From: Joe Simmons-Talbott <josimmon@redhat.com>
To: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
Cc: libc-alpha@sourceware.org, Andreas Schwab <schwab@suse.de>,
"Carlos O'Donell" <carlos@redhat.com>
Subject: Re: [PATCH v3] elf/rtld: Count skipped environment variables for enable_secure
Date: Fri, 12 Apr 2024 11:23:48 -0400 [thread overview]
Message-ID: <CAAQBMsOxdKtOfDxgv+kkcaaghb0A8n0XwG=h=ih3=7w-4+o2hw@mail.gmail.com> (raw)
In-Reply-To: <2c2922df-4900-4efb-b317-6d6760c5a61c@linaro.org>
On Fri, Apr 12, 2024 at 10:14 AM Adhemerval Zanella Netto
<adhemerval.zanella@linaro.org> wrote:
>
>
>
> On 11/04/24 13:53, Joe Simmons-Talbott wrote:
> > When using the glibc.rtld.enable_secure tunable we need to keep track of
> > the count of environment variables we skip due to __libc_enable_secure
> > being set and adjust the auxv section of the stack. This fixes an
> > assertion when running ld.so directly with glibc.rtld.enable_secure set.
> > Add a testcase that ensures the assert is not hit.
> >
> > elf/rtld.c:1324 assert (auxv == sp + 1);
> > ---
> > Changes to v2:
> > * Add testcase that triggers the assertion before this patch and does
> > not after the patch.
> >
> > Changes to v1:
> > * Add comment explaining how and why skip_env is adjusted.
> >
> > elf/Makefile | 3 +++
> > elf/rtld.c | 31 +++++++++++++++++++-------
> > elf/tst-tunables-enable_secure-env.c | 33 ++++++++++++++++++++++++++++
> > 3 files changed, 59 insertions(+), 8 deletions(-)
> > create mode 100644 elf/tst-tunables-enable_secure-env.c
> >
> > diff --git a/elf/Makefile b/elf/Makefile
> > index 6dad11bcfb..5c35db1799 100644
> > --- a/elf/Makefile
> > +++ b/elf/Makefile
> > @@ -315,6 +315,7 @@ tests := \
> > tst-leaks1 \
> > tst-stringtable \
> > tst-tls9 \
> > + tst-tunables-enable_secure-env \
> > # tests
> >
> > tests-internal := \
> > @@ -336,6 +337,8 @@ tst-tls9-static-ENV = $(static-dlopen-environment)
> > tst-single_threaded-static-dlopen-ENV = $(static-dlopen-environment)
> > tst-rootdir-ENV = LD_LIBRARY_PATH=/
> >
> > +tst-tunables-enable_secure-env-ENV = GLIBC_TUNABLES=glibc.rtld.enable_secure=1
> > +
>
> I think since the idea is to check for running ld.so directly, you will need to
> explicit issue the loader:
>
> diff --git a/elf/Makefile b/elf/Makefile
> index 5c35db1799..1598fbd6d7 100644
> --- a/elf/Makefile
> +++ b/elf/Makefile
> @@ -337,8 +337,6 @@ tst-tls9-static-ENV = $(static-dlopen-environment)
> tst-single_threaded-static-dlopen-ENV = $(static-dlopen-environment)
> tst-rootdir-ENV = LD_LIBRARY_PATH=/
>
> -tst-tunables-enable_secure-env-ENV = GLIBC_TUNABLES=glibc.rtld.enable_secure=1
> -
> tests += \
> argv0test \
> constload1 \
> @@ -1192,6 +1190,7 @@ tests-special += \
> $(objpfx)tst-trace3.out \
> $(objpfx)tst-trace4.out \
> $(objpfx)tst-trace5.out \
> + $(objpfx)tst-tunables-enable_secure-env.out \
> $(objpfx)tst-unused-dep-cmp.out \
> $(objpfx)tst-unused-dep.out \
> # tests-special
> @@ -2219,6 +2218,13 @@ $(objpfx)tst-unused-dep-cmp.out: $(objpfx)tst-unused-dep.out
> cmp $< /dev/null > $@; \
> $(evaluate-test)
>
> +$(objpfx)tst-tunables-enable_secure-env.out: $(objpfx)tst-tunables-enable_secure-env
> + $(test-wrapper-env) \
> + GLIBC_TUNABLES=glibc.rtld.enable_secure=1 \
> + $(rtld-prefix) \
> + $< > $@; \
> + $(evaluate-test)
> +
> $(objpfx)tst-audit11.out: $(objpfx)tst-auditmod11.so $(objpfx)tst-audit11mod1.so
> tst-audit11-ENV = LD_AUDIT=$(objpfx)tst-auditmod11.so
> $(objpfx)tst-audit11mod1.so: $(objpfx)tst-audit11mod2.so
>
> Otherwise this won't check the bug with --enable-hardcoded-path-in-tests.
>
> The rest looks ok.
Thanks for the review. I've posted a v4 [1] with this change.
Thanks,
Joe
[1] https://sourceware.org/pipermail/libc-alpha/2024-April/156040.html
>
> > tests += \
> > argv0test \
> > constload1 \
> > diff --git a/elf/rtld.c b/elf/rtld.c
> > index 18d73f19c6..d116a436f5 100644
> > --- a/elf/rtld.c
> > +++ b/elf/rtld.c
> > @@ -155,7 +155,7 @@ static void dl_main_state_init (struct dl_main_state *state);
> > Since all of them start with `LD_' we are a bit smarter while finding
> > all the entries. */
> > extern char **_environ attribute_hidden;
> > -static void process_envvars (struct dl_main_state *state);
> > +static int process_envvars (struct dl_main_state *state);
> >
> > int _dl_argc attribute_relro attribute_hidden;
> > char **_dl_argv attribute_relro = NULL;
> > @@ -1289,7 +1289,7 @@ rtld_setup_main_map (struct link_map *main_map)
> > _dl_argv and _dl_argc accordingly. Those arguments are removed from
> > argv here. */
> > static void
> > -_dl_start_args_adjust (int skip_args)
> > +_dl_start_args_adjust (int skip_args, int skip_env)
> > {
> > void **sp = (void **) (_dl_argv - skip_args - 1);
> > void **p = sp + skip_args;
> > @@ -1321,7 +1321,7 @@ _dl_start_args_adjust (int skip_args)
> > while (*p != NULL);
> >
> > #ifdef HAVE_AUX_VECTOR
> > - void **auxv = (void **) GLRO(dl_auxv) - skip_args;
> > + void **auxv = (void **) GLRO(dl_auxv) - skip_args - skip_env;
> > GLRO(dl_auxv) = (ElfW(auxv_t) *) auxv; /* Aliasing violation. */
> > assert (auxv == sp + 1);
> >
> > @@ -1352,6 +1352,7 @@ dl_main (const ElfW(Phdr) *phdr,
> > unsigned int i;
> > bool rtld_is_main = false;
> > void *tcbp = NULL;
> > + int skip_env = 0;
> >
> > struct dl_main_state state;
> > dl_main_state_init (&state);
> > @@ -1365,7 +1366,7 @@ dl_main (const ElfW(Phdr) *phdr,
> > #endif
> >
> > /* Process the environment variable which control the behaviour. */
> > - process_envvars (&state);
> > + skip_env = process_envvars (&state);
> >
> > #ifndef HAVE_INLINED_SYSCALLS
> > /* Set up a flag which tells we are just starting. */
> > @@ -1630,7 +1631,7 @@ dl_main (const ElfW(Phdr) *phdr,
> > _dl_argv[0] = argv0;
> >
> > /* Adjust arguments for the application entry point. */
> > - _dl_start_args_adjust (_dl_argv - orig_argv);
> > + _dl_start_args_adjust (_dl_argv - orig_argv, skip_env);
> > }
> > else
> > {
> > @@ -2534,11 +2535,12 @@ a filename can be specified using the LD_DEBUG_OUTPUT environment variable.\n");
> > }
> > }
> >
> > -static void
> > +static int
> > process_envvars_secure (struct dl_main_state *state)
> > {
> > char **runp = _environ;
> > char *envline;
> > + int skip_env = 0;
> >
> > while ((envline = _dl_next_ld_env_entry (&runp)) != NULL)
> > {
> > @@ -2580,6 +2582,14 @@ process_envvars_secure (struct dl_main_state *state)
> > const char *nextp = UNSECURE_ENVVARS;
> > do
> > {
> > + /* Keep track of the number of environment variables that were set in
> > + the environment and are unset below. Use getenv() which returns
> > + non-NULL if the variable is set in the environment. This count is
> > + needed if we need to adjust the location of the AUX vector on the
> > + stack when running ld.so directly. */
> > + if (getenv (nextp) != NULL)
> > + skip_env++;
> > +
> > unsetenv (nextp);
> > nextp = strchr (nextp, '\0') + 1;
> > }
> > @@ -2592,6 +2602,8 @@ process_envvars_secure (struct dl_main_state *state)
> > || state->mode != rtld_mode_normal
> > || state->version_info)
> > _exit (5);
> > +
> > + return skip_env;
> > }
> >
> > static void
> > @@ -2745,13 +2757,16 @@ process_envvars_default (struct dl_main_state *state)
> > }
> > }
> >
> > -static void
> > +static int
> > process_envvars (struct dl_main_state *state)
> > {
> > + int skip_env = 0;
> > if (__glibc_unlikely (__libc_enable_secure))
> > - process_envvars_secure (state);
> > + skip_env += process_envvars_secure (state);
> > else
> > process_envvars_default (state);
> > +
> > + return skip_env;
> > }
> >
> > #if HP_TIMING_INLINE
> > diff --git a/elf/tst-tunables-enable_secure-env.c b/elf/tst-tunables-enable_secure-env.c
> > new file mode 100644
> > index 0000000000..24e846f299
> > --- /dev/null
> > +++ b/elf/tst-tunables-enable_secure-env.c
> > @@ -0,0 +1,33 @@
> > +/* Check enable_secure tunable handles removed ENV variables without
> > + assertions.
> > + Copyright (C) 2024 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 <support/capture_subprocess.h>
> > +#include <support/check.h>
> > +
> > +static int
> > +do_test (int argc, char *argv[])
> > +{
> > + /* Ensure that no assertions are hit when a dynamically linked application
> > + runs. This test requires that GLIBC_TUNABLES=glibc.rtld.enable_secure=1
> > + is set. */
> > + return 0;
> > +}
> > +
> > +#define TEST_FUNCTION_ARGV do_test
> > +#include <support/test-driver.c>
>
--
Joe Simmons-Talbott
prev parent reply other threads:[~2024-04-12 15:24 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-11 16:53 Joe Simmons-Talbott
2024-04-12 14:13 ` Adhemerval Zanella Netto
2024-04-12 15:23 ` Joe Simmons-Talbott [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAAQBMsOxdKtOfDxgv+kkcaaghb0A8n0XwG=h=ih3=7w-4+o2hw@mail.gmail.com' \
--to=josimmon@redhat.com \
--cc=adhemerval.zanella@linaro.org \
--cc=carlos@redhat.com \
--cc=libc-alpha@sourceware.org \
--cc=schwab@suse.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).