Re: [PATCH] bpftool: Add missing libgen.h for basename()

[Jiri Olsa <olsajiri@gmail.com>]

On Fri, Jan 05, 2024 at 05:53:18PM -0300, Arnaldo Carvalho de Melo wrote:
> Em Fri, Jan 05, 2024 at 04:01:37PM +0100, Jiri Olsa escreveu:
> > On Fri, Jan 05, 2024 at 09:48:31AM +0100, Jiri Olsa wrote:
> > > On Thu, Jan 04, 2024 at 10:01:35AM -0300, Arnaldo Carvalho de Melo wrote:
> > > 
> > > SNIP
> > > 
> > > >    9    51.66 amazonlinux:2                 : Ok   gcc (GCC) 7.3.1 20180712 (Red Hat 7.3.1-17) , clang version 11.1.0 (Amazon Linux 2 11.1.0-1.amzn2.0.2) flex 2.5.37
> > > >   10    60.77 amazonlinux:2023              : Ok   gcc (GCC) 11.4.1 20230605 (Red Hat 11.4.1-2) , clang version 15.0.7 (Amazon Linux 15.0.7-3.amzn2023.0.1) flex 2.6.4
> > > >   11    61.29 amazonlinux:devel             : Ok   gcc (GCC) 11.3.1 20221121 (Red Hat 11.3.1-4) , clang version 15.0.6 (Amazon Linux 15.0.6-3.amzn2023.0.2) flex 2.6.4
> > > >   12    74.72 archlinux:base                : Ok   gcc (GCC) 13.2.1 20230801 , clang version 16.0.6 flex 2.6.4
> > > > 
> > > > / $ grep -B8 -A2 -w basename /usr/include/string.h
> > > > #ifdef _GNU_SOURCE
> > > > #define	strdupa(x)	strcpy(alloca(strlen(x)+1),x)
> > > > int strverscmp (const char *, const char *);
> > > > char *strchrnul(const char *, int);
> > > > char *strcasestr(const char *, const char *);
> > > > void *memrchr(const void *, int, size_t);
> > > > void *mempcpy(void *, const void *, size_t);
> > > > #ifndef __cplusplus
> > > > char *basename();
> > > > #endif
> > > > #endif
> > > > / $ cat /etc/os-release
> > > > NAME="Alpine Linux"
> > > > ID=alpine
> > > > VERSION_ID=3.19.0
> > > > PRETTY_NAME="Alpine Linux v3.19"
> > > > HOME_URL="https://alpinelinux.org/"
> > > > BUG_REPORT_URL="https://gitlab.alpinelinux.org/alpine/aports/-/issues"
> > > > / $
> > > > 
> > > > Weird, they had it and now removed the _GNU_SOURCE bits (edge is their
> > > > devel distro, like rawhide is for fedora, tumbleweed for opensuse, etc).
> > > 
> > > let's see, I asked them in here: https://gitlab.alpinelinux.org/alpine/aports/-/issues/15643
> > 
> > it got removed in musl libc recently:
> >   https://git.musl-libc.org/cgit/musl/commit/?id=725e17ed6dff4d0cd22487bb64470881e86a92e7
> > 
> > so perhaps switching to POSIX version of basename is the easiest way out?
> 
> I think so, in all of perf we use the POSIX one, strdup'ing the arg,
> etc.
> 
> Something like the patch below?
> 
> - Arnaldo
> 
> diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c
> index ee3ce2b8000d75d2..a5cc5938c3d7951e 100644
> --- a/tools/bpf/bpftool/gen.c
> +++ b/tools/bpf/bpftool/gen.c
> @@ -7,6 +7,7 @@
>  #include <ctype.h>
>  #include <errno.h>
>  #include <fcntl.h>
> +#include <libgen.h>
>  #include <linux/err.h>
>  #include <stdbool.h>
>  #include <stdio.h>
> @@ -56,9 +57,10 @@ static bool str_has_suffix(const char *str, const char *suffix)
>  
>  static void get_obj_name(char *name, const char *file)
>  {
> -	/* Using basename() GNU version which doesn't modify arg. */
> -	strncpy(name, basename(file), MAX_OBJ_NAME_LEN - 1);
> -	name[MAX_OBJ_NAME_LEN - 1] = '\0';
> +	char file_copy[PATH_MAX];

ok, probably better then checking for strdup error

> +	/* Using basename() POSIX version to be more portable. */
> +	strncpy(file_copy, file, PATH_MAX - 1)[PATH_MAX - 1] = '\0';
> +	strncpy(name, basename(file_copy), MAX_OBJ_NAME_LEN - 1)[MAX_OBJ_NAME_LEN - 1] = '\0';

I've never used it like that.. had to zoom in twice ;-)
but extra line with that might be more readable

jirka

>  	if (str_has_suffix(name, ".o"))
>  		name[strlen(name) - 2] = '\0';
>  	sanitize_identifier(name);