linux-kernel - Re: [PATCH v6] tpm: fix a race condition in tpm2_unseal

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 30 Aug 2016 12:07:01 -0600
From:   Jason Gunthorpe <jgunthorpe@...idianresearch.com>
To:     Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>
Cc:     Peter Huewe <peterhuewe@....de>, stable@...r.kernel.org,
        Marcel Selhorst <tpmdd@...horst.net>,
        "moderated list:TPM DEVICE DRIVER" 
        <tpmdd-devel@...ts.sourceforge.net>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v6] tpm: fix a race condition in tpm2_unseal_trusted()

On Tue, Aug 30, 2016 at 01:51:06AM +0300, Jarkko Sakkinen wrote:
> Unseal and load operations should be done as an atomic operation. This
> commit introduces unlocked tpm_transmit() so that tpm2_unseal_trusted()
> can do the locking by itself.
> 
> v2: Introduced an unlocked unseal operation instead of changing locking
>     strategy in order to make less intrusive bug fix and thus more
>     backportable.
> 
> v3: Have also separate __tpm_transmit() that takes 'flags' in order to
>     better localize the bug fix and make it easier to backport.
> 
> v4: Cleaned up the control flow in tpm2_unseal_trusted. Added the
>     missing 'Fixes' line.
> 
> v5: Previous versions did not set the enum to any value, so it was set
>     to zero. Changed enum to TPM_TRANSMIT_UNLOCKED because it is
>     guaranteed to catch such situation. With change the common case
>     adds so little clutter that the wrappers do not make sense anymore.
> 
> v6: Added flags parameter to internal TPM command wrappers to have
>     an expected default behavior and documented them.

Yeah, this is nice.

BTW, do not put the vX: stuff into the commit message, and it would be
best to include it after the diffstat the patch email

Reviewed-by: Jason Gunthorpe <jgunthorpe@...idianresearch.com>

Jason