From: Namhyung Kim <namhyung@kernel.org>
To: Athira Rajeev <atrajeev@linux.ibm.com>
Cc: acme@kernel.org, jolsa@kernel.org, adrian.hunter@intel.com,
vmolnaro@redhat.com, mpetlan@redhat.com, tmricht@linux.ibm.com,
maddy@linux.ibm.com, irogers@google.com,
linux-perf-users@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
hbathini@linux.vnet.ibm.com, Tejas.Manhas1@ibm.com,
Tanushree.Shah@ibm.com, Shivani.Nittor@ibm.com
Subject: Re: [PATCH V2] tools/perf/tests: Update test_adding_kernel.sh to handle proper debuginfo check
Date: Mon, 30 Mar 2026 23:56:40 -0700 [thread overview]
Message-ID: <actwKNAzk6PnoBDX@google.com> (raw)
In-Reply-To: <3E20DA19-2C2A-45EC-A2AA-7C05CE89C4FE@linux.ibm.com>
On Fri, Mar 27, 2026 at 11:26:48PM +0530, Athira Rajeev wrote:
>
>
> > On 27 Mar 2026, at 4:43 AM, Namhyung Kim <namhyung@kernel.org> wrote:
> >
> > On Mon, Mar 23, 2026 at 05:54:24PM +0530, Athira Rajeev wrote:
> >> Perf test perftool-testsuite_probe fails as below:
> >>
> >> Regexp not found: "\s*probe:inode_permission(?:_\d+)?\s+\(on inode_permission(?:[:\+][0-9A-Fa-f]+)?@.+\)"
> >> -- [ FAIL ] -- perf_probe :: test_adding_kernel :: listing added probe :: perf probe -l (output regexp parsing)
> >> -- [ PASS ] -- perf_probe :: test_adding_kernel :: removing multiple probes
> >> Regexp not found: "probe:vfs_mknod"
> >> Regexp not found: "probe:vfs_create"
> >> Regexp not found: "probe:vfs_rmdir"
> >> Regexp not found: "probe:vfs_link"
> >> Regexp not found: "probe:vfs_write"
> >> -- [ FAIL ] -- perf_probe :: test_adding_kernel :: wildcard adding support (command exitcode + output regexp parsing)
> >> Regexp not found: "somenonexistingrandomstuffwhichisalsoprettylongorevenlongertoexceed64"
> >> Regexp not found: "in this function|at this address"
> >> -- [ FAIL ] -- perf_probe :: test_adding_kernel :: non-existing variable (output regexp parsing)
> >> ## [ FAIL ] ## perf_probe :: test_adding_kernel SUMMARY :: 3 failures found
> >>
> >> Further analysing, the failed testcase is for "test_adding_kernel".
> >> If the kernel debuginfo is missing, perf probe fails as below:
> >>
> >> perf probe -nf --max-probes=512 -a 'vfs_* $params'
> >> Failed to find the path for the kernel: No such file or directory
> >> Error: Failed to add events.
> >>
> >> skip_if_no_debuginfo has check to handle whether debuginfo is present
> >> and the testcase checks for debuginfo since this :
> >> commit 90d32e92011e ("tools/perf: Handle perftool-testsuite_probe
> >> testcases fail when kernel debuginfo is not present")
> >>
> >> Recently a change got added in "tests/shell/lib/probe_vfs_getname.sh"
> >> via this another fix:
> >> commit 92b664dcefab ("perf test probe_vfs_getname: Skip if no suitable
> >> line detected")
> >> Since this commit, first add_probe_vfs_getname is used to prevent false
> >> failures. And based on return code of add_probe_vfs_getname, skip_if_no_debuginfo
> >> is used to skip testcase if debuginfo is present. And this modified other
> >> testcases to call add_probe_vfs_getname first and invoke
> >> skip_if_no_debuginfo based on return value.
> >>
> >> The tests in test_adding_kernel.sh which depends on presence of
> >> debuginfo are:
> >> 1. probe add for inode_permission
> >> 2. probe max-probes option using 'vfs_* $params'
> >> 3. non-existing variable probing
> >>
> >> For these tests, probe check for specific line is not required.
> >> So call skip_if_no_debuginfo with argument to say if line check is
> >> needed. This is to convey to skip_if_no_debuginfo() function
> >> that test only needs to check for debuginfo, and not specifically
> >> line number. Update skip_if_no_debuginfo to use simple "perf probe"
> >> check if test only needs to check for debuginfo. And for other
> >> tests which rely on line number, use add_probe_vfs_getname()
> >>
> >> With the change, verified that only three which required debuginfo only
> >> is skipped and others ran successfully. Also tested with debuginfo
> >> to make sure tests are not skipped.
> >>
> >> Reported-by: Tejas Manhas <Tejas.Manhas1@ibm.com>
> >> Reviewed-by: Ian Rogers <irogers@google.com>
> >> Signed-off-by: Athira Rajeev <atrajeev@linux.ibm.com>
> >> ---
> >> Changelog:
> >> - First version used "perf probe -v -L getname_flags" for debuginfo
> >> check. This will not catch fail string "Debuginfo-analysis is not
> >> supported" which is used in cases when perf is built without dwarf.
> >> So use "perf probe -vn add inode_permission" to capture cases when
> >> tools built with NO_LIBDWARF=1. This will capture debuginfo missing as
> >> well as tool built without dwarf case.
> >>
> >> .../tests/shell/base_probe/test_adding_kernel.sh | 15 ++++++++++++++-
> >> tools/perf/tests/shell/lib/probe_vfs_getname.sh | 13 ++++++++++++-
> >> 2 files changed, 26 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/tools/perf/tests/shell/base_probe/test_adding_kernel.sh b/tools/perf/tests/shell/base_probe/test_adding_kernel.sh
> >> index 555a825d55f2..f3db125c8669 100755
> >> --- a/tools/perf/tests/shell/base_probe/test_adding_kernel.sh
> >> +++ b/tools/perf/tests/shell/base_probe/test_adding_kernel.sh
> >> @@ -23,10 +23,23 @@ TEST_RESULT=0
> >> . "$DIR_PATH/../lib/probe_vfs_getname.sh"
> >>
> >> TEST_PROBE=${TEST_PROBE:-"inode_permission"}
> >> +PROBE_NO_LINE_CHECK=1
> >
> > I'm not sure I follow but shouldn't it depend on the return value of
> > `perf check feature -q dwarf`?
> >
> > Thanks,
> > Namhyung
>
> Hi Namhyung
>
> Thanks for checking the patch.
>
> Yes, "perf check feature -q dwarf” checks if perf is compiled with dwarf support.
> But for “test_adding_kernel” , we need to check if kernel debuginfo is missing.
> The check this path is trying to, also includes “debuginfo”
>
> In environment without debuginfo: perftool-testsuite_probe fails as below:
>
> Regexp not found: "\s*probe:inode_permission(?:_\d+)?\s+\(on inode_permission(?:[:\+][0-9A-Fa-f]+)?@.+\)"
> -- [ FAIL ] -- perf_probe :: test_adding_kernel :: listing added probe :: perf probe -l (output regexp parsing)
> -- [ PASS ] -- perf_probe :: test_adding_kernel :: removing multiple probes
> Regexp not found: "probe:vfs_mknod"
> Regexp not found: "probe:vfs_create"
> Regexp not found: "probe:vfs_rmdir"
> Regexp not found: "probe:vfs_link"
> Regexp not found: "probe:vfs_write"
> -- [ FAIL ] -- perf_probe :: test_adding_kernel :: wildcard adding support (command exitcode + output regexp parsing)
> Regexp not found: "somenonexistingrandomstuffwhichisalsoprettylongorevenlongertoexceed64"
> Regexp not found: "in this function|at this address"
> -- [ FAIL ] -- perf_probe :: test_adding_kernel :: non-existing variable (output regexp parsing)
> ## [ FAIL ] ## perf_probe :: test_adding_kernel SUMMARY :: 3 failures found
>
> We have “skip_if_no_debuginfo” check in tools/perf/tests/shell/lib/probe_vfs_getname.sh and
> “test_adding_kernel” calls this to find out presence of debuginfo.
>
> Recently a change got added via commit 92b664dcefab ("perf test probe_vfs_getname: Skip if no suitable
> line detected”) where we return if specific line is not found using perf probe. And hence debuginfo check is not
> happening.
>
>
> The tests in test_adding_kernel.sh which depends on presence of
> debuginfo are:
> 1. probe add for inode_permission
> 2. probe max-probes option using 'vfs_* $params'
> 3. non-existing variable probing
>
> For these tests, probe check for specific line is not required. So this patch updates
> skip_if_no_debuginfo to use an argument that communicates whether line check is needed.
> If only debuginfo/dwarf check is required, use simple "perf probe”. And for other
> tests which rely on line number, fallback to default add_probe_vfs_getname().
Ok, thanks for the explanation. But I'm afraid you didn't change other
callsites to pass 0 explicitly.
> >
> >>
> >> # set NO_DEBUGINFO to skip testcase if debuginfo is not present
> >> # skip_if_no_debuginfo returns 2 if debuginfo is not present
> >> -skip_if_no_debuginfo
> >> +#
> >> +# The perf probe checks which depends on presence of debuginfo and
> >> +# used in this testcase are:
> >> +# 1. probe add for inode_permission
> >> +# 2. probe max-probes option using 'vfs_* $params'
> >> +# 3. non-existing variable probing
> >> +#
> >> +# For these tests, probe check for specific line is not
> >> +# required ( add_probe_vfs_getname does that ). So call
> >> +# skip_if_no_debuginfo with argument as 1. This is to convey
> >> +# that test only needs to check for debuginfo, and not specifically
> >> +# line number
> >> +skip_if_no_debuginfo $PROBE_NO_LINE_CHECK
> >> if [ $? -eq 2 ]; then
> >> NO_DEBUGINFO=1
> >> fi
> >> diff --git a/tools/perf/tests/shell/lib/probe_vfs_getname.sh b/tools/perf/tests/shell/lib/probe_vfs_getname.sh
> >> index 88cd0e26d5f6..8ae2ea2bc8a2 100644
> >> --- a/tools/perf/tests/shell/lib/probe_vfs_getname.sh
> >> +++ b/tools/perf/tests/shell/lib/probe_vfs_getname.sh
> >> @@ -39,7 +39,18 @@ add_probe_vfs_getname() {
> >> }
> >>
> >> skip_if_no_debuginfo() {
> >> - add_probe_vfs_getname -v 2>&1 | grep -E -q "^(Failed to find the path for the kernel|Debuginfo-analysis is not supported)|(file has no debug information)" && return 2
> >> + no_line_check=$1
> >> + debug_str="^(Failed to find the path for the kernel|Debuginfo-analysis is not supported)|(file has no debug information)"
> >> +
> >> + # search for debug_str using simple perf probe if the
> >> + # test only needs to check for debuginfo, and not specifically
> >> + # line number.
> >> + if [ $no_line_check -eq 1 ]; then
> >> + perf probe -vn add inode_permission 2>&1 | grep -E -q "$debug_str" && return 2
Sashiko found that it should be '--add' or '-a'.
Thanks,
Namhyung
> >> + else
> >> + add_probe_vfs_getname -v 2>&1 | grep -E -q "$debug_str" && return 2
> >> + fi
> >> +
> >> return 1
> >> }
> >>
> >> --
> >> 2.47.3
>
>
next prev parent reply other threads:[~2026-03-31 6:56 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-23 12:24 [PATCH V2] tools/perf/tests: Update test_adding_kernel.sh to handle proper debuginfo check Athira Rajeev
2026-03-26 23:13 ` Namhyung Kim
2026-03-27 17:56 ` Athira Rajeev
2026-03-31 6:56 ` Namhyung Kim [this message]
2026-03-31 10:12 ` Athira Rajeev
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=actwKNAzk6PnoBDX@google.com \
--to=namhyung@kernel.org \
--cc=Shivani.Nittor@ibm.com \
--cc=Tanushree.Shah@ibm.com \
--cc=Tejas.Manhas1@ibm.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=atrajeev@linux.ibm.com \
--cc=hbathini@linux.vnet.ibm.com \
--cc=irogers@google.com \
--cc=jolsa@kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=maddy@linux.ibm.com \
--cc=mpetlan@redhat.com \
--cc=tmricht@linux.ibm.com \
--cc=vmolnaro@redhat.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