linux-kernel - Re: [PATCH 2/2] tpm: Fix crash on tmprm release

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, 17 Jun 2021 07:44:58 +0200
From:   Vincent Whitchurch <vincent.whitchurch@...s.com>
To:     Jarkko Sakkinen <jarkko@...nel.org>
CC:     Peter Huewe <peterhuewe@....de>, Jason Gunthorpe <jgg@...pe.ca>,
        kernel <kernel@...s.com>,
        "linux-integrity@...r.kernel.org" <linux-integrity@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        <James.Bottomley@...senPartnership.com>, <l.sanfilippo@...bus.com>
Subject: Re: [PATCH 2/2] tpm: Fix crash on tmprm release

On Tue, Jun 15, 2021 at 03:18:48PM +0200, Jarkko Sakkinen wrote:
> On Tue, Jun 15, 2021 at 11:14:09AM +0200, Vincent Whitchurch wrote:
> > If the tpm_tis module is removed (or a system shutdown is triggered)
> > while the tpmrm device is use, the kernel crashes due to chip->ops being
> > NULL:
> > 
> >  # exec 3<>/dev/tpmrm0
> >  # rmmod tpm_tis
> >  # exit
> >  ==================================================================
> >  BUG: KASAN: null-ptr-deref in tpm_chip_start+0x2d/0x120 [tpm]
> >  Read of size 8 at addr 0000000000000060 by task sh/994
> > 
> >  Call Trace:
> >   kasan_report.cold.13+0x10f/0x111
> >   tpm_chip_start+0x2d/0x120 [tpm]
> >   tpm2_del_space+0x2c/0xa0 [tpm]
> >   tpmrm_release+0x3f/0x50 [tpm]
> >   __fput+0x110/0x3c0
> >   task_work_run+0x94/0xd0
> >   do_exit+0x683/0x13e0
> >   do_group_exit+0x8b/0x140
> >   do_syscall_64+0x3c/0x80
> >  ==================================================================
> > 
> > Fix this by making tpm2_del_space() use tpm_try_get_ops().  The latter
> > already includes the calls to tpm_chip_start() and tpm_chip_stop().
> 
> This lacks explanation why migrating to tpm_try_get_ops() fixes the
> issue. Saying that doing something fixes something is not good enough
> explanation. So, can you extend this paragraph just a bit in the next
> version?

Thank you for the review, but I see now that James already proposed a
more-or-less identical fix earlier, so I'll let that patch run its due
course:

 https://lore.kernel.org/linux-integrity/e7566e1e48f5be9dca034b4bfb67683b5d3cb88f.camel@HansenPartnership.com/
 https://lore.kernel.org/linux-integrity/327f4c87-e652-6cbe-c624-16a6edf0c31d@kunbus.com/