linux-kernel - Re: [patch 16/17] Immediate Values

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:	Mon, 14 Apr 2008 20:12:19 -0400
From:	Mathieu Desnoyers <mathieu.desnoyers@...ymtl.ca>
To:	Rusty Russell <rusty@...tcorp.com.au>
Cc:	akpm@...ux-foundation.org, Ingo Molnar <mingo@...e.hu>,
	linux-kernel@...r.kernel.org, Andi Kleen <andi@...stfloor.org>,
	Adrian Bunk <bunk@...sta.de>,
	Alexey Dobriyan <adobriyan@...il.com>,
	Christoph Hellwig <hch@...radead.org>, akpm@...l.org
Subject: Re: [patch 16/17] Immediate Values - Documentation

* Rusty Russell (rusty@...tcorp.com.au) wrote:
> On Friday 11 April 2008 11:16:47 Mathieu Desnoyers wrote:
> > * Rusty Russell (rusty@...tcorp.com.au) wrote:
> > > On Thursday 10 April 2008 01:08:45 Mathieu Desnoyers wrote:
> > > > If you have to read the immediate values from a function declared as
> > > > __init or __exit, you should explicitly use _imv_read(), which will
> > > > fall back on a global variable read. Failing to do so will leave a
> > > > reference to the __init section after it is freed (it would generate a
> > > > modpost warning).
> > >
> > > That's a real usability wart.  Couldn't we skip these in the patching
> > > loop if required and revert so noone can make this mistake?
> >
> > Yeah, I know :(
> >
> > Well, only if we can find a way to detect the macro is put within a init
> > or exit section. Is there some assembly trickery that would permit us to
> > do that ?
> >
> > Otherwise, given the memory freed from the init section could be reused
> > later by the kernel, I don't see how we can detect the pointer leads to
> > a freed init section and, say, a module.
> 
> In theory although not in practice, since everyone vmallocs modules.  Let's 
> not rely on that tho.
> 
> How about we sweep the immediate table on init discard and remove/mark all the 
> init and exit references?
> 
> Cheers,
> Rusty.

I already posted a patch which nullifies the immediate values pointing
to init code after the init phase of the core kernel and the init phase
of modules, just before the init section is freed.

For the exit section, I could add some code which nullifies the
immediate values pointing to exit section for !CONFIG_MODULE_UNLOAD.
However, I would need to get the equivalent of init and init_size for
.exit too.

I wonder what would happen if someone declares a __exit function in a
builtin object, with an immediate value in it ? Is there a possibility
that it leaves a reference to code not even linked in ?

Mathieu


-- 
Mathieu Desnoyers
Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68
--
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/