Re: [PATCH] Use strscpy() instead of the deprecated strlcpy()
From: Athira Rajeev
Date: Tue Nov 05 2024 - 07:25:29 EST
> On 5 Nov 2024, at 12:30 PM, Daniel Yang <danielyangkang@xxxxxxxxx> wrote:
>
> The latest kernel doc:
> https://www.kernel.org/doc/html/latest/process/deprecated.html
> recommends replacing strlcpy with strscpy. Since the return value of
> strlcpy is not used, replacing with strscpy should be safe.
>
> Signed-off-by: Daniel Yang <danielyangkang@xxxxxxxxx>
> ---
> tools/perf/util/symbol.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> index 3bbf173ad..660ed5626 100644
> --- a/tools/perf/util/symbol.c
> +++ b/tools/perf/util/symbol.c
> @@ -1893,7 +1893,7 @@ int dso__load(struct dso *dso, struct map *map)
> char *new_name = dso__filename_with_chroot(dso, name);
> if (new_name) {
> is_reg = is_regular_file(new_name);
> - strlcpy(name, new_name, PATH_MAX);
> + strscpy(name, new_name, PATH_MAX);
> free(new_name);
Hi,
This hits a compilation fail:
In file included from util/symbol.c:10:
util/symbol.c: In function ‘dso__load’:
/home/athira/perf-tools-next/tools/include/linux/string.h:15:17: error: too many arguments to function ‘strcpy’
15 | #define strscpy strcpy
| ^~~~~~
util/symbol.c:1896:33: note: in expansion of macro ‘strscpy’
1896 | strscpy(name, new_name, PATH_MAX);
| ^~~~~~~
In file included from /usr/include/features.h:490,
from /usr/include/dirent.h:25,
from util/symbol.c:2:
/usr/include/bits/string_fortified.h:77:1: note: declared here
77 | __NTH (strcpy (char *__restrict __dest, const char *__restrict __src))
| ^~~~~
“strscpy" in perf tools is defined here: tools/include/linux/string.h as “strcpy”
#define strscpy strcpy
And this is introduced from below commit:
commit 9e3d665384fca2a1c56283c7a79a968243ef4614
Author: Wei Yang <richard.weiyang@xxxxxxxxx>
Date: Tue Aug 6 01:03:19 2024 +0000
memblock test: fix implicit declaration of function 'strscpy'
Commit 1e4c64b71c9b ("mm/memblock: Add "reserve_mem" to reserved named
memory at boot up") introduce the usage of strscpy, which breaks the
memblock test.
Let's define it as strcpy in userspace to fix it.
Signed-off-by: Wei Yang <richard.weiyang@xxxxxxxxx>
Link: https://lore.kernel.org/r/20240806010319.29194-5-richard.weiyang@xxxxxxxxx
Signed-off-by: Mike Rapoport (Microsoft) <rppt@xxxxxxxxxx>
diff --git a/tools/include/linux/string.h b/tools/include/linux/string.h
index fb8eda3019b5..41e7fa734922 100644
--- a/tools/include/linux/string.h
+++ b/tools/include/linux/string.h
@@ -12,6 +12,8 @@ void argv_free(char **argv);
int strtobool(const char *s, bool *res);
+#define strscpy strcpy
+
/*
* glibc based builds needs the extern while uClibc doesn't.
* However uClibc headers also define __GLIBC__ hence the hack below
So the strscpy here is defined as strcpy itself.
Ian discussed about having an strscpy shim for perf here :
https://lore.kernel.org/linux-perf-users/172892779710.897882.14949543082561189584.b4-ty@xxxxxxxxxx/T/#meaa2f8d993c6db435f9ed399dc1bf10132a31292
So IIUC, we are yet to have the safe strscpy in the tools perf.
Please correct if I am wrong. Were you able to compile the patch in your setup or am I missing something ?
Thanks
Athira
> }
> }
> --
> 2.39.2
>
>