Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1570804 > unrolled thread

[PATCH] jbd2: Fix use after free in kjournald2()

Started bySahitya Tummala <stummala@codeaurora.org>
First post2017-01-31 15:50 +0100
Last post2017-02-01 05:30 +0100
Articles 3 — 3 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH] jbd2: Fix use after free in kjournald2() Sahitya Tummala <stummala@codeaurora.org> - 2017-01-31 15:50 +0100
    Re: [PATCH] jbd2: Fix use after free in kjournald2() Jan Kara <jack@suse.cz> - 2017-01-31 17:00 +0100
      Re: [PATCH] jbd2: Fix use after free in kjournald2() "Tummala, Sahitya" <stummala@codeaurora.org> - 2017-02-01 05:30 +0100

#1570804 — [PATCH] jbd2: Fix use after free in kjournald2()

FromSahitya Tummala <stummala@codeaurora.org>
Date2017-01-31 15:50 +0100
Subject[PATCH] jbd2: Fix use after free in kjournald2()
Message-ID<t5HUS-1S1-29@gated-at.bofh.it>
Below is the synchronization issue between unmount and kjournald2
contexts, which results into use after free issue in kjournald2().
Fix this issue by using journal->j_state_lock to synchronize the
wait_event() done in journal_kill_thread() and the wake_up() done
in kjournald2().

TASK 1:
umount cmd:
   |--jbd2_journal_destroy() {
       |--journal_kill_thread() {
            write_lock(&journal->j_state_lock);
	    journal->j_flags |= JBD2_UNMOUNT;
	    ...
	    write_unlock(&journal->j_state_lock);
	    wake_up(&journal->j_wait_commit);	   TASK 2 wakes up here:
	    					   kjournald2() {
						     ...
						     checks JBD2_UNMOUNT flag and calls goto end-loop;
						     ...
						     end_loop:
						       write_unlock(&journal->j_state_lock);
						       journal->j_task = NULL; --> If this thread gets
						       pre-empted here, then TASK 1 wait_event will
						       exit even before this thread is completely
						       done.
	    wait_event(journal->j_wait_done_commit, journal->j_task == NULL);
	    ...
	    write_lock(&journal->j_state_lock);
	    write_unlock(&journal->j_state_lock);
	  }
       |--kfree(journal);
     }
}
						       wake_up(&journal->j_wait_done_commit); --> this step
						       now results into use after free issue.
						   }

Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
---
 fs/jbd2/journal.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index a097048..f5cd3c0 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -278,9 +278,11 @@ static int kjournald2(void *arg)
 end_loop:
 	write_unlock(&journal->j_state_lock);
 	del_timer_sync(&journal->j_commit_timer);
+	write_lock(&journal->j_state_lock);
 	journal->j_task = NULL;
 	wake_up(&journal->j_wait_done_commit);
 	jbd_debug(1, "Journal thread exiting.\n");
+	write_unlock(&journal->j_state_lock);
 	return 0;
 }
 
-- 
Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

[toc] | [next] | [standalone]


#1570848

FromJan Kara <jack@suse.cz>
Date2017-01-31 17:00 +0100
Message-ID<t5J0C-2tO-31@gated-at.bofh.it>
In reply to#1570804
On Tue 31-01-17 20:08:57, Sahitya Tummala wrote:
> Below is the synchronization issue between unmount and kjournald2
> contexts, which results into use after free issue in kjournald2().
> Fix this issue by using journal->j_state_lock to synchronize the
> wait_event() done in journal_kill_thread() and the wake_up() done
> in kjournald2().
> 
> TASK 1:
> umount cmd:
>    |--jbd2_journal_destroy() {
>        |--journal_kill_thread() {
>             write_lock(&journal->j_state_lock);
> 	    journal->j_flags |= JBD2_UNMOUNT;
> 	    ...
> 	    write_unlock(&journal->j_state_lock);
> 	    wake_up(&journal->j_wait_commit);	   TASK 2 wakes up here:
> 	    					   kjournald2() {
> 						     ...
> 						     checks JBD2_UNMOUNT flag and calls goto end-loop;
> 						     ...
> 						     end_loop:
> 						       write_unlock(&journal->j_state_lock);
> 						       journal->j_task = NULL; --> If this thread gets
> 						       pre-empted here, then TASK 1 wait_event will
> 						       exit even before this thread is completely
> 						       done.
> 	    wait_event(journal->j_wait_done_commit, journal->j_task == NULL);
> 	    ...
> 	    write_lock(&journal->j_state_lock);
> 	    write_unlock(&journal->j_state_lock);
> 	  }
>        |--kfree(journal);
>      }
> }
> 						       wake_up(&journal->j_wait_done_commit); --> this step
> 						       now results into use after free issue.
> 						   }
> 
> Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>

Yeah, what you write looks possible (although rather unlikely). Thanks for
catching this. One small nit below:

> diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
> index a097048..f5cd3c0 100644
> --- a/fs/jbd2/journal.c
> +++ b/fs/jbd2/journal.c
> @@ -278,9 +278,11 @@ static int kjournald2(void *arg)
>  end_loop:
>  	write_unlock(&journal->j_state_lock);
>  	del_timer_sync(&journal->j_commit_timer);
> +	write_lock(&journal->j_state_lock);

There's no good reason to do del_timer_sync() outside of j_state_lock. This
is not performance critical code and commit_timeout is trivial and cannot
block on anything. So just keep j_state_lock locked upto the place where
you unlock it now...

								Honza
>  	journal->j_task = NULL;
>  	wake_up(&journal->j_wait_done_commit);
>  	jbd_debug(1, "Journal thread exiting.\n");
> +	write_unlock(&journal->j_state_lock);
>  	return 0;
>  }
>  
> -- 
> Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc.
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.
> 
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

[toc] | [prev] | [next] | [standalone]


#1571263

From"Tummala, Sahitya" <stummala@codeaurora.org>
Date2017-02-01 05:30 +0100
Message-ID<t5UIp-1aG-1@gated-at.bofh.it>
In reply to#1570848
On 1/31/2017 9:21 PM, Jan Kara wrote:
> On Tue 31-01-17 20:08:57, Sahitya Tummala wrote:
>> Below is the synchronization issue between unmount and kjournald2
>> contexts, which results into use after free issue in kjournald2().
>> Fix this issue by using journal->j_state_lock to synchronize the
>> wait_event() done in journal_kill_thread() and the wake_up() done
>> in kjournald2().
>>
>> TASK 1:
>> umount cmd:
>>     |--jbd2_journal_destroy() {
>>         |--journal_kill_thread() {
>>              write_lock(&journal->j_state_lock);
>> 	    journal->j_flags |= JBD2_UNMOUNT;
>> 	    ...
>> 	    write_unlock(&journal->j_state_lock);
>> 	    wake_up(&journal->j_wait_commit);	   TASK 2 wakes up here:
>> 	    					   kjournald2() {
>> 						     ...
>> 						     checks JBD2_UNMOUNT flag and calls goto end-loop;
>> 						     ...
>> 						     end_loop:
>> 						       write_unlock(&journal->j_state_lock);
>> 						       journal->j_task = NULL; --> If this thread gets
>> 						       pre-empted here, then TASK 1 wait_event will
>> 						       exit even before this thread is completely
>> 						       done.
>> 	    wait_event(journal->j_wait_done_commit, journal->j_task == NULL);
>> 	    ...
>> 	    write_lock(&journal->j_state_lock);
>> 	    write_unlock(&journal->j_state_lock);
>> 	  }
>>         |--kfree(journal);
>>       }
>> }
>> 						       wake_up(&journal->j_wait_done_commit); --> this step
>> 						       now results into use after free issue.
>> 						   }
>>
>> Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
> Yeah, what you write looks possible (although rather unlikely). Thanks for
> catching this. One small nit below:
Yes, it was observed only once and is very hard to reproduce.
>> diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
>> index a097048..f5cd3c0 100644
>> --- a/fs/jbd2/journal.c
>> +++ b/fs/jbd2/journal.c
>> @@ -278,9 +278,11 @@ static int kjournald2(void *arg)
>>   end_loop:
>>   	write_unlock(&journal->j_state_lock);
>>   	del_timer_sync(&journal->j_commit_timer);
>> +	write_lock(&journal->j_state_lock);
> There's no good reason to do del_timer_sync() outside of j_state_lock. This
> is not performance critical code and commit_timeout is trivial and cannot
> block on anything. So just keep j_state_lock locked upto the place where
> you unlock it now...
>
Sure, I will update the patch.
> 								Honza
>>   	journal->j_task = NULL;
>>   	wake_up(&journal->j_wait_done_commit);
>>   	jbd_debug(1, "Journal thread exiting.\n");
>> +	write_unlock(&journal->j_state_lock);
>>   	return 0;
>>   }
>>   
>> -- 
>> Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc.
>> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.
>>
>>

-- 
Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web