Re: [PATCH] tools/perf/tests: Fix compilation error with strncpy in tests/tool_pmu

On Wed, Oct 16, 2024 at 11:23 AM Athira Rajeev
[off-list ref] wrote:

quoted

On 16 Oct 2024, at 8:36 PM, Ian Rogers [off-list ref] wrote:

On Wed, Oct 16, 2024 at 5:30 AM Athira Rajeev
[off-list ref] wrote:

quoted

quoted

On 14 Oct 2024, at 10:56 PM, Namhyung Kim [off-list ref] wrote:

Hello Athira,

On Sun, Oct 13, 2024 at 11:07:42PM +0530, Athira Rajeev wrote:

quoted

perf fails to compile on systems with GCC version11
as below:

In file included from /usr/include/string.h:519,
               from /home/athir/perf-tools-next/tools/include/linux/bitmap.h:5,
               from /home/athir/perf-tools-next/tools/perf/util/pmu.h:5,
               from /home/athir/perf-tools-next/tools/perf/util/evsel.h:14,
               from /home/athir/perf-tools-next/tools/perf/util/evlist.h:14,
               from tests/tool_pmu.c:3:
In function ‘strncpy’,
  inlined from ‘do_test’ at tests/tool_pmu.c:25:3:
/usr/include/bits/string_fortified.h:95:10: error: ‘__builtin_strncpy’ specified bound 128 equals destination size [-Werror=stringop-truncation]
 95 |   return __builtin___strncpy_chk (__dest, __src, __len,
    |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 96 |                                   __glibc_objsize (__dest));
    |                                   ~~~~~~~~~~~~~~~~~~~~~~~~~

The compile error is from strncpy refernce in do_test:
strncpy(str, tool_pmu__event_to_str(ev), sizeof(str));

This behaviour is not observed with GCC version 8, but observed
with GCC version 11 . This is message from gcc for detecting
truncation while using strncpu. Use snprintf instead of strncpy
here to be safe.

Signed-off-by: Athira Rajeev <redacted>

I found this issue now and thanks for the quick fix.  I will push to
perf-tools-next soon.

Thanks,
Namhyung

Sure

Thanks Namhyung for pulling in the change

Athira

Thanks for the fixes. As this is test code I don't think performance,
style, etc. matter much. The GCC strncpy warnings are annoying imo,
I'm not sure it makes sense for them to be enabled. I see in the
kernel lots of "sizeof(foo)-1" as a workaround. strlcpy looks like a
better alternative but it gets a checkpatch warning as in the kernel
strscpy is preferred. Perhaps we should create a strscpy shim.

IIUC, strscpy is defined in kernel source and is currently not used in tools side.

Ian,
Your point here is that we can create an interface in tools side so that we can later safely use strscpy in cases like this ?

Yeah. Tbh the string code is a bit chaotic in the code base. There's
some stuff using char*s, some asprintf, there's strbuf then there all
the kernel ways of doing things. I lose track of why strscpy is
superior to strlcpy, but strlcpy will get you a checkpath warning.

The standard kernel solution to the GCC strncpy warning I've seen,
changing "strncpy(a, b, sizeof(a))" to "strncpy(a, b, sizeof(a) - 1)",
looks dodgy and possibly worse to me. The point of the warning I think
relates to strncpy not necessarily \0 terminating the string and you
need to reserve at least 1 character for this, while strlcpy
guarantees the \0 termination. Doing "sizeof(..) - 1" likely confuses
the checker making it give up, but now you have less space to copy a
potential \0 as well as not getting a \0.

Thanks,
Ian

Thanks
Athira

quoted

Thanks,
Ian

quoted

quoted

quoted

---
tools/perf/tests/tool_pmu.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/tests/tool_pmu.c b/tools/perf/tests/tool_pmu.c
index 94d0dd8fd3cb..297cc8c55579 100644
--- a/tools/perf/tests/tool_pmu.c
+++ b/tools/perf/tests/tool_pmu.c

@@ -22,7 +22,7 @@ static int do_test(enum tool_pmu_event ev, bool with_pmu)

if (with_pmu)
snprintf(str, sizeof(str), "tool/%s/", tool_pmu__event_to_str(ev));
else
- strncpy(str, tool_pmu__event_to_str(ev), sizeof(str));
+ snprintf(str, sizeof(str), "%s", tool_pmu__event_to_str(ev));

parse_events_error__init(&err);
ret = parse_events(evlist, str, &err);
--
2.43.5