Re: [Xen-devel] [PATCH] vTPM: Fix Atmel timeout bug.

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Thu, 2014-10-30 at 15:48 +0200, Emil Condrea wrote:
> Of course we can use max, but I thought that it might be useful to
> have a prink to inform the user that the timeout was adjusted.
> In init_tpm_tis the default timeouts are set using:
> /* Set default timeouts */ tpm->timeout_a =
> MILLISECS(TIS_SHORT_TIMEOUT);//750*1000000UL tpm->timeout_b =
> MILLISECS(TIS_LONG_TIMEOUT);//2000*1000000UL tpm->timeout_c =
> MILLISECS(TIS_SHORT_TIMEOUT); tpm->timeout_d =
> MILLISECS(TIS_SHORT_TIMEOUT);
> 
> 
> 
> But in kernel fix they are set as 750*1000 instead of 750*1000000UL :
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/char/tpm/tpm_tis.c#n381
> So if we want to integrate kernel changes I think we should use
> MICROSECS(TIS_SHORT_TIMEOUT) which is 750000
> Also in kernel the default timeouts are initialized using
> msecs_to_jiffies which is different from MILLISECS
> macro.: 
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/char/tpm/tpm_tis.c#n548
> Is there a certain reason for not using msecs_to_jiffies ?

jiffies are a Linux specific concept which mini-os doesn't share.

Daniel, do you have any opinion on this patch?

It seems like the Linux fix is made only for the specifically broken
platform. That seems to make sense to me since presumably other systems
report short timeouts which they can indeed cope with. It's only Atmel
which brokenly reports something it cannot handle.

Ian.



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel