From: Luis Machado <luis.machado@arm.com>
To: Guinevere Larsen <blarsen@redhat.com>, gdb-patches@sourceware.org
Cc: Andrew Burgess <aburgess@redhat.com>
Subject: Re: [PINGv3][PATCH v4] gdb/testsuite: add test for backtracing for threaded inferiors from a corefile
Date: Wed, 24 Jan 2024 12:35:00 +0000 [thread overview]
Message-ID: <33d4e5ab-27dd-4008-962f-abd02436847d@arm.com> (raw)
In-Reply-To: <31ade58a-3cd2-4d7b-a282-d0160f4f7401@redhat.com>
Hi Guinevere,
Sorry for the delay.
On 1/24/24 09:34, Guinevere Larsen wrote:
> Ping!
> On 09/01/2024 12:50, Guinevere Larsen wrote:
>> Ping!
>> On 20/12/2023 10:40, Guinevere Larsen wrote:
>>> Ping!
>>>
>>> --
>>> Cheers,
>>> Guinevere Larsen
>>> She/Her/Hers
>>>
>>> On 04/12/2023 18:33, Guinevere Larsen wrote:
>>>> This patch is based on an out-of-tree patch that fedora has been
>>>> carrying for a while. It tests if GDB is able to properly unwind a
>>>> threaded program in the following situations:
>>>> * regular threads
>>>> * in a signal handler
>>>> * in a signal handler executing on an alternate stack
>>>>
>>>> And the final frame can either be in a syscall or in an infinite loop.
>>>>
>>>> The test works by running the inferior until a crash to generate a
>>>> corefile, or until right before the crash. Then applies a backtrace to
>>>> all threads to see if any frame can't be identified, and the order of
>>>> the threads in GDB. Finally, it goes thread by thread and tries to
>>>> collect a large part of the backtrace, to confirm that everything is
>>>> being unwound correctly.
>>>>
>>>> Co-Authored-By: Andrew Burgess <aburgess@redhat.com>
>>>> Reviewed-By: Luis Machado <luis.machado@arm.com>
>>>>
>>>> ---
>>>>
>>>> Changes for v4:
>>>> * Luis mentioned that my strategy for starting the inferior didn't work
>>>> with native-extended testing. Changed to use runto_main instead
>>>> * Improved comments in the exp file based on Andrew's comments
>>>> * Minor cleanups with regards to TCL usage
>>>> ---
>>>> gdb/testsuite/gdb.threads/threadcrash.c | 443 ++++++++++++++++++++++
>>>> gdb/testsuite/gdb.threads/threadcrash.exp | 233 ++++++++++++
>>>> 2 files changed, 676 insertions(+)
>>>> create mode 100644 gdb/testsuite/gdb.threads/threadcrash.c
>>>> create mode 100644 gdb/testsuite/gdb.threads/threadcrash.exp
>>>>
>>>> diff --git a/gdb/testsuite/gdb.threads/threadcrash.c b/gdb/testsuite/gdb.threads/threadcrash.c
>>>> new file mode 100644
>>>> index 00000000000..e476ae7b07d
>>>> --- /dev/null
>>>> +++ b/gdb/testsuite/gdb.threads/threadcrash.c
>>>> @@ -0,0 +1,443 @@
>>>> +/* This testcase is part of GDB, the GNU debugger.
>>>> +
>>>> + Copyright 2023 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/>. */
>>>> +
>>>> +#include <pthread.h>
>>>> +#include <assert.h>
>>>> +#include <stdlib.h>
>>>> +#include <signal.h>
>>>> +#include <unistd.h>
>>>> +
>>>> +/* The delay that the main thread gives once all the worker threads have
>>>> + reached the barrier before the main thread enters the function on which
>>>> + GDB will have placed a breakpoint. */
>>>> +
>>>> +#define MAIN_THREAD_DELAY 2
>>>> +
>>>> +/* The maximum time we allow this test program to run for before an alarm
>>>> + signal is sent and everything will exit. */
>>>> +#define WATCHDOG_ALARM_TIME 600
>>>> +
>>>> +/* Aliases for the signals used within this script. Each signal
>>>> + corresponds to an action (from the FINAL_ACTION enum) that the signal
>>>> + handler will perform. */
>>>> +
>>>> +#define SPIN_SIGNAL SIGUSR1
>>>> +#define SYSCALL_SIGNAL SIGUSR2
>>>> +
>>>> +/* Describe the final action that a thread should perform. */
>>>> +
>>>> +enum final_action
>>>> + {
>>>> + /* Thread should spin in an infinite loop. */
>>>> + SPIN = 0,
>>>> +
>>>> + /* Thread should block in a syscall. */
>>>> + SYSCALL,
>>>> +
>>>> + /* This is just a marker to allow for looping over the enum. */
>>>> + LAST_ACTION
>>>> + };
>>>> +
>>>> +/* Where should the thread perform this action? */
>>>> +
>>>> +enum exec_location
>>>> + {
>>>> + /* Just a normal thread, on a normal stack. */
>>>> + NORMAL = 0,
>>>> +
>>>> + /* In a signal handler, but use the normal stack. */
>>>> + SIGNAL_HANDLER,
>>>> +
>>>> + /* In a signal handler using an alternative stack. */
>>>> + SIGNAL_ALT_STACK,
>>>> +
>>>> + /* This is just a marker to allow for looping over the enum. */
>>>> + LAST_LOCACTION
>>>> + };
>>>> +
>>>> +/* A descriptor for a single thread job. We create a new thread for each
>>>> + job_description. */
>>>> +
>>>> +struct job_description
>>>> +{
>>>> + /* What action should this thread perform. */
>>>> + enum final_action action;
>>>> +
>>>> + /* Where should the thread perform the action. */
>>>> + enum exec_location location;
>>>> +
>>>> + /* The actual thread handle, so we can join with the thread. */
>>>> + pthread_t thread;
>>>> +};
>>>> +
>>>> +/* A pthread barrier, used to (try) and synchronise the threads. */
>>>> +pthread_barrier_t global_barrier;
>>>> +
>>>> +/* Return a list of jobs, and place the length of the list in *COUNT. */
>>>> +
>>>> +struct job_description *
>>>> +get_job_list (int *count)
>>>> +{
>>>> + /* The number of jobs. */
>>>> + int num = LAST_ACTION * LAST_LOCACTION;
>>>> +
>>>> + /* The uninitialised array of jobs. */
>>>> + struct job_description *list
>>>> + = malloc (num * sizeof (struct job_description));
>>>> + assert (list != NULL);
>>>> +
>>>> + /* Fill the array with all possible jobs. */
>>>> + for (int i = 0; i < (int) LAST_ACTION; ++i)
>>>> + for (int j = 0; j < (int) LAST_LOCACTION; ++j)
>>>> + {
>>>> + int idx = (i * LAST_LOCACTION) + j;
>>>> + list[idx].action = (enum final_action) i;
>>>> + list[idx].location = (enum exec_location) j;
>>>> + }
>>>> +
>>>> + /* Return the array of jobs. */
>>>> + *count = num;
>>>> + return list;
>>>> +}
>>>> +
>>>> +/* This function should never be called. If it is then an assertion will
>>>> + trigger. */
>>>> +
>>>> +void
>>>> +assert_not_reached (void)
>>>> +{
>>>> + assert (0);
>>>> +}
>>>> +
>>>> +/* The function for a SPIN action. Just spins in a loop. The LOCATION
>>>> + argument exists so GDB can identify the expected context for this
>>>> + function. */
>>>> +
>>>> +void
>>>> +do_spin_task (enum exec_location location)
>>>> +{
>>>> + (void) location;
>>>> +
>>>> + /* Let everyone know that we're about to perform our action. */
>>>> + int res = pthread_barrier_wait (&global_barrier);
>>>> + assert (res == PTHREAD_BARRIER_SERIAL_THREAD || res == 0);
>>>> +
>>>> + while (1)
>>>> + {
>>>> + /* Nothing. */
>>>> + }
>>>> +}
>>>> +
>>>> +/* The function for a SYSCALL action. Just spins in a loop. The LOCATION
>>>> + argument exists so GDB can identify the expected context for this
>>>> + function. */
>>>> +
>>>> +void
>>>> +do_syscall_task (enum exec_location location)
>>>> +{
>>>> + (void) location;
>>>> +
>>>> + /* Let everyone know that we're about to perform our action. */
>>>> + int res = pthread_barrier_wait (&global_barrier);
>>>> + assert (res == PTHREAD_BARRIER_SERIAL_THREAD || res == 0);
>>>> +
>>>> + sleep (600);
>>>> +}
>>>> +
>>>> +/* Return the required size for a sigaltstack. We start with a single
>>>> + page, but do check against the system defined minimums. We don't run
>>>> + much on the alternative stacks, so we don't need a huge one. */
>>>> +
>>>> +size_t
>>>> +get_stack_size (void)
>>>> +{
>>>> + size_t size = getpagesize (); /* Arbitrary starting size. */
>>>> + if (size < SIGSTKSZ)
>>>> + size = SIGSTKSZ;
>>>> + if (size < MINSIGSTKSZ)
>>>> + size = MINSIGSTKSZ;
>>>> + return size;
>>>> +}
>>>> +
>>>> +/* A descriptor for an alternative stack. */
>>>> +
>>>> +struct stack_descriptor
>>>> +{
>>>> + /* The base address of the alternative stack. This is the address that
>>>> + must be freed to release the memory used by this stack. */
>>>> + void *base;
>>>> +
>>>> + /* The size of this alternative stack. Tracked just so we can query this
>>>> + from GDB. */
>>>> + size_t size;
>>>> +};
>>>> +
>>>> +/* Install an alternative signal stack. Return a descriptor for the newly
>>>> + allocated alternative stack. */
>>>> +
>>>> +struct stack_descriptor
>>>> +setup_alt_stack (void)
>>>> +{
>>>> + size_t stack_size = get_stack_size ();
>>>> +
>>>> + void *stack_area = malloc (stack_size);
>>>> +
>>>> + stack_t stk;
>>>> + stk.ss_sp = stack_area;
>>>> + stk.ss_flags = 0;
>>>> + stk.ss_size = stack_size;
>>>> +
>>>> + int res = sigaltstack (&stk, NULL);
>>>> + assert (res == 0);
>>>> +
>>>> + struct stack_descriptor desc;
>>>> + desc.base = stack_area;
>>>> + desc.size = stack_size;
>>>> +
>>>> + return desc;
>>>> +}
>>>> +
>>>> +/* Return true (non-zero) if we are currently on the alternative stack,
>>>> + otherwise, return false (zero). */
>>>> +
>>>> +int
>>>> +on_alt_stack_p (void)
>>>> +{
>>>> + stack_t stk;
>>>> + int res = sigaltstack (NULL, &stk);
>>>> + assert (res == 0);
>>>> +
>>>> + return (stk.ss_flags & SS_ONSTACK) != 0;
>>>> +}
>>>> +
>>>> +/* The signal handler function. All signals call here, so we use SIGNO
>>>> + (the signal that was delivered) to decide what action to perform. This
>>>> + function might, or might not, have been called on an alternative signal
>>>> + stack. */
>>>> +
>>>> +void
>>>> +signal_handler (int signo)
>>>> +{
>>>> + enum exec_location location
>>>> + = on_alt_stack_p () ? SIGNAL_ALT_STACK : SIGNAL_HANDLER;
>>>> +
>>>> + switch (signo)
>>>> + {
>>>> + case SPIN_SIGNAL:
>>>> + do_spin_task (location);
>>>> + break;
>>>> +
>>>> + case SYSCALL_SIGNAL:
>>>> + do_syscall_task (location);
>>>> + break;
>>>> +
>>>> + default:
>>>> + assert_not_reached ();
>>>> + }
>>>> +}
>>>> +
>>>> +/* The thread worker function. ARG is a job_description pointer which
>>>> + describes what this thread is expected to do. This function always
>>>> + returns a NULL pointer. */
>>>> +
>>>> +void *
>>>> +thread_function (void *arg)
>>>> +{
>>>> + struct job_description *job = (struct job_description *) arg;
>>>> + struct stack_descriptor desc = { NULL, 0 };
>>>> + int sa_flags = 0;
>>>> +
>>>> + switch (job->location)
>>>> + {
>>>> + case NORMAL:
>>>> + /* This thread performs the worker action on the current thread,
>>>> + select the correct worker function based on the requested
>>>> + action. */
>>>> + switch (job->action)
>>>> + {
>>>> + case SPIN:
>>>> + do_spin_task (NORMAL);
>>>> + break;
>>>> +
>>>> + case SYSCALL:
>>>> + do_syscall_task (NORMAL);
>>>> + break;
>>>> +
>>>> + default:
>>>> + assert_not_reached ();
>>>> + }
>>>> + break;
>>>> +
>>>> + case SIGNAL_ALT_STACK:
>>>> + /* This thread is to perform its action in a signal handler on the
>>>> + alternative stack. Install the alternative stack now, and then
>>>> + fall through to the normal signal handler location code. */
>>>> + desc = setup_alt_stack ();
>>>> + assert (desc.base != NULL);
>>>> + assert (desc.size > 0);
>>>> + sa_flags = SA_ONSTACK;
>>>> +
>>>> + /* Fall through. */
>>>> + case SIGNAL_HANDLER:
>>>> + {
>>>> + /* This thread is to perform its action in a signal handler. We
>>>> + might have just installed an alternative signal stack. */
>>>> + int signo, res;
>>>> +
>>>> + /* Select the correct signal number so that the signal handler will
>>>> + perform the required action. */
>>>> + switch (job->action)
>>>> + {
>>>> + case SPIN:
>>>> + signo = SPIN_SIGNAL;
>>>> + break;
>>>> +
>>>> + case SYSCALL:
>>>> + signo = SYSCALL_SIGNAL;
>>>> + break;
>>>> +
>>>> + default:
>>>> + assert_not_reached ();
>>>> + }
>>>> +
>>>> + /* Now setup the signal handler. */
>>>> + struct sigaction sa;
>>>> + sa.sa_handler = signal_handler;
>>>> + sigfillset (&sa.sa_mask);
>>>> + sa.sa_flags = sa_flags;
>>>> + res = sigaction (signo, &sa, NULL);
>>>> + assert (res == 0);
>>>> +
>>>> + /* Send the signal to this thread. */
>>>> + res = pthread_kill (job->thread, signo);
>>>> + assert (res == 0);
>>>> + }
>>>> + break;
>>>> +
>>>> + default:
>>>> + assert_not_reached ();
>>>> + };
>>>> +
>>>> + /* Free the alt-stack if we allocated one, if not DESC.BASE will be
>>>> + NULL so this call is fine. */
>>>> + free (desc.base);
>>>> +
>>>> + /* Thread complete. */
>>>> + return NULL;
>>>> +}
>>>> +
>>>> +void
>>>> +start_job (struct job_description *job)
>>>> +{
>>>> + int res;
>>>> +
>>>> + res = pthread_create (&job->thread, NULL, thread_function, job);
>>>> + assert (res == 0);
>>>> +}
>>>> +
>>>> +/* Join with the thread for JOB. This will block until the thread for JOB
>>>> + has finished. */
>>>> +
>>>> +void
>>>> +finalise_job (struct job_description *job)
>>>> +{
>>>> + int res;
>>>> + void *retval;
>>>> +
>>>> + res = pthread_join (job->thread, &retval);
>>>> + assert (res == 0);
>>>> + assert (retval == NULL);
>>>> +}
>>>> +
>>>> +/* Function that GDB can place a breakpoint on. */
>>>> +
>>>> +void
>>>> +breakpt (void)
>>>> +{
>>>> + /* Nothing. */
>>>> +}
>>>> +
>>>> +/* Function that triggers a crash, if the user has setup their environment
>>>> + correctly this will dump a core file, which GDB can then examine. */
>>>> +
>>>> +void
>>>> +crash_function (void)
>>>> +{
>>>> + volatile int *p = 0;
>>>> + volatile int n = *p;
>>>> + (void) n;
>>>> +}
>>>> +
>>>> +/* Entry point. */
>>>> +
>>>> +int
>>>> +main ()
>>>> +{
>>>> + int job_count, res;
>>>> + struct job_description *jobs = get_job_list (&job_count);
>>>> +
>>>> + /* This test is going to park some threads inside infinite loops. Just
>>>> + in case this program is left running, install an alarm that will cause
>>>> + everything to exit. */
>>>> + alarm (WATCHDOG_ALARM_TIME);
>>>> +
>>>> + /* We want each worker thread (of which there are JOB_COUNT) plus the
>>>> + main thread (hence + 1) to wait at the barrier. */
>>>> + res = pthread_barrier_init (&global_barrier, NULL, job_count + 1);
>>>> + assert (res == 0);
>>>> +
>>>> + /* Start all the jobs. */
>>>> + for (int i = 0; i < job_count; ++i)
>>>> + start_job (&jobs[i]);
>>>> +
>>>> + /* Notify all the worker threads that we're waiting for them. */
>>>> + res = pthread_barrier_wait (&global_barrier);
>>>> + assert (res == PTHREAD_BARRIER_SERIAL_THREAD || res == 0);
>>>> +
>>>> + /* All we know at this point is that all the worker threads have reached
>>>> + the barrier, which is just before they perform their action. But we
>>>> + really want them to start their action.
>>>> +
>>>> + There's really no way we can be 100% certain that the worker threads
>>>> + have started their action, all we can do is wait for a short while and
>>>> + hope that the machine we're running on is not too slow. */
>>>> + sleep (MAIN_THREAD_DELAY);
>>>> +
>>>> + /* A function that GDB can place a breakpoint on. By the time we get
>>>> + here we are as sure as we can be that all of the worker threads have
>>>> + started and are in their worker action (spinning, or syscall). */
>>>> + breakpt ();
>>>> +
>>>> + /* If GDB is not attached then this function will cause a crash, which
>>>> + can be used to dump a core file, which GDB can then analyse. */
>>>> + crash_function ();
>>>> +
>>>> + /* Due to the crash we never expect to get here. Plus the worker actions
>>>> + never terminate. But for completeness, here's where we join with all
>>>> + the worker threads. */
>>>> + for (int i = 0; i < job_count; ++i)
>>>> + finalise_job (&jobs[i]);
>>>> +
>>>> + /* Cleanup the barrier. */
>>>> + res = pthread_barrier_destroy (&global_barrier);
>>>> + assert (res == 0);
>>>> +
>>>> + /* And clean up the jobs list. */
>>>> + free (jobs);
>>>> +
>>>> + return 0;
>>>> +}
>>>> diff --git a/gdb/testsuite/gdb.threads/threadcrash.exp b/gdb/testsuite/gdb.threads/threadcrash.exp
>>>> new file mode 100644
>>>> index 00000000000..996e020d1e8
>>>> --- /dev/null
>>>> +++ b/gdb/testsuite/gdb.threads/threadcrash.exp
>>>> @@ -0,0 +1,233 @@
>>>> +# This testcase is part of GDB, the GNU debugger.
>>>> +
>>>> +# Copyright 2023 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/>.
>>>> +
>>>> +# This test case looks at GDB's ability to get correct backtraces for a
>>>> +# crashed inferior, recreating it from a live inferior, a corefile and
>>>> +# a gcore.
>>>> +
>>>> +
>>>> +# Check that the inferior has 7 threads, and return the number of threads (7).
>>>> +# We return the thread count so that, even if there is some error in the test,
>>>> +# the final log doesn't get flooded with failures.
>>>> +
>>>> +proc test_thread_count {} {
>>>> + set thread_count 0
>>>> +
>>>> + gdb_test_multiple "info threads" "getting thread count" -lbl {
>>>> + -re "Thread" {
>>>> + incr thread_count
>>>> + exp_continue
>>>> + }
>>>> + -re "$::gdb_prompt " {
>>>> + gdb_assert {$thread_count == 7}
>>>> + }
>>>> + }
>>>> +
>>>> + return $thread_count
>>>> +}
>>>> +
>>>> +# Use 'thread apply all backtrace' to check if all expected threads
>>>> +# are present, and stopped in the expected locations. Set the global
>>>> +# TEST_LIST to be the a list of regexps expected to match all the
>>>> +# threads. We generate it now so that the list is in the order that
>>>> +# GDB sees the threads.
>>>> +
>>>> +proc thread_apply_all {} {
>>>> + global test_list
>>>> +
>>>> + set test_list { }
>>>> +
>>>> + set unwind_fail false
>>>> +
>>>> + gdb_test_multiple "thread apply all backtrace" \
>>>> + "Get thread information" -lbl {
>>>> + -re "#\[0-9\]+\\\?\\\?\[^\n\]*" {
>>>> + set unwind_fail true
>>>> + exp_continue
>>>> + }
>>>> + -re "\[^\n\]*syscall_task .location=SIGNAL_ALT_STACK\[^\n\]*" {
>>>> + lappend test_list [multi_line ".*sleep.*" \
>>>> + ".*do_syscall_task .location=SIGNAL_ALT_STACK.*" \
>>>> + ".*signal_handler.*" \
>>>> + ".*signal handler called.*" \
>>>> + ".*pthread_kill.*" \
>>>> + ".*thread_function.*"]
>>>> + exp_continue
>>>> + }
>>>> + -re "\[^\n\]*syscall_task .location=SIGNAL_HANDLER\[^\n\]*" {
>>>> + lappend test_list [multi_line ".*sleep.*" \
>>>> + ".*do_syscall_task .location=SIGNAL_HANDLER.*" \
>>>> + ".*signal_handler.*" \
>>>> + ".*signal handler called.*" \
>>>> + ".*pthread_kill.*" \
>>>> + ".*thread_function.*"]
>>>> + exp_continue
>>>> + }
>>>> + -re "\[^\n\]*syscall_task .location=NORMAL\[^\n\]*" {
>>>> + lappend test_list [multi_line ".*sleep.*" \
>>>> + ".*do_syscall_task .location=NORMAL.*" \
>>>> + ".*thread_function.*"]
>>>> + exp_continue
>>>> + }
>>>> + -re "\[^\n\]*spin_task .location=SIGNAL_ALT_STACK\[^\n\]*" {
>>>> + lappend test_list [multi_line ".*do_spin_task .location=SIGNAL_ALT_STACK.*" \
>>>> + ".*signal_handler.*" \
>>>> + ".*signal handler called.*" \
>>>> + ".*pthread_kill.*" \
>>>> + ".*thread_function.*"]
>>>> + exp_continue
>>>> + }
>>>> + -re "\[^\n\]*spin_task .location=SIGNAL_HANDLER\[^\n\]*" {
>>>> + lappend test_list [multi_line ".*do_spin_task .location=SIGNAL_HANDLER.*" \
>>>> + ".*signal_handler.*" \
>>>> + ".*signal handler called.*" \
>>>> + ".*pthread_kill.*" \
>>>> + ".*thread_function.*"]
>>>> + exp_continue
>>>> + }
>>>> + -re "\[^\n\]*spin_task .location=NORMAL\[^\n\]*" {
>>>> + lappend test_list [multi_line ".*do_spin_task .location=NORMAL..*" \
>>>> + ".*thread_function.*"]
>>>> + exp_continue
>>>> + }
>>>> + -re "\[^\n\]*main\[^\n\]*" {
>>>> + lappend test_list ".*main.*"
>>>> + exp_continue
>>>> + }
>>>> + -re "$::gdb_prompt " {
>>>> + pass $gdb_test_name
>>>> + }
>>>> + }
>>>> +
>>>> + gdb_assert {$unwind_fail == false}
>>>> +}
>>>> +
>>>> +# Perform all the tests we're interested in. They are:
>>>> +# * test if we have 7 threads
>>>> +# * Creating the list of backtraces for all threads seen
>>>> +# * testing if GDB recreated the full backtrace we expect for all threads
>>>> +
>>>> +proc do_full_test {} {
>>>> + global test_list
>>>> + set thread_count [test_thread_count]
>>>> +
>>>> + thread_apply_all
>>>> +
>>>> + gdb_assert {$thread_count == [llength $test_list]}
>>>> +
>>>> + for {set i 0} {$i < $thread_count } {incr i} {
>>>> + set thread_num [expr [llength $test_list] - $i]
>>>> +
>>>> + gdb_test "thread apply $thread_num backtrace" [lindex $test_list $i]
>>>> + }
>>>> +}
>>>> +
>>>> +# Do all preparation steps for running the corefile tests, then
>>>> +# call do_full_test to actually run the tests.
>>>> +
>>>> +proc_with_prefix test_live_inferior {} {
>>>> + gdb_test "handle SIGUSR1 nostop print pass" \
>>>> + ".*SIGUSR1.*No.*Yes.*Yes.*User defined signal 1" \
>>>> + "setup SIGUSR1"
>>>> + gdb_test "handle SIGUSR2 nostop print pass" \
>>>> + ".*SIGUSR2.*No.*Yes.*Yes.*User defined signal 2" \
>>>> + "setup SIGUSR2"
>>>> +
>>>> + if {![runto_main]} {
>>>> + return
>>>> + }
>>>> +
>>>> + gdb_breakpoint "breakpt"
>>>> + gdb_continue_to_breakpoint "running to breakpoint" ".*"
>>>> +
>>>> + do_full_test
>>>> +}
>>>> +
>>>> +# Do all preparation steps for running the corefile tests, then
>>>> +# call do_full_test to actually run the tests.
>>>> +
>>>> +proc_with_prefix test_corefile {} {
>>>> + set corefile [core_find $::binfile]
>>>> + if { $corefile == "" } {
>>>> + untested "couldn't generate corefile"
>>>> + return
>>>> + }
>>>> + set corefile [gdb_remote_download host $corefile]
>>>> +
>>>> + gdb_test "core-file $corefile" \
>>>> + "" \
>>>> + "loading_corefile" \
>>>> + "A program is being debugged already\\\. Kill it\\\? \\\(y or n\\\) " \
>>>> + "y"
>>>> +
>>>> + do_full_test
>>>> +}
>>>> +
>>>> +# Do all preparation steps for running the gcore tests, then
>>>> +# call do_full_test to actually run the tests.
>>>> +
>>>> +proc_with_prefix test_gcore {} {
>>>> +
>>>> + clean_restart "$::binfile"
>>>> +
>>>> + gdb_test "handle SIGUSR1 nostop print pass" \
>>>> + ".*SIGUSR1.*No.*Yes.*Yes.*User defined signal 1" \
>>>> + "setup SIGUSR1"
>>>> + gdb_test "handle SIGUSR2 nostop print pass" \
>>>> + ".*SIGUSR2.*No.*Yes.*Yes.*User defined signal 2" \
>>>> + "setup SIGUSR2"
>>>> +
>>>> + if {![runto_main]} {
>>>> + return -1
>>>> + }
>>>> + gdb_test "continue" ".*Segmentation fault.*" "continue to crash"
>>>> +
>>>> + set gcore_name "${::binfile}.gcore"
>>>> + set gcore_supported [gdb_gcore_cmd "$gcore_name" "saving gcore"]
>>>> +
>>>> + if {!$gcore_supported} {
>>>> + unsupported "couldn't generate gcore file"
>>>> + return
>>>> + }
>>>> +
>>>> + set corefile [gdb_remote_download host $gcore_name]
>>>> +
>>>> + gdb_test "core-file $corefile" \
>>>> + "" \
>>>> + "loading_corefile" \
>>>> + "A program is being debugged already\\\. Kill it\\\? \\\(y or n\\\) " \
>>>> + "y"
>>>> +
>>>> + do_full_test
>>>> +}
>>>> +
>>>> +standard_testfile
>>>> +
>>>> +if [prepare_for_testing "failed to prepare" $testfile $srcfile \
>>>> + {debug pthreads}] {
>>>> + return -1
>>>> +}
>>>> +
>>>> +clean_restart ${binfile}
>>>> +
>>>> +gdb_test_no_output "set backtrace limit unlimited"
>>>> +
>>>> +test_live_inferior
>>>> +
>>>> +test_corefile
>>>> +
>>>> +test_gcore
>>>
>>>
>>
>
Thanks for the patch.
I re-tested this on aarch64-linux, for native gdb/gdbserver and extended native gdbserver.
I'm getting full passes on my end. The code looks good to me.
Approved-By: Luis Machado <luis.machado@arm.com>
next prev parent reply other threads:[~2024-01-24 12:35 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-04 17:33 [PATCH " Guinevere Larsen
2023-12-20 9:40 ` Guinevere Larsen
2024-01-02 10:41 ` [PINGv2] " Guinevere Larsen
2024-01-09 11:50 ` [PINGv2][PATCH " Guinevere Larsen
2024-01-17 9:04 ` [PINGv3][PATCH " Guinevere Larsen
2024-01-24 9:34 ` Guinevere Larsen
2024-01-24 12:35 ` Luis Machado [this message]
2024-01-24 12:54 ` Guinevere Larsen
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=33d4e5ab-27dd-4008-962f-abd02436847d@arm.com \
--to=luis.machado@arm.com \
--cc=aburgess@redhat.com \
--cc=blarsen@redhat.com \
--cc=gdb-patches@sourceware.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).