Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Guinevere Larsen <guinevere@redhat.com>
To: Ezra.Sitorus@arm.com, gdb-patches@sourceware.org
Cc: luis.machado.foss@gmail.com, thiago.bauermann@linaro.org
Subject: Re: [PATCH v2 1/4] gdb/aarch64: record/replay support for LRCPC3
Date: Tue, 26 May 2026 11:30:10 -0300	[thread overview]
Message-ID: <f741185a-f505-439f-9951-53c2a4d6003a@redhat.com> (raw)
In-Reply-To: <20260512231707.18123-2-Ezra.Sitorus@arm.com>

On 5/12/26 8:17 PM, Ezra.Sitorus@arm.com wrote:
> From: Ezra Sitorus <ezra.sitorus@arm.com>
>
> FEAT_LRCPC3 introduces various load/store instructions with release
> consistency for cases where ordering is required. This patch teaches GDB
> to decode these instructions for recording and reversing.
>
> The gdb.reverse/aarch64-lrcpc3.exp testcase verifies that the
> instructions are recorded and correctly reversed. In particular, there
> are some interesting cases to note:
> * ldapur/stlur are SIMD instructions, but are not decoded in the simd
>    function.
> * There are writeback cases to cover too. These were taken from the
>    binutils testcases: gas/testsuite/gas/aarch64/rcpc3.s.
>
> The full testsuite was done on aarch64-none-linux-gnu without LRCPC3.
> The gdb.arch and gdb.reverse tests were run on Shrinkwrap with LRCPC3
> support.
>
> Please note:
> 1) There is no support for LRCPC and LRCPC2 instructions
> 2) LRCPC3 is gated with +rcpc3 in GCC/binutils and LLVM.
> ---

Hi Ezra! I have one tiny comment with regards to the test (that can 
truthfully probably be ignored), but you shouldn't need to send a v3 
just to fix that. Feel free to add my review tag

Reviewed-By: Guinevere Larsen <guinevere@redhat.com

>   gdb/aarch64-tdep.c                           |  87 ++++++++
>   gdb/testsuite/gdb.reverse/aarch64-lrcpc3.c   | 216 +++++++++++++++++++
>   gdb/testsuite/gdb.reverse/aarch64-lrcpc3.exp | 203 +++++++++++++++++
>   gdb/testsuite/lib/gdb.exp                    |  63 ++++++
>   4 files changed, 569 insertions(+)
>   create mode 100644 gdb/testsuite/gdb.reverse/aarch64-lrcpc3.c
>   create mode 100644 gdb/testsuite/gdb.reverse/aarch64-lrcpc3.exp
>
<snip>
> diff --git a/gdb/testsuite/gdb.reverse/aarch64-lrcpc3.exp b/gdb/testsuite/gdb.reverse/aarch64-lrcpc3.exp
> new file mode 100644
> index 00000000..adcf2906
> --- /dev/null
> +++ b/gdb/testsuite/gdb.reverse/aarch64-lrcpc3.exp
> @@ -0,0 +1,203 @@
> +# Copyright 2024-2026 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 AArch64 FEAT_LRCPC3 instructions.
> +# Based on gdb.reverse/aarch64-mops.exp
> +#
> +# 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.
> +
> +require allow_aarch64_lrcpc3_tests
> +
> +standard_testfile
> +
> +if {[prepare_for_testing "failed to prepare" ${testfile} ${srcfile} \
> +	[list debug additional_flags=-march=armv8-a+rcpc3]]} {
> +    return
> +}
> +
> +if {![runto_main]} {
> +    return
> +}
> +
> +gdb_test_no_output "record full"
> +
> +proc test_single_asm {name diff_reg diff_var diff_mem same_reg same_var same_mem} {
> +    with_test_prefix $name {
> +
> +	set before_seq [gdb_get_line_number "Before ${name}"]
> +	set after_seq [gdb_get_line_number "After ${name}"]
> +
> +	set insn [lindex [split $name "-"] 0]
> +
> +	gdb_test "break $before_seq" \
> +	    "Breakpoint ${::decimal} at ${::hex}: file .*/aarch64-lrcpc3.c, line ${::decimal}\\." \
> +	    "$insn before instruction sequence"
> +
> +	gdb_continue_to_breakpoint "about to execute instruction sequence" \
> +	    [multi_line ".*/aarch64-lrcpc3.c:${::decimal}" \
> +		"${::decimal}\[ \t\]+__asm__ volatile \\(\"${insn} .*\".*"]
> +
> +	# Depending on the compiler, the line number information may put GDB a few
> +	# instructions before the beginning of the asm statement.
> +	arrive_at_instruction $insn
> +	# Add a breakpoint that we're sure is at the prologue instruction.
> +	gdb_test "break *\$pc" \
> +	    "Breakpoint ${::decimal} at ${::hex}: file .*/aarch64-lrcpc3.c, line ${::decimal}\\." \
> +	    "break at prologue instruction"
> +
> +	# Record the initial memory and register values.
> +	set regs [concat $diff_reg $same_reg]
> +	set vars [concat $diff_var $same_var]
> +	set mem [concat $diff_mem $same_mem]
> +	foreach r $regs {
> +	    set ${r}_initial [capture_command_output "info register $r" ""]
> +	}
> +	foreach v $vars {
> +	    set ${v}_initial [get_valueof "/x" "$v" "unable to read $v" \
> +		"get $v initial value"]
> +	}
> +	foreach m $mem {
> +	    set ${m}_initial [capture_command_output "x/x $m" ""]
> +	}
> +
> +	gdb_test "break $after_seq" \
> +	    "Breakpoint ${::decimal} at ${::hex}: file .*/aarch64-lrcpc3.c, line ${::decimal}\\." \
> +	    "$insn after instruction sequence"
> +	gdb_continue_to_breakpoint "executed instruction sequence" \
> +	    [multi_line ".*/aarch64-lrcpc3.c:${::decimal}" ".*"]
> +
> +	# Record the new memory and register values.
> +	foreach r $regs {
> +	    set ${r}_new [capture_command_output "info register $r" ""]
> +	}
> +	foreach v $vars {
> +	    set ${v}_new [get_valueof "/x" "$v" "unable to read $v" \
> +		"get $v new value"]
> +	}
> +	foreach m $mem {
> +	    set ${m}_new [capture_command_output "x/x $m" ""]
> +	}
> +
> +	# Execute in reverse to before the instruction sequence.
> +	gdb_test_no_output "set exec-direction reverse"
> +
> +	gdb_continue_to_breakpoint "reversed execution of instruction sequence" \
> +	    [multi_line ".*/aarch64-lrcpc3.c:${::decimal}" \
> +		"${::decimal}\[ \t\]+__asm__ volatile \\(\"${insn} .*\".*"]
> +
> +	# Record the final memory and register values.
> +	foreach r $regs {
> +	    set ${r}_final [capture_command_output "info register $r" ""]
> +	}
> +	foreach v $vars {
> +	    set ${v}_final [get_valueof "/x" "$v" "unable to read $v" \
> +		"get $v final value"]
> +	}
> +	foreach m $mem {
> +	    set ${m}_final [capture_command_output "x/x $m" ""]
> +	}
> +
> +	foreach v [concat $same_reg $same_var $same_mem] {
> +	    gdb_assert ![string compare [set ${v}_initial] [set ${v}_new]] \
> +		"check $v initial value versus $v new value"
> +	    gdb_assert ![string compare [set ${v}_initial] [set ${v}_final]] \
> +		"check $v initial value versus $v final value"
> +	}
> +
> +	foreach v [concat $diff_reg $diff_var $diff_mem] {
> +	    gdb_assert [string compare [set ${v}_initial] [set ${v}_new]] \
> +		"check $v initial value versus $v new value"
> +	    gdb_assert ![string compare [set ${v}_initial] [set ${v}_final]] \
> +		"check $v initial value versus $v final value"
> +	}
> +
> +	# Restore forward execution and go to end of recording.
> +	gdb_test_no_output "set exec-direction forward"
> +	gdb_test "record goto end" \
> +	    [multi_line \
> +		"Go forward to insn number ${::decimal}" \
> +		"#0  main \\(\\) at .*/aarch64-lrcpc3.c:${::decimal}" \
> +		".*"]
> +    }
> +}
> +
> +set ldiapp_cases {
> +	{ ldiapp-0 { x19 x20     } { } { } { x21 } { } { src } }
> +	{ ldiapp-1 { w19 w20     } { } { } { x21 } { } { src } }
> +	{ ldiapp-2 { x19 x20 x21 } { } { } {     } { } { src } }
> +	{ ldiapp-3 { w19 w20 x21 } { } { } {     } { } { src } }
> +	{ ldiapp-4 { x21 x20     } { } { } {     } { } { src } }
> +	{ ldiapp-5 { w21 w20     } { } { } {     } { } { src } }
> +}
> +
> +set stilp_cases {
> +	{ stilp-0  {     } { } { src } { x19 x20 x21 } { } { } }
> +	{ stilp-1  {     } { } { src } { w19 w20 x21 } { } { } }
> +	{ stilp-2  { x21 } { } { src } { x19 x20     } { } { } }
> +	{ stilp-3  { x21 } { } { src } { x19 x20     } { } { } }
> +	{ stilp-4  {     } { } { src } { x20     x21 } { } { } }
> +	{ stilp-5  {     } { } { src } { w20     x21 } { } { } }
> +}
> +
> +set ldapr_stlr_cases {
> +	{ ldapr-0  { x19 x21 } { } {     } {     } { } { src } }
> +	{ ldapr-1  { w19 x21 } { } {     } {     } { } { src } }
> +	{ stlr-0   {     x21 } { } { src } { x19 } { } {     } }
> +	{ stlr-1   {     x21 } { } { src } { w19 } { } {     } }
> +}
> +
> +set ldap1_stl1_cases {
> +	{ ldap1-0  { v22 } { } {     } { x21 } { } { src } }
> +	{ stl1-0   {     } { } { src } { v22 } { } {     } }
> +}
> +
> +set ldapur_stlur_cases {
> +	{ ldapur-0 { v22 } { } {     } { x21     } { } { src } }
> +	{ stlur-0  {     } { } { src } { x21 v22 } { } {     } }
> +	{ ldapur-1 { v22 } { } {     } {     x21 } { } { src } }
> +	{ stlur-1  {     } { } { src } { x21 v22 } { } {     } }
> +	{ ldapur-2 { v22 } { } {     } {     x21 } { } { src } }
> +	{ stlur-2  {     } { } { src } { x21 v22 } { } {     } }
> +	{ ldapur-3 { v22 } { } {     } {     x21 } { } { src } }
> +	{ stlur-3  {     } { } { src } { x21 v22 } { } {     } }
> +	{ ldapur-4 { v22 } { } {     } {     x21 } { } { src } }
> +	{ stlur-4  {     } { } { src } { x21 v22 } { } {     } }
> +	{ ldapur-5 { v22 } { } {     } {     x21 } { } { src } }
> +	{ stlur-5  {     } { } { src } { x21 v22 } { } {     } }
> +	{ ldapur-6 { v22 } { } {     } {     x21 } { } { src } }
> +	{ stlur-6  {     } { } { src } { x21 v22 } { } {     } }
> +}
All of the cases have the third and second-to-last argument list as 
empty. They can probably be dropped there, and dropped as arguments to 
the test_single_asm proc.
> +
> +set all_cases [concat \
> +    $ldiapp_cases $stilp_cases $ldapr_stlr_cases \
> +    $ldap1_stl1_cases $ldapur_stlur_cases]
> +
> +foreach c $all_cases {
> +    lassign $c name diff_reg diff_var diff_mem same_reg same_var same_mem
> +    test_single_asm $name $diff_reg $diff_var $diff_mem $same_reg $same_var $same_mem
> +}


-- 
Cheers,
Guinevere Larsen
It/she


  parent reply	other threads:[~2026-05-26 14:30 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-12 23:17 [PATCH v2 0/4] gdb/aarch64: record/replay support for AArch64 Ezra.Sitorus
2026-05-12 23:17 ` [PATCH v2 1/4] gdb/aarch64: record/replay support for LRCPC3 Ezra.Sitorus
2026-05-15  3:50   ` Thiago Jung Bauermann
2026-05-15 11:45     ` Ezra Sitorus
2026-05-15 23:03       ` Thiago Jung Bauermann
2026-05-26 14:30   ` Guinevere Larsen [this message]
2026-05-26 23:36     ` Ezra Sitorus
2026-05-27  0:05       ` Thiago Jung Bauermann
2026-05-12 23:17 ` [PATCH v2 2/4] gdb/aarch64: Test record/replay support for CSSC Ezra.Sitorus
2026-05-15  3:50   ` Thiago Jung Bauermann
2026-05-26 19:40     ` Guinevere Larsen
2026-05-26 16:40   ` Guinevere Larsen
2026-05-12 23:17 ` [PATCH v2 3/4] gdb/aarch64: record/replay support for RPRFM, PRFM (reg), PRFUM Ezra.Sitorus
2026-05-15  3:56   ` Thiago Jung Bauermann
2026-05-12 23:17 ` [PATCH v2 4/4] gdb/aarch64: record/replay support for LSE128 Ezra.Sitorus
2026-05-15  3:58   ` Thiago Jung Bauermann
2026-05-26 19:40   ` Guinevere Larsen
2026-05-15  3:59 ` [PATCH v2 0/4] gdb/aarch64: record/replay support for AArch64 Thiago Jung Bauermann

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=f741185a-f505-439f-9951-53c2a4d6003a@redhat.com \
    --to=guinevere@redhat.com \
    --cc=Ezra.Sitorus@arm.com \
    --cc=gdb-patches@sourceware.org \
    --cc=luis.machado.foss@gmail.com \
    --cc=thiago.bauermann@linaro.org \
    /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