From: Steven Rostedt <rostedt@goodmis.org>
To: Metin Kaya <metin.kaya@arm.com>
Cc: linux-trace-devel@vger.kernel.org
Subject: Re: [PATCH] trace-cmd utest: Add test cases for trace-cmd reset
Date: Tue, 15 Oct 2024 12:38:31 -0400 [thread overview]
Message-ID: <20241015123831.347ff0f4@gandalf.local.home> (raw)
In-Reply-To: <20241015140840.4183007-1-metin.kaya@arm.com>
On Tue, 15 Oct 2024 15:08:40 +0100
Metin Kaya <metin.kaya@arm.com> wrote:
> Implement test cases to confirm existing dynamic event types (kprobe,
> kretprobe, uprobe, uretprobe and eprobe) are really preserved during
> "trace-cmd reset".
>
> Verify synthetic events by inserting a hook to
> test_trace_sqlhist_hist().
>
> While we are in the neighborhood, fix typo "ceated".
>
> TODO: I'm unable to prevent uprobes/uretprobes from being destroyed
> during reset.
That's not a bug with this code. It's a bug with libtracefs. I'll fix it.
>
> Link: https://lore.kernel.org/linux-trace-devel/20241014175905.03bec85a@gandalf.local.home
>
> Signed-off-by: Metin Kaya <metin.kaya@arm.com>
> ---
> utest/tracecmd-utest.c | 161 ++++++++++++++++++++++++++++++++++++++++-
> 1 file changed, 160 insertions(+), 1 deletion(-)
>
> diff --git a/utest/tracecmd-utest.c b/utest/tracecmd-utest.c
> index 164f5da1..b2f6b65b 100644
> --- a/utest/tracecmd-utest.c
> +++ b/utest/tracecmd-utest.c
> @@ -352,6 +352,11 @@ static void test_trace_sqlhist_hist(void)
> sleep(1);
> ret = grep_match(SYNTH_EVENT ":", "show", NULL);
> CU_TEST(ret == 0);
> + /* Ensure synthetic events remain untouched after "trace-cmd reset -k synth". */
> + ret = run_trace("reset", "-k", "synth", NULL);
> + CU_TEST(ret == 0);
> + ret = grep_match(SYNTH_EVENT, "stat", NULL);
> + CU_TEST(ret == 0);
>
> tracefs_instance_reset(NULL);
> }
> @@ -598,6 +603,159 @@ static void test_trace_library_read_back(void)
> tracecmd_close(handle);
> }
>
> +static void test_trace_reset_kprobe(void)
> +{
> + int ret;
> + FILE *fp;
> +
> + /* Create a simple kprobe for do_sys_open */
> + fp = fopen("/sys/kernel/tracing/kprobe_events", "w");
> + CU_TEST(fp != NULL);
> + fprintf(fp, "p do_sys_open");
> + fclose(fp);
This is open coded of:
ret = tracefs_instance_file_write(NULL, "kprobe_events", "p do_sys_open");
CU_TEST(ret > 0);
And there's no guarantee that tracefs will be mounted at
/sys/kernel/tracing, so the above is incorrect anyway.
> +
> + /* Ensure the kprobe is listed in "trace-cmd stat" output. */
> + ret = grep_match("p:kprobes/p_do_sys_open_0 do_sys_open", "stat", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Issue "trace-cmd reset", but keep kprobes. */
> + ret = run_trace("reset", "-k", "kprobe", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Verify the kprobe's existence after reset. */
> + ret = grep_match("p:kprobes/p_do_sys_open_0 do_sys_open", "stat", NULL);
> + CU_TEST(ret == 0);
> +}
> +
> +static void test_trace_reset_kretprobe(void)
> +{
> + int ret;
> + FILE *fp;
> +
> + /* Create a simple kretprobe for do_sys_open */
> + fp = fopen("/sys/kernel/tracing/kprobe_events", "w");
> + CU_TEST(fp != NULL);
> + fprintf(fp, "r do_sys_open");
> + fclose(fp);
Use tracefs_instance_file_write() here too.
> +
> + /* Ensure the kretprobe is listed in "trace-cmd stat" output. */
> + ret = grep_match("r128:kprobes/r_do_sys_open_0 do_sys_open", "stat", NULL);
Note, this failed for me, because mine had "r16:kprobes..."
> + CU_TEST(ret == 0);
> +
> + /* Issue "trace-cmd reset", but keep kretprobes. */
> + ret = run_trace("reset", "-k", "kretprobe", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Verify the kretprobe's existence after reset. */
> + ret = grep_match("r128:kprobes/r_do_sys_open_0 do_sys_open", "stat", NULL);
> + CU_TEST(ret == 0);
> +}
> +
> +static void test_trace_reset_uprobe(void)
> +{
> + int ret;
> + FILE *fp;
> +
> + /* Create a simple uprobe for do_sys_open */
> + fp = fopen("/sys/kernel/tracing/uprobe_events", "w");
> + CU_TEST(fp != NULL);
> + fprintf(fp, "p /bin/bash:0x4245c0");
> + fclose(fp);
Ditto.
> +
> + /* Ensure the uprobe is listed in "trace-cmd stat" output. */
> + ret = grep_match("p:uprobes/p_bash_0x4245c0 /bin/bash:0x00000000004245c0", "stat", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Issue "trace-cmd reset", but keep uprobes. */
> + ret = run_trace("reset", "-k", "uprobe", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Verify the uprobe's existence after reset. */
> + ret = grep_match("p:uprobes/p_bash_0x4245c0 /bin/bash:0x00000000004245c0", "stat", NULL);
> + CU_TEST(ret == 0);
> +}
> +
> +static void test_trace_reset_uretprobe(void)
> +{
> + int ret;
> + FILE *fp;
> +
> + /* Create a simple uretprobe for do_sys_open */
> + fp = fopen("/sys/kernel/tracing/uprobe_events", "w");
> + CU_TEST(fp != NULL);
> + fprintf(fp, "r /bin/bash:0x4245c0");
> + fclose(fp);
Ditto.
> +
> + /* Ensure the uretprobe is listed in "trace-cmd stat" output. */
> + ret = grep_match("r:uprobes/p_bash_0x4245c0 /bin/bash:0x00000000004245c0", "stat", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Issue "trace-cmd reset", but keep uretprobes. */
> + ret = run_trace("reset", "-k", "uretprobe", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Verify the uretprobe's existence after reset. */
> + ret = grep_match("r:uprobes/p_bash_0x4245c0 /bin/bash:0x00000000004245c0", "stat", NULL);
> + CU_TEST(ret == 0);
> +}
> +
> +static void test_trace_reset_eprobe(void)
> +{
> + int ret;
> + bool matched = false;
> + size_t l = 0;
> + ssize_t n;
> + char *buf = NULL;
> + struct tracefs_dynevent *deprobe;
> + FILE *fp;
> +
> + deprobe = tracefs_eprobe_alloc(NULL, "sopen_in", "syscalls", "sys_enter_openat", NULL);
> + CU_TEST(deprobe != NULL);
> +
> + ret = tracefs_dynevent_create(deprobe);
> + CU_TEST(ret == 0);
> +
> + /* Issue "trace-cmd reset", but keep eprobes. */
> + ret = run_trace("reset", "-k", "eprobe", NULL);
> + CU_TEST(ret == 0);
> +
> + /* Verify the eprobe's existence after reset. */
> + fp = fopen("/sys/kernel/tracing/dynamic_events", "r");
> + CU_TEST(fp != NULL);
int fd;
fd = tracefs_instance_file_open(NULL, "dynamic_events", O_RDONLY);
fp = fdopen(fd, "r");
> +
> + while ((n = getline(&buf, &l, fp)) != -1) {
> + if (!strcmp(buf, "e:eprobes/sopen_in syscalls.sys_enter_openat\n")) {
> + matched = true;
> + break;
> + }
> + }
> + free(buf);
-- Steve
> +
> + fclose(fp);
> +
> + CU_TEST(matched == true);
> +
> + ret = tracefs_dynevent_destroy(deprobe, false);
> + CU_TEST(ret == 0);
> +
> + tracefs_dynevent_free(deprobe);
> +}
> +
> +static void test_trace_reset(void)
> +{
> + int ret;
> +
> + test_trace_reset_kprobe();
> + test_trace_reset_kretprobe();
> + test_trace_reset_uprobe();
> + test_trace_reset_uretprobe();
> + test_trace_reset_eprobe();
> +
> + /* Destroy all dynamic events. */
> + ret = run_trace("reset", NULL);
> + CU_TEST(ret == 0);
> +}
> +
> static int test_suite_destroy(void)
> {
> unlink(TRACECMD_FILE);
> @@ -639,7 +797,7 @@ void test_tracecmd_lib(void)
>
> suite = CU_add_suite(TRACECMD_SUITE, test_suite_init, test_suite_destroy);
> if (suite == NULL) {
> - fprintf(stderr, "Suite \"%s\" cannot be ceated\n", TRACECMD_SUITE);
> + fprintf(stderr, "Suite \"%s\" cannot be created\n", TRACECMD_SUITE);
> return;
> }
> CU_add_test(suite, "Simple record and report",
> @@ -656,4 +814,5 @@ void test_tracecmd_lib(void)
> test_trace_library_read_back);
> CU_add_test(suite, "Test max length",
> test_trace_record_max);
> + CU_add_test(suite, "Simple reset", test_trace_reset);
> }
next prev parent reply other threads:[~2024-10-15 16:38 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-15 14:08 [PATCH] trace-cmd utest: Add test cases for trace-cmd reset Metin Kaya
2024-10-15 16:38 ` Steven Rostedt [this message]
2024-10-15 17:38 ` Steven Rostedt
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=20241015123831.347ff0f4@gandalf.local.home \
--to=rostedt@goodmis.org \
--cc=linux-trace-devel@vger.kernel.org \
--cc=metin.kaya@arm.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 NNTP newsgroup(s).