Messages in this thread |  | | | Subject | Re: [PATCH 3/4][resubmit] TPM: rcu locking | | From | Mimi Zohar <> | | Date | Tue, 23 Sep 2008 17:27:34 -0400 | |
On Tue, 2008-09-23 at 15:01 -0600, Jonathan Corbet wrote:
> On Tue, 23 Sep 2008 17:19:33 -0300> Rajiv Andrade <srajiv@linux.vnet.ibm.com> wrote:
>
> > Protects tpm_chip_list when transversing it.
>
> It all looks like it should work (though it might make sense to include
> <linux/rcupdate.h> and <linux/rculist.h>. But I have to ask: do you
> really need the added complexity of RCU here? I suspect that TPM
> devices don't come and go very often...once most of us have ripped it
> out, we tend to leave it out...:)
>
> jon
True, the RCU locking is overkill for the current TPM code. It will be
used in the integrity-tpm-internal-interface patch to resolve the
locking issue that Christoph Helwig noted for the case when the TPM
driver is built as a module, which it shouldn't be, instead of being
built-in.
Mimi Zohar
|  |