember now.  I may have been and
got distracted and never got back to it????   Yea, the
gdb_test_no_output works just as well as the gdb_test.  Changed to
following tests as well.  Note, looks like I did use the
gdb_test_no_output in the ISA 2.07 test.

Again, retested this patch to make sure I didn't break anything.

Thanks for the detailed review.

                                Carl 
-----------------------------------------------------------
GDB PowerPC record test cases for ISA 2.06 and ISA 3.1

This patch adds PowerPC specific tests to verify recording of various
instructions.  The first test case checks the ISA 2.06 lxvd2x instruction.
The second test case tests several of the ISA 3.01 instructions.  Specifically,
it checks the word and prefixed instructions and some of the Matrix
Multiply Assist (MMA) instructions.

The patch has been run on both Power 10 and Power 9 to verify the ISA
2.06 test case runs on both platforms without errors.  The ISA 3.1 test
runs without errors on Power 10 and is skipped as expected on Power 9.
---
 .../gdb.reverse/ppc_record_test_isa_2_06.c    |  39 ++
 .../gdb.reverse/ppc_record_test_isa_2_06.exp  | 105 ++++++
 .../gdb.reverse/ppc_record_test_isa_3_1.c     |  95 +++++
 .../gdb.reverse/ppc_record_test_isa_3_1.exp   | 341 ++++++++++++++++++
 4 files changed, 580 insertions(+)
 create mode 100644 gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.c
 create mode 100644 gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.exp
 create mode 100644 gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.c
 create mode 100644 gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.exp

diff --git a/gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.c b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.c
new file mode 100644
index 00000000000..45e58413da2
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.c
@@ -0,0 +1,39 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+   Copyright 2012-2022 Free Software Foundation, Inc.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+/* globals used for vector tests */
+static vector unsigned long vec_xb;
+static unsigned long ra, rb, rs;
+
+int
+main ()
+{
+  ra = 0xABCDEF012;
+  rb = 0;
+  rs = 0x012345678;
+
+  /* 9.0, 16.0, 25.0, 36.0 */
+  vec_xb = (vector unsigned long){0x4110000041800000, 0x41c8000042100000};
+
+  /* Test ISA 2.06 instructions.  Load source into vs1, result of sqrt
+     put into vs0.  */
+  ra = (unsigned long) & vec_xb;        /* stop 1 */
+  __asm__ __volatile__ ("lxvd2x 1, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("xvsqrtsp 0, 1");
+  ra = 0;                               /* stop 2 */
+}
+
diff --git a/gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.exp b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.exp
new file mode 100644
index 00000000000..d68dd7b9049
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_2_06.exp
@@ -0,0 +1,105 @@
+# Copyright 2008-2022 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# Test instruction record for PowerPC, ISA 2.06.
+#
+
+# The basic flow of the record tests are:
+#    1) Stop before executing the instructions of interest.  Record
+#       the initial value of the registers that the instruction will
+#       change, i.e. the destination register.
+#    2) Execute the instructions.  Record the new value of the
+#       registers that changed.
+#    3) Reverse the direction of the execution and execute back to
+#       just before the instructions of interest.  Record the final
+#       value of the registers of interest.
+#    4) Check that the initial and new values of the registers are
+#       different, i.e. the instruction changed the registers as expected.
+#    5) Check that the initial and final values of the registers are
+#       the same, i.e. gdb record restored the registers to their
+#       original values.
+
+standard_testfile
+
+set gen_src record_test_isa_2_06.c
+set executable record_test_isa_2_06
+set options [list debug]
+
+if {![istarget "powerpc*"]} then {
+    verbose "Skipping PowerPC ISA 2.06 instruction record_test_2_06."
+    return
+}
+
+if {[build_executable "failed to prepare" $executable $srcfile $options] == -1} then {
+    return -1
+}
+
+clean_restart $executable
+
+if ![runto_main] then {
+    untested "could not run to main"
+    continue
+}
+
+gdb_test_no_output "record"
+
+###### Test: Test an ISA 2.06 load (lxvd2x) and square root instruction
+###### (xvsqrtsp).  The load instruction will load vs1.  The sqrt instruction
+###### will put its result into vs0.
+
+set stop1 [gdb_get_line_number "stop 1"]
+set stop2 [gdb_get_line_number "stop 2"]
+
+gdb_test "break $stop1" ".*Breakpoint .*" "about to execute test"
+gdb_test "continue"  ".*Breakpoint .*" "at stop 1"
+
+# Record the initial values in vs0, vs1.
+set vs0_initial [capture_command_output "info register vs0" ""]
+set vs1_initial [capture_command_output "info register vs1" ""]
+
+gdb_test "break $stop2" ".*Breakpoint .*" "executed lxvd2x, xvsqrtsp"
+gdb_test "continue"  ".*Breakpoint .*" "at stop 2"
+
+# Record the new values of vs0 and vs1.
+set vs0_new [capture_command_output "info register vs0" ""]
+set vs1_new [capture_command_output "info register vs1" ""]
+
+# Reverse the execution direction.
+gdb_test_no_output "set exec-direction reverse"
+gdb_test "break $stop1" ".*Breakpoint .*" "un executed lxvd2x, xvsqrtsp"
+
+# Execute in reverse to before the lxvd2x instruction.
+gdb_test "continue"  ".*Breakpoint.*" "at stop 1 in reverse"
+
+# Record the final values of vs0, vs1.
+set vs0_final [capture_command_output "info register vs0" ""]
+set vs1_final [capture_command_output "info register vs1" ""]
+
+# Check initial and new of vs0 are different.
+gdb_assert [string compare $vs0_initial $vs1_new] \
+    "check vs0 initial versus vs0 new"
+
+# Check initial and new of vs1 are different.
+gdb_assert [string compare $vs1_initial $vs1_new] \
+    "check vs0 initial versus vs1 new"
+
+# Check initial and final are the same.
+gdb_assert ![string compare $vs0_initial $vs0_final] \
+    "check vs0 initial versus vs0 final"
+
+# Check initial and final are the same.
+gdb_assert ![string compare $vs1_initial $vs1_final] \
+    "check vs1 initial versus vs1 final"
+
diff --git a/gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.c b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.c
new file mode 100644
index 00000000000..c0d65d944af
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.c
@@ -0,0 +1,95 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+   Copyright 2012-2022 Free Software Foundation, Inc.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+/* Globals used for vector tests.  */
+static vector unsigned long vec_xa, vec_xb, vec_xt;
+static unsigned long ra, rb, rs;
+
+int
+main ()
+{
+  ra = 0xABCDEF012;
+  rb = 0;
+  rs = 0x012345678;
+
+  /* 9.0, 16.0, 25.0, 36.0 */
+  vec_xb = (vector unsigned long){0x4110000041800000, 0x41c8000042100000};
+
+  vec_xt = (vector unsigned long){0xFF00FF00FF00FF00, 0xAA00AA00AA00AA00};
+
+  /* Test 1, ISA 3.1 word instructions. Load source into r1, result of brh
+     put in r0.  */
+  ra = 0xABCDEF012;                     /* stop 1 */
+  __asm__ __volatile__ ("pld 1, %0" :: "r" (ra ));
+  __asm__ __volatile__ ("brh 0, 1" );
+  ra = 0;                               /* stop 2 */
+
+  /* Test 2, ISA 3.1 MMA instructions with results in various ACC entries
+     xxsetaccz    - ACC[3]
+     xvi4ger8     - ACC[4]
+     xvf16ger2pn  - ACC[5]
+     pmxvi8ger4   - ACC[6]
+     pmxvf32gerpp - ACC[7] and fpscr */
+  /* Need to initialize the vs registers to a non zero value.  */
+  ra = (unsigned long) & vec_xb;
+  __asm__ __volatile__ ("lxvd2x 12, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 13, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 14, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 15, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xa = (vector unsigned long){0x333134343987601, 0x9994bbbc9983307};
+  vec_xb = (vector unsigned long){0x411234041898760, 0x41c833042103400};
+  __asm__ __volatile__ ("lxvd2x 16, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xb = (vector unsigned long){0x123456789987650, 0x235676546989807};
+  __asm__ __volatile__ ("lxvd2x 17, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xb = (vector unsigned long){0x878363439823470, 0x413434c99839870};
+  __asm__ __volatile__ ("lxvd2x 18, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xb = (vector unsigned long){0x043765434398760, 0x419876555558850};
+  __asm__ __volatile__ ("lxvd2x 19, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xb = (vector unsigned long){0x33313434398760, 0x9994bbbc99899330};
+  __asm__ __volatile__ ("lxvd2x 20, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 21, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 22, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 23, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 24, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 25, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 26, %0, %1" :: "r" (ra ), "r" (rb));
+  __asm__ __volatile__ ("lxvd2x 27, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xa = (vector unsigned long){0x33313434398760, 0x9994bbbc998330};
+  vec_xb = (vector unsigned long){0x4110000041800000, 0x41c8000042100000};
+  __asm__ __volatile__ ("lxvd2x 28, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xb = (vector unsigned long){0x4567000046800000, 0x4458000048700000};
+  __asm__ __volatile__ ("lxvd2x 29, %0, %1" :: "r" (ra ), "r" (rb));
+  vec_xb = (vector unsigned long){0x41dd000041e00000, 0x41c8000046544400};
+  __asm__ __volatile__ ("lxvd2x 30, %0, %1" :: "r" (ra ), "r" (rb));
+
+  /* SNAN */
+  vec_xb = (vector unsigned long){0x7F8F00007F8F0000, 0x7F8F00007F8F0000};
+
+  __asm__ __volatile__ ("lxvd2x 31, %0, %1" :: "r" (ra ), "r" (rb));
+
+  ra = 0xAB;                            /* stop 3 */
+  __asm__ __volatile__ ("xxsetaccz 3");
+  __asm__ __volatile__ ("xvi4ger8 4, %x0, %x1" :: "wa" (vec_xa), \
+			"wa" (vec_xb) );
+  __asm__ __volatile__ ("xvf16ger2pn 5, %x0, %x1" :: "wa" (vec_xa),\
+			"wa" (vec_xb) );
+  __asm__ __volatile__ ("pmxvi8ger4spp  6, %x0, %x1, 11, 13, 5"
+                                :: "wa" (vec_xa), "wa" (vec_xb) );
+  __asm__ __volatile__ ("pmxvf32gerpp  7, %x0, %x1, 11, 13"
+                                :: "wa" (vec_xa), "wa" (vec_xb) );
+  ra = 0;                               /* stop 4 */
+}
diff --git a/gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.exp b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.exp
new file mode 100644
index 00000000000..9ea8931e9cf
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/ppc_record_test_isa_3_1.exp
@@ -0,0 +1,341 @@
+# Copyright 2008-2022 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# Test instruction record for PowerPC, ISA 3.1.
+#
+
+# The basic flow of the record tests are:
+#    1) Stop before executing the instructions of interest.  Record
+#       the initial value of the registers that the instruction will
+#       change, i.e. the destination register.
+#    2) Execute the instructions.  Record the new value of the
+#       registers that changed.
+#    3) Reverse the direction of the execution and execute back to
+#       just before the instructions of interest.  Record the final
+#       value of the registers of interest.
+#    4) Check that the initial and new values of the registers are
+#       different, i.e. the instruction changed the registers as expected.
+#    5) Check that the initial and final values of the registers are
+#       the same, i.e. gdb record restored the registers to their
+#       original values.
+
+
+standard_testfile
+
+set gen_src record_test_isa_3_1.c
+set executable record_test_isa_3_1
+
+if {![istarget "powerpc*"] || [skip_power_isa_3_1_tests] } then  {
+    verbose "Skipping PowerPC ISA 3.1 instruction record_test."
+    return
+}
+
+set options [list additional_flags=-mcpu=power10 debug]
+if {[build_executable "failed to prepare" $executable $srcfile $options] == -1} then {
+    return -1
+}
+
+clean_restart $executable
+
+if ![runto_main] then {
+    untested "could not run to main"
+    continue
+}
+
+gdb_test_no_output "record"
+
+######  Test 1:  Test an ISA 3.1 byte reverse word instruction (brd) and a
+######  prefixed load double (pld) instruction.
+set stop1 [gdb_get_line_number "stop 1"]
+set stop2 [gdb_get_line_number "stop 2"]
+
+gdb_test "break $stop1" ".*Breakpoint .*" "about to execute test 1"
+gdb_test "continue"  ".*Breakpoint .*" "at stop 1"
+
+# Record the initial values in r0, r1
+# Load the argument into r1, result of byte reverse is put into r0.
+set r0_initial [capture_command_output "info register r0" ""]
+set r1_initial [capture_command_output "info register r1" ""]
+
+gdb_test "break $stop2" ".*Breakpoint .*" "executed test 1"
+gdb_test "continue"  ".*Breakpoint .*" "at stop 2"
+
+# Record the new values of r0 and r1
+set r0_new [capture_command_output "info register r0" ""]
+set r1_new [capture_command_output "info register r1" ""]
+
+# Execute in reverse to before test 1
+gdb_test_no_output "set exec-direction reverse"
+
+gdb_test "break $stop1" ".*Breakpoint .*" "reverse stop at test 1 start"
+gdb_test "continue"  ".*Breakpoint.*" "at stop 1 in reverse"
+
+# Record the final values of r0, r1
+set r0_final [capture_command_output "info register r0" ""]
+set r1_final [capture_command_output "info register r1" ""]
+
+# Check initial and new of r0 are different.
+gdb_assert [string compare $r0_initial $r0_new] \
+    "check r0 initial versus r0 new"
+
+# Check initial and new of r1 are different.
+gdb_assert [string compare $r1_initial $r1_new] \
+    "check r0 initial versus r1 new"
+
+# Check initial and final are the same.
+gdb_assert ![string compare $r0_initial $r0_final] \
+    "check r0 initial versus r0 final"
+
+# Check initial and final are the same.
+gdb_assert ![string compare $r1_initial $r1_final] \
+    "check r1 initial versus r1 final"
+
+
+# Change execution direction to forward for next test.
+gdb_test_no_output "set exec-direction forward" "start forward test3"
+gdb_test "record stop" ".*Process record is stopped.*" "stopped recording 2"
+set test_del_bkpts "delete breakpoints, answer prompt 2"
+
+# Delete all breakpoints and catchpoints.
+delete_breakpoints
+
+gdb_test_no_output "record" "start recording test2"
+
+
+######  Test 2:  Test the ISA 3.1 MMA instructions xxsetaccz, xvi4ger8,
+######  xvf16ger2pn, pmxvi8ger4, and pmxvf32gerpp.  Goal here is to hit all
+######  the places where ppc_record_ACC_fpscr() gets called.
+##
+##       xxsetaccz    - ACC[3], vs[12] to vs[15]
+##       xvi4ger8     - ACC[4], vs[16] to vs[19]
+##       xvf16ger2pn  - ACC[5], vs[20] to vs[23]
+##       pmxvi8ger4   - ACC[6], vs[21] to vs[27]
+##       pmxvf32gerpp - ACC[7], vs[28] to vs[31] and fpscr
+
+set stop3 [gdb_get_line_number "stop 3"]
+set stop4 [gdb_get_line_number "stop 4"]
+
+gdb_test "break $stop3" ".*Breakpoint .*" "about to execute test 2"
+gdb_test "continue"  ".*Breakpoint .*" "at stop 3"
+
+# Record the initial values of vs's that correspond to the ACC entries,
+# and fpscr.
+set acc_3_0_initial [capture_command_output "info register vs12" ""]
+set acc_3_1_initial [capture_command_output "info register vs13" ""]
+set acc_3_2_initial [capture_command_output "info register vs14" ""]
+set acc_3_3_initial [capture_command_output "info register vs15" ""]
+set acc_4_0_initial [capture_command_output "info register vs16" ""]
+set acc_4_1_initial [capture_command_output "info register vs17" ""]
+set acc_4_2_initial [capture_command_output "info register vs18" ""]
+set acc_4_3_initial [capture_command_output "info register vs19" ""]
+set acc_5_0_initial [capture_command_output "info register vs20" ""]
+set acc_5_1_initial [capture_command_output "info register vs21" ""]
+set acc_5_2_initial [capture_command_output "info register vs22" ""]
+set acc_5_3_initial [capture_command_output "info register vs23" ""]
+set acc_6_0_initial [capture_command_output "info register vs24" ""]
+set acc_6_1_initial [capture_command_output "info register vs25" ""]
+set acc_6_2_initial [capture_command_output "info register vs26" ""]
+set acc_6_3_initial [capture_command_output "info register vs27" ""]
+set acc_7_0_initial [capture_command_output "info register vs28" ""]
+set acc_7_1_initial [capture_command_output "info register vs29" ""]
+set acc_7_2_initial [capture_command_output "info register vs30" ""]
+set acc_7_3_initial [capture_command_output "info register vs31" ""]
+set fpscr_initial [capture_command_output "info register fpscr" ""]
+
+gdb_test "break $stop4" ".*Breakpoint .*" "executed test 2"
+gdb_test "continue"  ".*Breakpoint .*" "at stop 4"
+
+# Record the new values of the ACC entries and fpscr.
+set acc_3_0_new [capture_command_output "info register vs12" ""]
+set acc_3_1_new [capture_command_output "info register vs13" ""]
+set acc_3_2_new [capture_command_output "info register vs14" ""]
+set acc_3_3_new [capture_command_output "info register vs15" ""]
+set acc_4_0_new [capture_command_output "info register vs16" ""]
+set acc_4_1_new [capture_command_output "info register vs17" ""]
+set acc_4_2_new [capture_command_output "info register vs18" ""]
+set acc_4_3_new [capture_command_output "info register vs19" ""]
+set acc_5_0_new [capture_command_output "info register vs20" ""]
+set acc_5_1_new [capture_command_output "info register vs21" ""]
+set acc_5_2_new [capture_command_output "info register vs22" ""]
+set acc_5_3_new [capture_command_output "info register vs23" ""]
+set acc_6_0_new [capture_command_output "info register vs24" ""]
+set acc_6_1_new [capture_command_output "info register vs25" ""]
+set acc_6_2_new [capture_command_output "info register vs26" ""]
+set acc_6_3_new [capture_command_output "info register vs27" ""]
+set acc_7_0_new [capture_command_output "info register vs28" ""]
+set acc_7_1_new [capture_command_output "info register vs29" ""]
+set acc_7_2_new [capture_command_output "info register vs30" ""]
+set acc_7_3_new [capture_command_output "info register vs31" ""]
+set fpscr_new [capture_command_output "info register fpscr" ""]
+
+# Execute in reverse to before test 2.
+gdb_test_no_output "set exec-direction reverse" "reverse to start of test 2"
+
+gdb_test "break $stop3" ".*Breakpoint .*" "reverse stop at test 2 start"
+gdb_test "continue"  ".*Breakpoint.*" "at stop 3 in reverse"
+
+# Record the final values of the ACC entries and fpscr.
+set acc_3_0_final [capture_command_output "info register vs12" ""]
+set acc_3_1_final [capture_command_output "info register vs13" ""]
+set acc_3_2_final [capture_command_output "info register vs14" ""]
+set acc_3_3_final [capture_command_output "info register vs15" ""]
+set acc_4_0_final [capture_command_output "info register vs16" ""]
+set acc_4_1_final [capture_command_output "info register vs17" ""]
+set acc_4_2_final [capture_command_output "info register vs18" ""]
+set acc_4_3_final [capture_command_output "info register vs19" ""]
+set acc_5_0_final [capture_command_output "info register vs20" ""]
+set acc_5_1_final [capture_command_output "info register vs21" ""]
+set acc_5_2_final [capture_command_output "info register vs22" ""]
+set acc_5_3_final [capture_command_output "info register vs23" ""]
+set acc_6_0_final [capture_command_output "info register vs24" ""]
+set acc_6_1_final [capture_command_output "info register vs25" ""]
+set acc_6_2_final [capture_command_output "info register vs26" ""]
+set acc_6_3_final [capture_command_output "info register vs27" ""]
+set acc_7_0_final [capture_command_output "info register vs28" ""]
+set acc_7_1_final [capture_command_output "info register vs29" ""]
+set acc_7_2_final [capture_command_output "info register vs30" ""]
+set acc_7_3_final [capture_command_output "info register vs31" ""]
+set fpscr_final [capture_command_output "info register fpscr" ""]
+
+# check initial and new ACC entries are different.
+gdb_assert [string compare $acc_3_0_initial $acc_3_0_new] \
+    "check vs12 initial versus new"
+
+gdb_assert [string compare $acc_3_1_initial $acc_3_1_new] \
+    "check vs13 initial versus new"
+
+gdb_assert [string compare $acc_3_2_initial $acc_3_2_new] \
+    "check vs14 initial versus new"
+
+gdb_assert [string compare $acc_3_3_initial $acc_3_3_new] \
+    "check vs15 initial versus new"
+
+gdb_assert [string compare $acc_4_0_initial $acc_4_0_new] \
+    "check vs16 initial versus new"
+
+gdb_assert [string compare $acc_4_1_initial $acc_4_1_new] \
+    "check vs17 initial versus new"
+
+gdb_assert [string compare $acc_4_2_initial $acc_4_2_new] \
+    "check vs18 initial versus new"
+
+gdb_assert [string compare $acc_4_3_initial $acc_4_3_new] \
+    "check vs19 initial versus new"
+
+gdb_assert [string compare $acc_5_0_initial $acc_5_0_new] \
+    "check vs20 initial versus new"
+
+gdb_assert [string compare $acc_5_1_initial $acc_5_1_new] \
+    "check vs21 initial versus new"
+
+gdb_assert [string compare $acc_5_2_initial $acc_5_2_new] \
+    "check vs22 initial versus new"
+
+gdb_assert [string compare $acc_5_3_initial $acc_5_3_new] \
+    "check vs23 initial versus new"
+
+gdb_assert [string compare $acc_6_0_initial $acc_6_0_new] \
+    "check vs24 initial versus new"
+
+gdb_assert [string compare $acc_6_1_initial $acc_6_1_new] \
+    "check vs25 initial versus new"
+
+gdb_assert [string compare $acc_6_2_initial $acc_6_2_new] \
+    "check vs26 initial versus new"
+
+gdb_assert [string compare $acc_6_3_initial $acc_6_3_new] \
+    "check vs27 initial versus new"
+
+gdb_assert [string compare $acc_7_0_initial $acc_7_0_new] \
+    "check vs28 initial versus new"
+
+gdb_assert [string compare $acc_7_1_initial $acc_7_1_new] \
+    "check vs29 initial versus new"
+
+gdb_assert [string compare $acc_7_2_initial $acc_7_2_new] \
+    "check vs30 initial versus new"
+
+gdb_assert [string compare $acc_7_3_initial $acc_7_3_new] \
+    "check vs31 initial versus new"
+
+gdb_assert [string compare $fpscr_initial $fpscr_new] \
+    "check fpscr initial versus new"
+
+
+# Check initial and final ACC entries are the same.
+gdb_assert ![string compare $acc_3_0_initial $acc_3_0_final] \
+    "check vs12 initial versus final"
+
+gdb_assert ![string compare $acc_3_1_initial $acc_3_1_final] \
+    "check vs13 initial versus final"
+
+gdb_assert ![string compare $acc_3_2_initial $acc_3_2_final] \
+    "check vs14 initial versus final"
+
+gdb_assert ![string compare $acc_3_3_initial $acc_3_3_final] \
+    "check vs15 initial versus final"
+
+gdb_assert ![string compare $acc_4_0_initial $acc_4_0_final] \
+    "check vs16 initial versus final"
+
+gdb_assert ![string compare $acc_4_1_initial $acc_4_1_final] \
+    "check vs17 initial versus final"
+
+gdb_assert ![string compare $acc_4_2_initial $acc_4_2_final] \
+    "check vs18 initial versus final"
+
+gdb_assert ![string compare $acc_4_3_initial $acc_4_3_final] \
+    "check vs19 initial versus final"
+
+gdb_assert ![string compare $acc_5_0_initial $acc_5_0_final] \
+    "check vs20 initial versus final"
+
+gdb_assert ![string compare $acc_5_1_initial $acc_5_1_final] \
+    "check vs21 initial versus final"
+
+gdb_assert ![string compare $acc_5_2_initial $acc_5_2_final] \
+    "check vs22 initial versus final"
+
+gdb_assert ![string compare $acc_5_3_initial $acc_5_3_final] \
+    "check vs23 initial versus final"
+
+gdb_assert ![string compare $acc_6_0_initial $acc_6_0_final] \
+    "check vs24 initial versus final"
+
+gdb_assert ![string compare $acc_6_1_initial $acc_6_1_final] \
+    "check vs25 initial versus final"
+
+gdb_assert ![string compare $acc_6_2_initial $acc_6_2_final] \
+    "check vs26 initial versus final"
+
+gdb_assert ![string compare $acc_6_3_initial $acc_6_3_final] \
+    "check vs27 initial versus final"
+
+gdb_assert ![string compare $acc_7_0_initial $acc_7_0_final] \
+    "check vs28 initial versus final"
+
+gdb_assert ![string compare $acc_7_1_initial $acc_7_1_final] \
+    "check vs29 initial versus final"
+
+gdb_assert ![string compare $acc_7_2_initial $acc_7_2_final] \
+    "check !vs30 initial versus final"
+
+gdb_assert ![string compare $acc_7_3_initial $acc_7_3_final] \
+    "check !vs31 initial versus final"
+
+gdb_assert ![string compare $fpscr_initial $fpscr_final] \
+    "check fpscr initial versus final"
+
+
-- 
2.31.1



  reply	other threads:[~2022-04-18 19:22 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-04 19:46 [PATCH 0/2] " Carl Love
2022-03-04 19:52 ` [PATCH 1/2] " Carl Love
2022-03-06 11:53   ` Joel Brobecker
2022-04-12 17:09     ` [PATCH 1/2 Version 2] " Carl Love
2022-04-12 21:50       ` will schmidt
2022-04-13 17:26         ` [PATCH 1/2 Version 3] " Carl Love
2022-04-17 16:23           ` Joel Brobecker
2022-04-18 19:21             ` [PATCH 1/2 Version 5] " Carl Love
2022-04-22 19:49               ` [PATCH 1/2 Version 5 Ping] " Carl Love
2022-04-24 16:50               ` [PATCH 1/2 Version 5] " Joel Brobecker
2022-04-25 15:58                 ` [PATCH 1/2 Version 6] " Carl Love
2022-04-26 18:05                   ` Joel Brobecker
2022-03-04 19:53 ` [PATCH 2/2] " Carl Love
2022-03-06 12:42   ` Joel Brobecker
2022-04-12 17:09     ` [PATCH 2/2 Version 2] " Carl Love
2022-04-13 14:12       ` will schmidt
2022-04-13 21:38         ` [PATCH 2/2 Version 3] " Carl Love
2022-04-14 13:05           ` Pedro Alves
2022-04-14 20:20             ` [PATCH 2/2 Version 4] " Carl Love
2022-04-14 20:43               ` Carl Love
2022-04-17 16:48                 ` Joel Brobecker
2022-04-18 19:21                   ` Carl Love [this message]
2022-04-22 19:47                     ` [PATCH 2/2 Version 5 PING] " Carl Love
2022-04-24 16:56                     ` [PATCH 2/2 Version 5] " Joel Brobecker
2022-04-12 17:09 ` [PATCH 0/2 Version 2] " Carl Love

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=9c137bdf94b78fa5fb3f43c6b63df29a5f91751f.camel@us.ibm.com \
    --to=cel@us.ibm.com \
    --cc=Ulrich.Weigand@de.ibm.com \
    --cc=brobecker@adacore.com \
    --cc=gdb-patches@sourceware.org \
    --cc=pedro@palves.net \
    --cc=rogealve@br.ibm.com \
    --cc=tuliom@linux.ibm.com \
    --cc=will_schmidt@vnet.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).