[v6] Allow for unpriviledged nested containers

[changes since v5: whether or not to isolate the pid namespace is now a
test-specific "pidns" option, support_needs_proc() takes a "why" message
for future readers]

If the build itself is run in a container, we may not be able to
fully set up a nested container for test-container testing.
Notably is the mounting of /proc, since it's critical that it
be mounted from within the same PID namespace as its users, and
thus cannot be bind mounted from outside the container like other
mounts.

This patch defaults to using the parent's PID namespace instead of
creating a new one, as this is more likely to be allowed.

If the test needs an isolated PID namespace, it should add the "pidns"
command to its init script.

On 3/28/22 23:53, DJ Delorie wrote:
> 
> [changes since v5: whether or not to isolate the pid namespace is now a
> test-specific "pidns" option, support_needs_proc() takes a "why" message
> for future readers]
> 
> If the build itself is run in a container, we may not be able to
> fully set up a nested container for test-container testing.
> Notably is the mounting of /proc, since it's critical that it
> be mounted from within the same PID namespace as its users, and
> thus cannot be bind mounted from outside the container like other
> mounts.
> 
> This patch defaults to using the parent's PID namespace instead of
> creating a new one, as this is more likely to be allowed.
> 
> If the test needs an isolated PID namespace, it should add the "pidns"
> command to its init script.

LGTM.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
 
> diff --git a/elf/tst-pldd.c b/elf/tst-pldd.c
> index f31f9956faa..378262504a5 100644
> --- a/elf/tst-pldd.c
> +++ b/elf/tst-pldd.c
> @@ -85,6 +85,8 @@ in_str_list (const char *libname, const char *const strlist[])
>  static int
>  do_test (void)
>  {
> +  support_need_proc ("needs /proc/sys/kernel/yama/ptrace_scope and /proc/$child");
> +

OK. Nice use of a message to print.

>    /* Check if our subprocess can be debugged with ptrace.  */
>    {
>      int ptrace_scope = support_ptrace_scope ();
> diff --git a/nptl/tst-pthread-getattr.c b/nptl/tst-pthread-getattr.c
> index d2ebf308ae7..3f6f76ea83b 100644
> --- a/nptl/tst-pthread-getattr.c
> +++ b/nptl/tst-pthread-getattr.c
> @@ -28,6 +28,8 @@
>  #include <unistd.h>
>  #include <inttypes.h>
>  
> +#include <support/support.h>
> +
>  /* There is an obscure bug in the kernel due to which RLIMIT_STACK is sometimes
>     returned as unlimited when it is not, which may cause this test to fail.
>     There is also the other case where RLIMIT_STACK is intentionally set as
> @@ -153,6 +155,8 @@ check_stack_top (void)
>  static int
>  do_test (void)
>  {
> +  support_need_proc ("Reads /proc/self/maps to get stack size.");

OK.

> +
>    pagesize = sysconf (_SC_PAGESIZE);
>    return check_stack_top ();
>  }
> diff --git a/nss/tst-reload2.c b/nss/tst-reload2.c
> index fb3b94a1fab..7df0ca740b6 100644
> --- a/nss/tst-reload2.c
> +++ b/nss/tst-reload2.c
> @@ -95,6 +95,8 @@ do_test (void)
>    char buf1[PATH_MAX];
>    char buf2[PATH_MAX];
>  
> +  support_need_proc ("Our xmkdirp fails if we can't map our uid, which requires /proc.");
> +

OK.

>    sprintf (buf1, "/subdir%s", support_slibdir_prefix);
>    xmkdirp (buf1, 0777);
>  
> diff --git a/support/Makefile b/support/Makefile
> index 5ddcb8d1581..f036a813048 100644
> --- a/support/Makefile
> +++ b/support/Makefile
> @@ -64,6 +64,7 @@ libsupport-routines = \
>    support_format_netent \
>    support_isolate_in_subprocess \
>    support_mutex_pi_monotonic \
> +  support_need_proc \

OK.

>    support_path_support_time64 \
>    support_process_state \
>    support_ptrace \
> diff --git a/support/support.h b/support/support.h
> index 73b9fc48f01..d20051da4d4 100644
> --- a/support/support.h
> +++ b/support/support.h
> @@ -91,6 +91,11 @@ char *support_quote_string (const char *);
>     regular file open for writing, and initially empty.  */
>  int support_descriptor_supports_holes (int fd);
>  
> +/* Predicates that a test requires a working /proc filesystem.  This
> +   call will exit with UNSUPPORTED if /proc is not available, printing
> +   WHY_MSG as part of the diagnostic.  */
> +void support_need_proc (const char *why_msg);

OK.

> +
>  /* Error-checking wrapper functions which terminate the process on
>     error.  */
>  
> diff --git a/support/support_need_proc.c b/support/support_need_proc.c
> new file mode 100644
> index 00000000000..9b4eab7539b
> --- /dev/null
> +++ b/support/support_need_proc.c
> @@ -0,0 +1,35 @@
> +/* Indicate that a test requires a working /proc.
> +   Copyright (C) 2022 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library; if not, see
> +   <https://www.gnu.org/licenses/>.  */
> +
> +#include <unistd.h>
> +#include <support/check.h>
> +#include <support/support.h>
> +
> +/* We test for /proc/self/maps since that's one of the files that one
> +   of our tests actually uses, but the general idea is if Linux's
> +   /proc/ (procfs) filesystem is mounted.  If not, the process exits
> +   with an UNSUPPORTED result code.  */
> +
> +void
> +support_need_proc (const char *why_msg)
> +{
> +#ifdef __linux__
> +  if (access ("/proc/self/maps", R_OK))
> +    FAIL_UNSUPPORTED ("/proc is not available, %s", why_msg);
> +#endif

OK. Only does anything on linux (fixed from v5).

> +}
> diff --git a/support/test-container.c b/support/test-container.c
> index 25e7f142193..c837c4d7580 100644
> --- a/support/test-container.c
> +++ b/support/test-container.c
> @@ -97,6 +97,7 @@ int verbose = 0;
>     * mytest.root/mytest.script has a list of "commands" to run:
>         syntax:
>           # comment
> +	 pidns <comment>

Perfect.

This is a great way to avoid bitrot of this code while still making it functional.

>           su
>           mv FILE FILE
>  	 cp FILE FILE
> @@ -122,6 +123,8 @@ int verbose = 0;
>  
>         details:
>           - '#': A comment.
> +	 - 'pidns': Require a separate PID namespace, prints comment if it can't
> +	    (default is a shared pid namespace)

OK.

>           - 'su': Enables running test as root in the container.
>           - 'mv': A minimal move files command.
>           - 'cp': A minimal copy files command.
> @@ -148,7 +151,7 @@ int verbose = 0;
>     * Simple, easy to review code (i.e. prefer simple naive code over
>       complex efficient code)
>  
> -   * The current implementation ist parallel-make-safe, but only in
> +   * The current implementation is parallel-make-safe, but only in
>       that it uses a lock to prevent parallel access to the testroot.  */
>  
>  
> @@ -227,11 +230,37 @@ concat (const char *str, ...)
>    return bufs[n];
>  }
>  
> +/* Like the above, but put spaces between words.  Caller frees.  */
> +static char *
> +concat_words (char **words, int num_words)
> +{
> +  int len = 0;
> +  int i;
> +  char *rv, *p;
> +
> +  for (i = 0; i < num_words; i ++)
> +    {
> +      len += strlen (words[i]);
> +      len ++;
> +    }
> +
> +  p = rv = (char *) xmalloc (len);
> +
> +  for (i = 0; i < num_words; i ++)
> +    {
> +      if (i > 0)
> +	p = stpcpy (p, " ");
> +      p = stpcpy (p, words[i]);
> +    }
> +
> +  return rv;
> +}
> +
>  /* Try to mount SRC onto DEST.  */
>  static void
>  trymount (const char *src, const char *dest)
>  {
> -  if (mount (src, dest, "", MS_BIND, NULL) < 0)
> +  if (mount (src, dest, "", MS_BIND | MS_REC, NULL) < 0)
>      FAIL_EXIT1 ("can't mount %s onto %s\n", src, dest);
>  }
>  
> @@ -726,6 +755,9 @@ main (int argc, char **argv)
>    gid_t original_gid;
>    /* If set, the test runs as root instead of the user running the testsuite.  */
>    int be_su = 0;
> +  int require_pidns = 0;
> +  const char *pidns_comment = NULL;
> +  int do_proc_mounts = 0;
>    int UMAP;
>    int GMAP;
>    /* Used for "%lld %lld 1" so need not be large.  */
> @@ -1011,6 +1043,12 @@ main (int argc, char **argv)
>  	      {
>  		be_su = 1;
>  	      }
> +	    else if (nt >= 1 && strcmp (the_words[0], "pidns") == 0)
> +	      {
> +		require_pidns = 1;
> +		if (nt > 1)
> +		  pidns_comment = concat_words (the_words + 1, nt - 1);

OK.

> +	      }
>  	    else if (nt == 3 && strcmp (the_words[0], "mkdirp") == 0)
>  	      {
>  		long int m;
> @@ -1068,7 +1106,8 @@ main (int argc, char **argv)
>  
>  #ifdef CLONE_NEWNS
>    /* The unshare here gives us our own spaces and capabilities.  */
> -  if (unshare (CLONE_NEWUSER | CLONE_NEWPID | CLONE_NEWNS) < 0)
> +  if (unshare (CLONE_NEWUSER | CLONE_NEWNS
> +	       | (require_pidns ? CLONE_NEWPID : 0)) < 0)
>      {
>        /* Older kernels may not support all the options, or security
>  	 policy may block this call.  */
> @@ -1079,6 +1118,11 @@ main (int argc, char **argv)
>  	    check_for_unshare_hints ();
>  	  FAIL_UNSUPPORTED ("unable to unshare user/fs: %s", strerror (saved_errno));
>  	}
> +      /* We're about to exit anyway, it's "safe" to call unshare again
> +	 just to see if the CLONE_NEWPID caused the error.  */
> +      else if (require_pidns && unshare (CLONE_NEWUSER | CLONE_NEWNS) >= 0)
> +	FAIL_EXIT1 ("unable to unshare pid ns: %s : %s", strerror (errno),
> +		    pidns_comment ? pidns_comment : "required by test");
>        else
>  	FAIL_EXIT1 ("unable to unshare user/fs: %s", strerror (errno));
>      }
> @@ -1094,6 +1138,15 @@ main (int argc, char **argv)
>    trymount (support_srcdir_root, new_srcdir_path);
>    trymount (support_objdir_root, new_objdir_path);
>  
> +  /* It may not be possible to mount /proc directly.  */
> +  if (! require_pidns)
> +  {
> +    char *new_proc = concat (new_root_path, "/proc", NULL);
> +    xmkdirp (new_proc, 0755);
> +    trymount ("/proc", new_proc);
> +    do_proc_mounts = 1;
> +  }
> +
>    xmkdirp (concat (new_root_path, "/dev", NULL), 0755);
>    devmount (new_root_path, "null");
>    devmount (new_root_path, "zero");
> @@ -1163,42 +1216,60 @@ main (int argc, char **argv)
>  
>    maybe_xmkdir ("/tmp", 0755);
>  
> -  /* Now that we're pid 1 (effectively "root") we can mount /proc  */
> -  maybe_xmkdir ("/proc", 0777);
> -  if (mount ("proc", "/proc", "proc", 0, NULL) < 0)
> -    FAIL_EXIT1 ("Unable to mount /proc: ");
> -
> -  /* We map our original UID to the same UID in the container so we
> -     can own our own files normally.  */
> -  UMAP = open ("/proc/self/uid_map", O_WRONLY);
> -  if (UMAP < 0)
> -    FAIL_EXIT1 ("can't write to /proc/self/uid_map\n");
> -
> -  sprintf (tmp, "%lld %lld 1\n",
> -	   (long long) (be_su ? 0 : original_uid), (long long) original_uid);
> -  write (UMAP, tmp, strlen (tmp));
> -  xclose (UMAP);
> -
> -  /* We must disable setgroups () before we can map our groups, else we
> -     get EPERM.  */
> -  GMAP = open ("/proc/self/setgroups", O_WRONLY);
> -  if (GMAP >= 0)
> +  if (require_pidns)
>      {
> -      /* We support kernels old enough to not have this.  */
> -      write (GMAP, "deny\n", 5);
> -      xclose (GMAP);
> +      /* Now that we're pid 1 (effectively "root") we can mount /proc  */
> +      maybe_xmkdir ("/proc", 0777);
> +      if (mount ("proc", "/proc", "proc", 0, NULL) != 0)
> +	{
> +	  /* This happens if we're trying to create a nested container,
> +	     like if the build is running under podman, and we lack
> +	     priviledges.
> +
> +	     Ideally we would WARN here, but that would just add noise to
> +	     *every* test-container test, and the ones that care should
> +	     have their own relevent diagnostics.
> +
> +	     FAIL_EXIT1 ("Unable to mount /proc: ");  */
> +	}
> +      else
> +	do_proc_mounts = 1;
>      }
>  
> -  /* We map our original GID to the same GID in the container so we
> -     can own our own files normally.  */
> -  GMAP = open ("/proc/self/gid_map", O_WRONLY);
> -  if (GMAP < 0)
> -    FAIL_EXIT1 ("can't write to /proc/self/gid_map\n");
> +  if (do_proc_mounts)
> +    {
> +      /* We map our original UID to the same UID in the container so we
> +	 can own our own files normally.  */
> +      UMAP = open ("/proc/self/uid_map", O_WRONLY);
> +      if (UMAP < 0)
> +	FAIL_EXIT1 ("can't write to /proc/self/uid_map\n");
> +
> +      sprintf (tmp, "%lld %lld 1\n",
> +	       (long long) (be_su ? 0 : original_uid), (long long) original_uid);
> +      write (UMAP, tmp, strlen (tmp));
> +      xclose (UMAP);
> +
> +      /* We must disable setgroups () before we can map our groups, else we
> +	 get EPERM.  */
> +      GMAP = open ("/proc/self/setgroups", O_WRONLY);
> +      if (GMAP >= 0)
> +	{
> +	  /* We support kernels old enough to not have this.  */
> +	  write (GMAP, "deny\n", 5);
> +	  xclose (GMAP);
> +	}
>  
> -  sprintf (tmp, "%lld %lld 1\n",
> -	   (long long) (be_su ? 0 : original_gid), (long long) original_gid);
> -  write (GMAP, tmp, strlen (tmp));
> -  xclose (GMAP);
> +      /* We map our original GID to the same GID in the container so we
> +	 can own our own files normally.  */
> +      GMAP = open ("/proc/self/gid_map", O_WRONLY);
> +      if (GMAP < 0)
> +	FAIL_EXIT1 ("can't write to /proc/self/gid_map\n");
> +
> +      sprintf (tmp, "%lld %lld 1\n",
> +	       (long long) (be_su ? 0 : original_gid), (long long) original_gid);
> +      write (GMAP, tmp, strlen (tmp));
> +      xclose (GMAP);
> +    }
>  
>    if (change_cwd)
>      {
>

Thanks!  Pushed.

On Mon, 4 Apr 2022, Carlos O'Donell via Libc-alpha wrote:

> > @@ -726,6 +755,9 @@ main (int argc, char **argv)
> >    gid_t original_gid;
> >    /* If set, the test runs as root instead of the user running the testsuite.  */
> >    int be_su = 0;
> > +  int require_pidns = 0;
> > +  const char *pidns_comment = NULL;
> > +  int do_proc_mounts = 0;

I'm seeing build failures for i686-gnu:

test-container.c: In function 'main':
test-container.c:759:15: error: variable 'pidns_comment' set but not used [-Werror=unused-but-set-variable]
  759 |   const char *pidns_comment = NULL;
      |               ^~~~~~~~~~~~~
cc1: all warnings being treated as errors

https://sourceware.org/pipermail/libc-testresults/2022q2/009529.html

Maybe we need a Hurd trybot?  ;-)

I see two options: #ifdef the variable, or mark it attribute-unused.
Preferences?

diff --git a/support/test-container.c b/support/test-container.c
index c837c4d7580..79d0e052d1c 100644
--- a/support/test-container.c
+++ b/support/test-container.c
@@ -756,7 +756,9 @@ main (int argc, char **argv)
   /* If set, the test runs as root instead of the user running the testsuite.  */
   int be_su = 0;
   int require_pidns = 0;
+#ifdef CLONE_NEWNS
   const char *pidns_comment = NULL;
+#endif
   int do_proc_mounts = 0;
   int UMAP;
   int GMAP;
@@ -1046,8 +1048,10 @@ main (int argc, char **argv)
 	    else if (nt >= 1 && strcmp (the_words[0], "pidns") == 0)
 	      {
 		require_pidns = 1;
+#ifdef CLONE_NEWNS
 		if (nt > 1)
 		  pidns_comment = concat_words (the_words + 1, nt - 1);
+#endif
 	      }
 	    else if (nt == 3 && strcmp (the_words[0], "mkdirp") == 0)
 	      {

* DJ Delorie via Libc-alpha:

> Maybe we need a Hurd trybot?  ;-)
>
> I see two options: #ifdef the variable, or mark it attribute-unused.
> Preferences?
>
> diff --git a/support/test-container.c b/support/test-container.c
> index c837c4d7580..79d0e052d1c 100644
> --- a/support/test-container.c
> +++ b/support/test-container.c
> @@ -756,7 +756,9 @@ main (int argc, char **argv)
>    /* If set, the test runs as root instead of the user running the testsuite.  */
>    int be_su = 0;
>    int require_pidns = 0;
> +#ifdef CLONE_NEWNS
>    const char *pidns_comment = NULL;
> +#endif
>    int do_proc_mounts = 0;
>    int UMAP;
>    int GMAP;
> @@ -1046,8 +1048,10 @@ main (int argc, char **argv)
>  	    else if (nt >= 1 && strcmp (the_words[0], "pidns") == 0)
>  	      {
>  		require_pidns = 1;
> +#ifdef CLONE_NEWNS
>  		if (nt > 1)
>  		  pidns_comment = concat_words (the_words + 1, nt - 1);
> +#endif
>  	      }
>  	    else if (nt == 3 && strcmp (the_words[0], "mkdirp") == 0)
>  	      {

This also needs:

diff --git a/support/test-container.c b/support/test-container.c
index 79d0e052d1..9dde784bf0 100644
--- a/support/test-container.c
+++ b/support/test-container.c
@@ -231,7 +231,7 @@ concat (const char *str, ...)
 }
 
 /* Like the above, but put spaces between words.  Caller frees.  */
-static char *
+static inline char *
 concat_words (char **words, int num_words)
 {
   int len = 0;

Or something like it.

Thanks,
Florian