linux-kernel - Re: Re: Re: [PATCH] lib/test_kmod: Fix a use after free in register_test_dev

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:   Thu, 11 Mar 2021 22:40:33 +0800 (GMT+08:00)
From:   lyl2019@...l.ustc.edu.cn
To:     "Luis Chamberlain" <mcgrof@...nel.org>
Cc:     linux-kernel@...r.kernel.org
Subject: Re: Re: Re: [PATCH] lib/test_kmod: Fix a use after free in
 register_test_dev_kmod




> -----原始邮件-----
> 发件人: "Luis Chamberlain" <mcgrof@...nel.org>
> 发送时间: 2021-03-11 21:58:33 (星期四)
> 收件人: lyl2019@...l.ustc.edu.cn
> 抄送: linux-kernel@...r.kernel.org
> 主题: Re: Re: [PATCH] lib/test_kmod: Fix a use after free in register_test_dev_kmod
> 
> On Thu, Mar 11, 2021 at 09:31:55PM +0800, lyl2019@...l.ustc.edu.cn wrote:
> > 
> > 
> > 
> > > -----原始邮件-----
> > > 发件人: "Luis Chamberlain" <mcgrof@...nel.org>
> > > 发送时间: 2021-03-11 21:01:08 (星期四)
> > > 收件人: "Lv Yunlong" <lyl2019@...l.ustc.edu.cn>
> > > 抄送: linux-kernel@...r.kernel.org
> > > 主题: Re: [PATCH] lib/test_kmod: Fix a use after free in register_test_dev_kmod
> > > 
> > > On Thu, Mar 11, 2021 at 12:02:46AM -0800, Lv Yunlong wrote:
> > > > In register_test_dev_kmod, it calls free_test_dev_kmod() to free
> > > > test_dev. But free_test_dev_kmod() can't set the original pointer
> > > > test_dev to NULL, because the test_dev was passed by it's value
> > > > not reference.
> > > 
> > > Did you actually get a crash or something? If so can you supply the
> > > actual log? If this is just an observation and you think this is an
> > > issue, specifying that would help during patch review.
> > > 
> > >   Luis
> > > 
> > > > Signed-off-by: Lv Yunlong <lyl2019@...l.ustc.edu.cn>
> > > > ---
> > > >  lib/test_kmod.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > 
> > > > diff --git a/lib/test_kmod.c b/lib/test_kmod.c
> > > > index 38c250fbace3..aa8a2a563d7e 100644
> > > > --- a/lib/test_kmod.c
> > > > +++ b/lib/test_kmod.c
> > > > @@ -1124,7 +1124,6 @@ static void free_test_dev_kmod(struct kmod_test_device *test_dev)
> > > >  		free_test_dev_info(test_dev);
> > > >  		kmod_config_free(test_dev);
> > > >  		vfree(test_dev);
> > > > -		test_dev = NULL;
> > > >  	}
> > > >  }
> > > >  
> > > > @@ -1149,6 +1148,7 @@ static struct kmod_test_device *register_test_dev_kmod(void)
> > > >  	if (ret) {
> > > >  		pr_err("could not register misc device: %d\n", ret);
> > > >  		free_test_dev_kmod(test_dev);
> > > > +		test_dev = NULL;
> > > >  		goto out;
> > > >  	}
> > > >  
> > > > -- 
> > > > 2.25.1
> > > > 
> > > > 
> > 
> > This problem was reported by source code analyzers developed by our Security Lab(Loccs).
> 
> This should be included in the commit log, please.
> 
> > We have confirmed this issue before submiting the patch.
> 
> How was this confirmed exactly.
> 
>   Luis

Thanks for your advice.

Ok, let's see the details. In the function test_kmod_init(), it calls register_test_dev_kmod()
to return a test_dev. Inside the callee function register_test_dev_kmod(), the test_dev is 
allocated by alloc_test_dev_kmod() and freed by free_test_dev_kmod(). 

Notice the implementation of free_test_dev_kmod, the 'test_dev' in this function actually is
a new variable in the parameter's stack space. Set the value of 'test_dev' inside this function 
will not affect the outside 'test_dev'.
"""
static void free_test_dev_kmod(struct kmod_test_device *test_dev)
{
	if (test_dev) {
		kfree_const(test_dev->misc_dev.name);
		test_dev->misc_dev.name = NULL;
		free_test_dev_info(test_dev);
		kmod_config_free(test_dev);
		vfree(test_dev);
		test_dev = NULL;
	}
}
"""

So, register_test_dev_kmod() will return a valid and freed test_dev, and cause use after free
in function test_kmod_init().
"""
static int __init test_kmod_init(void)
{
	test_dev = register_test_dev_kmod();
	if (!test_dev) {                     // freed and valid.
		pr_err("Cannot add first test kmod device\n");
		return -ENODEV;
	}

	ret = trigger_config_run_type(test_dev,  // cause use after free here!
					      TEST_KMOD_DRIVER, "tun");
"""

I think i should write more details in the first commit, i'm sorry for wasting your time.

Thanks.