public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Andrew Burgess <aburgess@redhat.com>
To: Simon Marchi <simon.marchi@polymtl.ca>, gdb-patches@sourceware.org
Cc: Simon Marchi <simon.marchi@efficios.com>
Subject: Re: [PATCH v2 2/5] gdb/testsuite: refactor gdb.threads/detach-step-over.exp
Date: Mon, 28 Nov 2022 12:05:22 +0000	[thread overview]
Message-ID: <87iliz2sjh.fsf@redhat.com> (raw)
In-Reply-To: <20221121171213.1414366-3-simon.marchi@polymtl.ca>

Simon Marchi <simon.marchi@polymtl.ca> writes:

> From: Andrew Burgess <aburgess@redhat.com>
>
> Factor out some bits of gdb.threads/detach-step-over.exp to procs in
> preparation to adding some new variations of the test.  Rename the
> existing "test" proc and make it use proc_with_prefix.
>
> Co-Authored-By: Simon Marchi <simon.marchi@efficios.com>
> Change-Id: Ib4412545c81c8556029e0f7bfa9dd48d7a9f3189

LGTM.  I guess if we're both happy then this is OK to go in without
anyone other review.

Thanks,
Andrew

> ---
>  .../gdb.threads/detach-step-over.exp          | 238 ++++++++++--------
>  1 file changed, 138 insertions(+), 100 deletions(-)
>
> diff --git a/gdb/testsuite/gdb.threads/detach-step-over.exp b/gdb/testsuite/gdb.threads/detach-step-over.exp
> index 917be2ef3782..ad9b08f549ea 100644
> --- a/gdb/testsuite/gdb.threads/detach-step-over.exp
> +++ b/gdb/testsuite/gdb.threads/detach-step-over.exp
> @@ -56,11 +56,11 @@ standard_testfile
>  
>  set bp_lineno [gdb_get_line_number "Set breakpoint here"]
>  
> -# The test proper.  See description above.
> -proc test {condition_eval target_non_stop non_stop displaced} {
> -    # Number of threads started by the program.
> -    set n_threads 10
> +# Number of threads started by the program.
> +set n_threads 10
>  
> +# Start GDB, configuring various settings according to the arguments.
> +proc start_gdb_for_test {condition_eval target_non_stop non_stop displaced} {
>      save_vars { ::GDBFLAGS } {
>  	append ::GDBFLAGS " -ex \"maint set target-non-stop $target_non_stop\""
>  	append ::GDBFLAGS " -ex \"set non-stop $non_stop\""
> @@ -69,10 +69,137 @@ proc test {condition_eval target_non_stop non_stop displaced} {
>  	clean_restart $::binfile
>      }
>  
> +    gdb_test_no_output "set breakpoint condition-evaluation $condition_eval"
> +}
> +
> +# Use the 'attach' command to attach to process with pid TESTPID.  Return true
> +# if we believe GDB has attached and we are back at the GDB prompt, otherwise,
> +# return false.
> +proc attach_to {testpid} {
> +    with_timeout_factor 2 {
> +	set attached 0
> +	set saw_attaching 0
> +	gdb_test_multiple "attach $testpid" "attach" {
> +	    -re "Attaching to program.*process $testpid\r\n" {
> +		set saw_attaching 1
> +		exp_continue
> +	    }
> +	    -re "new threads in iteration" {
> +		# Seen when "set debug libthread_db" is on.
> +		exp_continue
> +	    }
> +	    -re "Reading symbols from|Expanding full symbols from" {
> +		# Prevent -readnow timeout.
> +		exp_continue
> +	    }
> +	    -re "is a zombie - the process has already terminated.*$::gdb_prompt " {
> +		fail $gdb_test_name
> +	    }
> +	    -re "Unable to attach: .*$::gdb_prompt " {
> +		fail $gdb_test_name
> +	    }
> +	    -re "\r\n$::gdb_prompt " {
> +		if { $saw_attaching } {
> +		    set attached 1
> +		    pass $gdb_test_name
> +		} else {
> +		    fail $gdb_test_name
> +		}
> +	    }
> +	}
> +    }
> +
> +    return $attached
> +}
> +
> +# After attaching to a multi-threaded inferior in non-stop mode, we expect to
> +# see a stop message from each thread.  This proc waits for all of these stop
> +# messages.  TID_RE is a regexp used to match the thread-id of the stopped
> +# thread.
> +#
> +# Return true if we saw a stop from each of the expected threads (based on the
> +# global N_THREADS value), otherwise, return false.
> +proc check_stops_after_non_stop_attach {tid_re} {
> +    set any "\[^\r\n\]*"
> +
> +    # In non-stop, we will see one stop per thread after the prompt.
> +    set stops 0
> +    set test "seen all stops"
> +    for {set thread 1} { $thread <= $::n_threads } { incr thread } {
> +	if {[gdb_test_multiple "" $test {
> +	    -re "Thread ${tid_re} ${any} stopped" {
> +		incr stops
> +	    }
> +	}] != 0} {
> +	    break
> +	}
> +    }
> +
> +    # If we haven't seen all stops, then the
> +    # gdb_test_multiple in the loop above will have
> +    # already issued a FAIL.
> +    if {$stops != $::n_threads} {
> +	return false
> +    }
> +    pass $test
> +    return true
> +}
> +
> +# Prepare for a single test iteration.  TESTPID is the pid of the process GDB
> +# will be attached too.  NON_STOP indicates if GDB is configured in non-stop
> +# mode or not.  ATTEMPT is the current attempt number, and ATTEMPTS is the
> +# maximum number of attempts we plan to run.  TID_RE is a string used to match
> +# against a thread-id in GDB's stop messages.
> +#
> +# Return true if everything is prepared correctly, otherwise return false.
> +proc prepare_test_iter {testpid non_stop attempt attempts tid_re} {
> +    if {![attach_to $testpid]} {
> +	return false
> +    }
> +
> +    if {$non_stop} {
> +	if {![check_stops_after_non_stop_attach $tid_re]} {
> +	    return false
> +	}
> +    }
> +
> +    gdb_test "break ${::srcfile}:${::bp_lineno} if 0" "Breakpoint.*" \
> +	"break LOC if 0"
> +
> +    if {$attempt < $attempts} {
> +	# Kick the time out timer for another round.
> +	gdb_test "print again = 1" " = 1" "reset timer in the inferior"
> +	# Show the time we had left in the logs, in case
> +	# something goes wrong.
> +	gdb_test "print seconds_left" " = .*"
> +    }
> +
> +    if {$non_stop} {
> +	set cont_cmd "continue -a &"
> +    } else {
> +	set cont_cmd "continue &"
> +    }
> +
> +    set cont_cmd_re [string_to_regexp $cont_cmd]
> +    gdb_test_multiple $cont_cmd "" {
> +	-re "^$cont_cmd_re\r\nContinuing\.\r\n$::gdb_prompt " {
> +	    pass $gdb_test_name
> +	}
> +    }
> +
> +    # Wait a bit, to give time for the threads to hit the
> +    # breakpoint.
> +    sleep 1
> +
> +    return true
> +}
> +
> +# The test proper.  See the description at the top of the file.
> +proc_with_prefix test_detach_command {condition_eval target_non_stop non_stop displaced} {
>      set test_spawn_id [spawn_wait_for_attach $::binfile]
>      set testpid [spawn_id_get_pid $test_spawn_id]
>  
> -    set any "\[^\r\n\]*"
> +    start_gdb_for_test $condition_eval $target_non_stop $non_stop $displaced
>  
>      gdb_test "add-inferior" "Added inferior 2.*"
>      gdb_test "inferior 2" "Switching to .*"
> @@ -84,8 +211,6 @@ proc test {condition_eval target_non_stop non_stop displaced} {
>  	return
>      }
>  
> -    gdb_test_no_output "set breakpoint condition-evaluation $condition_eval"
> -
>      # Get the PID of the test process.
>      set pid_inf2 ""
>      gdb_test_multiple "p mypid" "get pid of inferior 2" {
> @@ -100,101 +225,12 @@ proc test {condition_eval target_non_stop non_stop displaced} {
>  	with_test_prefix "iter $attempt" {
>  	    gdb_test "inferior 1" "Switching to .*"
>  
> -	    with_timeout_factor 2 {
> -		set attached 0
> -		set saw_attaching 0
> -		set eperm 0
> -		set test "attach"
> -		gdb_test_multiple "attach $testpid" $test {
> -		    -re "Attaching to program.*process $testpid\r\n" {
> -			set saw_attaching 1
> -			exp_continue
> -		    }
> -		    -re "new threads in iteration" {
> -			# Seen when "set debug libthread_db" is on.
> -			exp_continue
> -		    }
> -		    -re "Reading symbols from|Expanding full symbols from" {
> -			# Prevent -readnow timeout.
> -			exp_continue
> -		    }
> -		    -re "is a zombie - the process has already terminated.*$::gdb_prompt " {
> -			fail $gdb_test_name
> -		    }
> -		    -re "Unable to attach: .*$::gdb_prompt " {
> -			fail $gdb_test_name
> -		    }
> -		    -re "\r\n$::gdb_prompt " {
> -			if { $saw_attaching } {
> -			    set attached 1
> -			    pass $test
> -			} else {
> -			    fail $test
> -			}
> -		    }
> -		}
> -	    }
> -
> -	    if {!$attached} {
> +	    if {![prepare_test_iter $testpid $non_stop \
> +		      $attempt $attempts "$::decimal\.$::decimal"]} {
>  		kill_wait_spawned_process $test_spawn_id
>  		return
>  	    }
>  
> -	    if {$non_stop} {
> -		# In non-stop, we will see one stop per thread after
> -		# the prompt.
> -		set stops 0
> -		set tid_re "$::decimal\.$::decimal"
> -		set test "seen all stops"
> -		for {set thread 1} { $thread <= $n_threads } { incr thread } {
> -		    if {[gdb_test_multiple "" $test {
> -			-re "Thread ${tid_re} ${any} stopped" {
> -			    incr stops
> -			}
> -		    }] != 0} {
> -			break
> -		    }
> -		}
> -
> -		# If we haven't seen all stops, then the
> -		# gdb_test_multiple in the loop above will have
> -		# already issued a FAIL.
> -		if {$stops != $n_threads} {
> -		    kill_wait_spawned_process $test_spawn_id
> -		    return
> -		}
> -		pass $test
> -	    }
> -
> -	    # Set threads stepping over a breakpoint continuously.
> -	    gdb_test "break $::srcfile:$::bp_lineno if 0" "Breakpoint.*" \
> -		"break LOC if 0"
> -
> -	    if {$attempt < $attempts} {
> -		# Kick the time out timer for another round.
> -		gdb_test "print again = 1" " = 1" "reset timer in the inferior"
> -		# Show the time we had left in the logs, in case
> -		# something goes wrong.
> -		gdb_test "print seconds_left" " = .*"
> -	    }
> -
> -	    if {$non_stop} {
> -		set cont_cmd "continue -a &"
> -	    } else {
> -		set cont_cmd "continue &"
> -	    }
> -
> -	    set cont_cmd_re [string_to_regexp $cont_cmd]
> -	    gdb_test_multiple $cont_cmd "" {
> -		-re "^$cont_cmd_re\r\nContinuing\.\r\n$::gdb_prompt " {
> -		    pass $gdb_test_name
> -		}
> -	    }
> -
> -	    # Wait a bit, to give time for the threads to hit the
> -	    # breakpoint.
> -	    sleep 1
> -
>  	    set running_count 0
>  	    set interrupted 0
>  	    gdb_test_multiple "info threads" "all threads running" {
> @@ -219,7 +255,8 @@ proc test {condition_eval target_non_stop non_stop displaced} {
>  		    }
>  		}
>  		-re "$::gdb_prompt $" {
> -		    gdb_assert {$running_count == ($n_threads + 1) * 2} $gdb_test_name
> +		    gdb_assert {$running_count == ($::n_threads + 1) * 2} \
> +			$gdb_test_name
>  		}
>  	    }
>  
> @@ -292,7 +329,8 @@ foreach_with_prefix breakpoint-condition-evaluation {"host" "target"} {
>  	    }
>  
>  	    foreach_with_prefix displaced {"off" "auto"} {
> -		test ${breakpoint-condition-evaluation} ${target-non-stop} ${non-stop} ${displaced}
> +		test_detach_command ${breakpoint-condition-evaluation} \
> +		    ${target-non-stop} ${non-stop} ${displaced}
>  	    }
>  	}
>      }
> -- 
> 2.38.1


  reply	other threads:[~2022-11-28 12:05 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-21 17:12 [PATCH v2 0/5] Fix some commit_resumed_state assertion failures (PR 28275) Simon Marchi
2022-11-21 17:12 ` [PATCH v2 1/5] gdb/testsuite: remove global declarations in gdb.threads/detach-step-over.exp Simon Marchi
2022-11-28 11:43   ` Andrew Burgess
2022-11-21 17:12 ` [PATCH v2 2/5] gdb/testsuite: refactor gdb.threads/detach-step-over.exp Simon Marchi
2022-11-28 12:05   ` Andrew Burgess [this message]
2022-11-21 17:12 ` [PATCH v2 3/5] gdb: fix assert when quitting GDB while a thread is stepping Simon Marchi
2022-11-28 12:11   ` Andrew Burgess
2022-11-28 13:03     ` Simon Marchi
2022-11-21 17:12 ` [PATCH v2 4/5] gdbserver: switch to right process in find_one_thread Simon Marchi
2022-11-28 13:30   ` Andrew Burgess
2022-11-28 14:09     ` Simon Marchi
2022-11-21 17:12 ` [PATCH v2 5/5] gdb: disable commit resumed in target_kill Simon Marchi
2022-11-28 13:44   ` Andrew Burgess
2022-11-28 14:12     ` Simon Marchi

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=87iliz2sjh.fsf@redhat.com \
    --to=aburgess@redhat.com \
    --cc=gdb-patches@sourceware.org \
    --cc=simon.marchi@efficios.com \
    --cc=simon.marchi@polymtl.ca \
    /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).