linux-kernel - Re: [PATCH] module: fix refptr allocation and release order

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]

Date:	Tue, 17 Mar 2009 04:24:36 +0100
From:	Eric Dumazet <dada1@...mosbay.com>
To:	Rusty Russell <rusty@...tcorp.com.au>
CC:	Linus Torvalds <torvalds@...ux-foundation.org>,
	Masami Hiramatsu <mhiramat@...hat.com>,
	LKML <linux-kernel@...r.kernel.org>,
	systemtap-ml <systemtap@...rces.redhat.com>
Subject: Re: [PATCH] module: fix refptr allocation and release order

Rusty Russell a écrit :
> From: Masami Hiramatsu <mhiramat@...hat.com>
> 
> Impact: fix ref-after-free crash on failed module load
> 
> Fix refptr bug: Change refptr allocation and release order not to access a module
> data structure pointed by 'mod' after freeing mod->module_core.
> This bug will cause kernel panic(e.g. failed to find undefined symbols).
> 
> This bug was reported on systemtap bugzilla.
> http://sources.redhat.com/bugzilla/show_bug.cgi?id=9927
> 
> Signed-off-by: Masami Hiramatsu <mhiramat@...hat.com>
> Cc: Eric Dumazet <dada1@...mosbay.com>
> Signed-off-by: Rusty Russell <rusty@...tcorp.com.au>


My original patch did not have this problem, because I used a local variable
to hold refptr.

http://www.archivum.info/linux-kernel@vger.kernel.org/2008-05/msg07400.html

A simpler patch could just use a local variable again, since we are very
late in rc phase ?


From: Masami Hiramatsu <mhiramat@...hat.com>

Impact: fix ref-after-free crash on failed module load

Fix refptr bug: Change refptr release not to access a module
data structure pointed by 'mod' after freeing mod->module_core.
This bug will cause kernel panic(e.g. failed to find undefined symbols).

This bug was reported on systemtap bugzilla.
http://sources.redhat.com/bugzilla/show_bug.cgi?id=9927

Signed-off-by: Masami Hiramatsu <mhiramat@...hat.com>
Signed-off-by: Eric Dumazet <dada1@...mosbay.com>
Signed-off-by: Rusty Russell <rusty@...tcorp.com.au>


diff --git a/kernel/module.c b/kernel/module.c
index ba22484..4dd1228 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -1874,6 +1874,9 @@ static noinline struct module *load_module(void __user *umod,
 	struct kernel_param *kp;
 	struct module *mod;
 	long err = 0;
+#if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP)
+	void *refptr = NULL;
+#endif
 	void *percpu = NULL, *ptr = NULL; /* Stops spurious gcc warning */
 	unsigned long *mseg;
 	mm_segment_t old_fs;
@@ -2018,6 +2021,7 @@ static noinline struct module *load_module(void __user *umod,
 #if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP)
 	mod->refptr = percpu_modalloc(sizeof(local_t), __alignof__(local_t),
 				      mod->name);
+	refptr = mod->refptr;
 	if (!mod->refptr) {
 		err = -ENOMEM;
 		goto free_mod;
@@ -2292,10 +2296,14 @@ static noinline struct module *load_module(void __user *umod,
  free_core:
 	module_free(mod, mod->module_core);
  free_percpu:
+	/*
+	 * Do not access to mod anymore, it's now freed
+	 */
 	if (percpu)
 		percpu_modfree(percpu);
 #if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP)
-	percpu_modfree(mod->refptr);
+	if (refptr)
+		percpu_modfree(refptr);
 #endif
  free_mod:
 	kfree(args);

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/