linux-kernel - RE: [PATCH] Fix rq->lock vs logbuf

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] [day] [month] [year] [list]

Date:	Wed, 20 Feb 2013 11:24:36 +0000
From:	"Bu, Yitian" <ybu@....qualcomm.com>
To:	Peter Zijlstra <a.p.zijlstra@...llo.nl>
CC:	"tglx@...utronix.de" <tglx@...utronix.de>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"mingo@...nel.org" <mingo@...nel.org>
Subject: RE: [PATCH] Fix rq->lock vs logbuf_lock unlock race

> Unfortunately that's not quite possible, rq->lock is really out of bounds. At
> one point I tried 'fixing' this but there's a whole bunch of nasty that's not
> going to go away.
> 
> I've since forgotten most of the details, but aside from logbuf problems
> there's a whole host of issues with the console drivers themselves as well.
> 
> If you really want to do this, use early_printk.

Hi Peter:

The patch  0b5e1c5255e("printk: Release console_sem after logbuf_lock" )
is as below:

			console_locked = 0;
-			up(&console_sem);
+			wake = 1;
 			retval = 0;
 		}
 	}
 	printk_cpu = UINT_MAX;
 	spin_unlock(&logbuf_lock);
+	if (wake)
+		up(&console_sem);
 	return retval;

the patch 07354eb1a74d1("locking, printk: Annotate logbuf_lock as raw ")
is as below:

	printk_cpu = UINT_MAX;
-	spin_unlock(&logbuf_lock);
 	if (wake)
 		up(&console_sem);
+	raw_spin_unlock(&logbuf_lock);
 	return retval;
 }

Note that the purpose of patch 07354eb1a74d1 is to use raw_spin_unlock
instead of spin_unlock, it is not supposed to change any lock sequences.

What I do is to revert this change to 0b5e1c5255e, and it is nothing to do with
how people use printk. It is just to recover a merge mistake.