linux-kernel - Re: [Patch 01/11] Introducing generic hardware breakpoint handler interfaces

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]

Message-ID: <Pine.LNX.4.44L0.0903231713360.3610-100000@iolanthe.rowland.org>
Date:	Mon, 23 Mar 2009 17:20:27 -0400 (EDT)
From:	Alan Stern <stern@...land.harvard.edu>
To:	"K.Prasad" <prasad@...ux.vnet.ibm.com>
cc:	Ingo Molnar <mingo@...e.hu>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Benjamin Herrenschmidt <benh@....ibm.com>,
	Frederic Weisbecker <fweisbec@...il.com>,
	Maneesh Soni <maneesh@...ibm.com>,
	Roland McGrath <roland@...hat.com>,
	Steven Rostedt <rostedt@...dmis.org>
Subject: Re: [Patch 01/11] Introducing generic hardware breakpoint handler
 interfaces

On Tue, 24 Mar 2009, K.Prasad wrote:

> > Isn't that exactly the check you need to implement?
> > 
> > 	addr >= TASK_SIZE && (addr + len) >= TASK_SIZE,
> > 
> > or perhaps better,
> > 
> > 	addr >= TASK_SIZE && (addr + len) >= addr.
> > 
> > In this case you _do_ know the length of the breakpoint.
> > 
> > Alan Stern
> >
> 
> Aren't we just checking if len is a positive number through the above
> checks? The validation checks in the patchset should take care of
> negative lengths. Or am I missing something?

Well, 0x60000000 is a positive number, and 0xd0000000 is >= TASK_SIZE.  
But their sum is 0x30000000, which lies in userspace.  In other words, 
you are missing the possibility that the addition might overflow and 
wrap around.

> I thought you wanted the code to check for an upper sane limit for addr
> in kernel-space, say something like this:
> 
> TASK_SIZE <= addr <= (Upper limit for Kernel Virtual Address)

No, the test should be

    TASK_SIZE <= addr <= addr + (len-1) <= (Upper limit for Kernel VA)

By the way, is TASK_SIZE the correct lower bound for kernel virtual
addresses on x86-64?

> When I referred to 'len' in my previous mail, it meant the length
> of the kernel virtual memory area (which can be used to find the upper
> bound).

Oh, sorry, I misunderstood.  Isn't that limit always 0xffffffff on 
x86-32?

Alan Stern

--
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/