From: DJ Delorie <dj@redhat.com>
To: libc-alpha@sourceware.org
Subject: nsswitch: do not reload if "/" changes
Date: Fri, 15 Jan 2021 19:59:12 -0500 [thread overview]
Message-ID: <xnsg71wvbz.fsf@greed.delorie.com> (raw)
[Note: I tried putting this functionality in the file_change_detection
module, but that didn't have enough persistence.]
[Note: tested by instrumenting test-container.c and observing the
instrumentation with test containers on the root fs and on a separate
fs]
https://sourceware.org/bugzilla/show_bug.cgi?id=27077
Before reloading nsswitch.conf, verify that the root directory
hasn't changed - if it has, it's likely that we've entered a
container and should not trust the nsswitch inside the container
nor load any shared objects therein.
diff --git a/nss/nss_database.c b/nss/nss_database.c
index e719ec0865..580ea7b963 100644
--- a/nss/nss_database.c
+++ b/nss/nss_database.c
@@ -33,6 +33,11 @@ struct nss_database_state
{
struct nss_database_data data;
__libc_lock_define (, lock);
+ /* If "/" changes, we switched into a container and do NOT want to
+ reload anything. This data must be persistent across
+ reloads. */
+ ino64_t root_ino;
+ ino64_t root_dev;
};
@@ -53,6 +58,8 @@ global_state_allocate (void *closure)
memset (result->data.services, 0, sizeof (result->data.services));
result->data.initialized = true;
result->data.reload_disabled = false;
+ result->root_ino = 0;
+ result->root_dev = 0;
__libc_lock_init (result->lock);
}
return result;
@@ -356,6 +363,8 @@ nss_database_check_reload_and_get (struct nss_database_state *local,
nss_action_list *result,
enum nss_database database_index)
{
+ struct stat64 str;
+
/* Acquire MO is needed because the thread that sets reload_disabled
may have loaded the configuration first, so synchronize with the
Release MO store there. */
@@ -379,6 +388,25 @@ nss_database_check_reload_and_get (struct nss_database_state *local,
__libc_lock_unlock (local->lock);
return true;
}
+
+ /* Before we reload, verify that "/" hasn't changed. We assume that
+ errors here are very unlikely, but the chance that we're entering
+ a container is also very unlikely, so we err on the side of both
+ very unlikely things not happening at the same time. */
+ if (__stat64 ("/", &str) == 0)
+ {
+ if (local->root_ino != 0
+ && (str.st_ino != local->root_ino
+ || str.st_dev != local->root_dev))
+ {
+ /* Change detected; disable reloading. */
+ atomic_store_release (&local->data.reload_disabled, 1);
+ __libc_lock_unlock (local->lock);
+ return true;
+ }
+ local->root_ino = str.st_ino;
+ local->root_dev = str.st_dev;
+ }
__libc_lock_unlock (local->lock);
/* Avoid overwriting the global configuration until we have loaded
diff --git a/nss/nss_database.h b/nss/nss_database.h
index 1f827e6def..f94c629174 100644
--- a/nss/nss_database.h
+++ b/nss/nss_database.h
@@ -75,6 +75,10 @@ struct nss_database_data
nss_action_list services[NSS_DATABASE_COUNT];
int reload_disabled; /* Actually bool; int for atomic access. */
bool initialized;
+ /* If "/" changes, we switched into a container and do NOT want to
+ reload anything. */
+ ino64_t root_ino;
+ ino64_t root_dev;
};
/* Called by fork in the parent process, before forking. */
next reply other threads:[~2021-01-16 0:59 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-16 0:59 DJ Delorie [this message]
2021-01-16 10:52 ` Florian Weimer
2021-01-18 1:13 ` DJ Delorie
2021-01-18 10:47 ` Florian Weimer
2021-01-18 18:20 ` DJ Delorie
2021-01-19 16:37 ` Florian Weimer
2021-01-22 19:10 ` [v2] " DJ Delorie
2021-01-26 9:58 ` Florian Weimer
2021-01-26 16:19 ` DJ Delorie
2021-01-26 16:30 ` Florian Weimer
2021-01-26 16:47 ` DJ Delorie
2021-01-27 17:28 ` Carlos O'Donell
2021-01-27 18:44 ` DJ Delorie
2021-01-28 0:31 ` Joseph Myers
2021-01-28 0:34 ` DJ Delorie
2021-01-28 0:39 ` Joseph Myers
2021-01-28 1:15 ` DJ Delorie
2021-01-18 12:42 ` Andreas Schwab
2021-01-18 12:53 ` Florian Weimer
2021-01-18 18:27 ` DJ Delorie
2021-01-18 15:59 ` Carlos O'Donell
2021-01-18 16:53 ` Florian Weimer
2021-01-19 14:30 ` Carlos O'Donell
2021-01-19 14:40 ` Florian Weimer
2021-01-18 18:35 ` DJ Delorie
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=xnsg71wvbz.fsf@greed.delorie.com \
--to=dj@redhat.com \
--cc=libc-alpha@sourceware.org \
/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).