On Sun, Nov 22, 2015 at 06:54:47PM -0500, Stefan Berger wrote:
On 11/22/2015 12:20 PM, Kevin O'Connor wrote:
Hi Stefan,
As part of trying to better understand the TPM code, I found some changes to tcgbios.c were helpful. It's mostly code movement. I've only compile tested these changes, but I think they are safe.
I don't want to conflict with any changes you may have pending. So, let me know if this is a problem.
I do have other changes and extensions pending, but go ahead and apply it.
The series is on top of your latest patches. (FYI, I made a minor change to a comment in your patch 2.) I've also put this series (and your series) up at:
I applied them locally. From what i can see from tests I have done, results are still like before.
ACK to series.
Thanks. I ran across a few other things - I sent an email series and put the changes up on github.
There were a few other things I noticed, but that I have not tried changing:
- It's odd that "Booting from CD ROM device" is added twice for cdrom boots. Is that intentional?
- It's odd that is_tpm_present() is called from tpm_interrupt_handler32() , as that probes the hardware. I would think that if the has_working_tpm() check in tpm_interrupt_handler32() passes then is_tpm_present() would always pass.
- It seems the "TPM hardware interface" group of functions could be moved to tpm_drivers.c . Doing that seems like it could simpify the software/hardware interface as only a handful of functions would need to be exported (instead of the function table that tpm_drivers.c currently exports). Could tpm_drivers.c just export something like: tpmhw_probe(), tpmhw_set_timeouts(), tpmhw_transmit() ?
I will hold off on further changes until you can merge any features you have pending.
-Kevin