linux-kernel - Re: [PATCH] tools/nolibc: add support for openat(2)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <D86I2TBQIPLZ.TCUUJWWM5DMO@kragniz.eu>
Date: Mon, 03 Mar 2025 08:49:37 +0000
From: "Louis Taylor" <louis@...gniz.eu>
To: Thomas Weißschuh <linux@...ssschuh.net>
Cc: "Willy Tarreau" <w@....eu>, "Shuah Khan" <shuah@...nel.org>,
 <linux-kernel@...r.kernel.org>, <linux-kselftest@...r.kernel.org>
Subject: Re: [PATCH] tools/nolibc: add support for openat(2)

On Sun Mar 2, 2025 at 10:24 PM GMT, Thomas Weißschuh wrote:
> Hi Louis,
>
> On 2025-03-02 20:25:23+0000, Louis Taylor wrote:
> > openat is useful to avoid needing to construct relative paths, so expose
> > a wrapper for using it directly.
>
> Can you say what you are using nolibc for? I'm curious :-)

An incredibly dumb side project which involves creating a very small
static binary. So far I've also needed a few other functions (for
example clock_gettime) which I've just added local wrappers for, since
I'm not sure if they're in scope for nolibc. If they are, I can also
send patches for those as well.

> > Signed-off-by: Louis Taylor <louis@...gniz.eu>
> > ---
> >  tools/include/nolibc/sys.h                   | 29 ++++++++++++++++++++
> >  tools/testing/selftests/nolibc/nolibc-test.c | 22 +++++++++++++++
> >  2 files changed, 51 insertions(+)
> > 
> > diff --git a/tools/include/nolibc/sys.h b/tools/include/nolibc/sys.h
> > index 8f44c33b1213..e5ff34df4aee 100644
> > --- a/tools/include/nolibc/sys.h
> > +++ b/tools/include/nolibc/sys.h
> > @@ -765,6 +765,35 @@ int mount(const char *src, const char *tgt,
> >  	return __sysret(sys_mount(src, tgt, fst, flags, data));
> >  }
> >  
> > +/*
> > + * int openat(int dirfd, const char *path, int flags[, mode_t mode]);
> > + */
> > +
> > +static __attribute__((unused))
> > +int sys_openat(int dirfd, const char *path, int flags, mode_t mode)
> > +{
> > +#ifdef __NR_openat
> > +	return my_syscall4(__NR_openat, dirfd, path, flags, mode);
> > +#else
> > +	return __nolibc_enosys(__func__, dirfd, path, flags, mode);
> > +#endif
>
> All architectures support openat(), so the #else could be dropped.

I agree, but I followed the existing implementation for sys_open which
only uses openat if it is defined. If openat can be assumed to always
exist, that other #ifdef should be dropped (which I can do in another
patch).

> > +}
> > +
> > +static __attribute__((unused))
> > +int openat(int dirfd, const char *path, int flags, ...)
> > +{
> > +	mode_t mode = 0;
> > +
> > +	if (flags & O_CREAT) {
> > +		va_list args;
> > +
> > +		va_start(args, flags);
> > +		mode = va_arg(args, int);
>
> mode_t instead of int?

This implementation is yoinked directly from open() below. I have no
opinions, but if it should be changed it should be changed in both
functions.

> > +		va_end(args);
> > +	}
> > +
> > +	return __sysret(sys_openat(dirfd, path, flags, mode));
> > +}
> >  
> >  /*
> >   * int open(const char *path, int flags[, mode_t mode]);
> > diff --git a/tools/testing/selftests/nolibc/nolibc-test.c b/tools/testing/selftests/nolibc/nolibc-test.c
> > index 79c3e6a845f3..97ded6c76f99 100644
> > --- a/tools/testing/selftests/nolibc/nolibc-test.c
> > +++ b/tools/testing/selftests/nolibc/nolibc-test.c
> > @@ -1028,6 +1028,26 @@ int test_rlimit(void)
> >  	return 0;
> >  }
> >  
> > +int test_openat(void)
>
> static.

I'll change this. Should the other test_ functions in this file also be
static?

> > +{
> > +	int dev;
> > +	int null;
> > +
> > +	dev = openat(AT_FDCWD, "/dev", O_DIRECTORY);
> > +	if (dev < 0)
> > +		return -1;
> > +
> > +	null = openat(dev, "null", 0);
> > +	if (null < 0) {
> > +		close(dev);
> > +		return -1;
> > +	}
> > +
> > +	close(dev);
> > +	close(null);
> > +
> > +	return 0;
> > +}
> >  
> >  /* Run syscall tests between IDs <min> and <max>.
> >   * Return 0 on success, non-zero on failure.
> > @@ -1116,6 +1136,8 @@ int run_syscall(int min, int max)
> >  		CASE_TEST(mmap_munmap_good);  EXPECT_SYSZR(1, test_mmap_munmap()); break;
> >  		CASE_TEST(open_tty);          EXPECT_SYSNE(1, tmp = open("/dev/null", 0), -1); if (tmp != -1) close(tmp); break;
> >  		CASE_TEST(open_blah);         EXPECT_SYSER(1, tmp = open("/proc/self/blah", 0), -1, ENOENT); if (tmp != -1) close(tmp); break;
> > +		CASE_TEST(openat_fdcwd);      EXPECT_SYSNE(1, tmp = openat(AT_FDCWD, "/dev/null", 0), -1); if (tmp != -1) close(tmp); break;
>
> AT_FDCWD is already used in test_openat(). What additional value does
> the test above add?

None, I just wrote that one first before thinking about how I'd test
opening something relative to an fd. I'll drop it.

>
> > +		CASE_TEST(openat_dir);        EXPECT_SYSNE(1, test_openat(), -1); break;
> >  		CASE_TEST(pipe);              EXPECT_SYSZR(1, test_pipe()); break;
> >  		CASE_TEST(poll_null);         EXPECT_SYSZR(1, poll(NULL, 0, 0)); break;
> >  		CASE_TEST(poll_stdout);       EXPECT_SYSNE(1, ({ struct pollfd fds = { 1, POLLOUT, 0}; poll(&fds, 1, 0); }), -1); break;
> > -- 
> > 2.45.2
> >