There are two ways to address it:
1) Call TPM2_Startup from the simulator-specific code.
2) Return back calling TPM2_Startup to the general case (this will also change it for the cases when for some reason the firmware didn't initialize TPM - with this fix trunksd will still start then).
WDYT?

Also, need to add 'open' to seccomp filters to cover the case when the check fails: https://chromium-review.googlesource.com/#/c/414952/.

I think the proper initialization should be called by the simulator startup code. 

It could also be tuned to do different things for normal and recovery mode boot, to make simulation environment even closer to a real chromes os device.

We can also update the CheckState() routine to print a warning but still return success if TPM is not initialized.
This would allow calling something like 'trunks_client --init_tpm' to initialize the TPM as firmware would. Otherwise, since trunksd is not running, trunks_client can't send any commands through it.

The following revision refers to this bug:
  https://chromium.googlesource.com/aosp/platform/system/tpm/+/b5f6248c030f2653d79ae3dec8f27aa8234d2216

commit b5f6248c030f2653d79ae3dec8f27aa8234d2216
Author: Andrey Pronin <apronin@chromium.org>
Date: Mon Nov 28 18:00:09 2016

trunks: add 'open' to seccomp filters

When an error is detected in TpmUtility::CheckState(), trunksd crashes
when minijail blocks 'open' syscall. Can be seen on the simulator, where
there is no firmware to initialize the TPM.

BUG= chromium:669109 
TEST=Start on the simulator, no "blocked syscall: open" in logs.

Change-Id: I0de70ec798b8da77da1d53b0297fe217ee935155
Reviewed-on: https://chromium-review.googlesource.com/414952
Commit-Ready: Andrey Pronin <apronin@chromium.org>
Tested-by: Andrey Pronin <apronin@chromium.org>
Reviewed-by: Darren Krahn <dkrahn@chromium.org>

[modify] https://crrev.com/b5f6248c030f2653d79ae3dec8f27aa8234d2216/trunks/trunksd-seccomp-mips.policy
[modify] https://crrev.com/b5f6248c030f2653d79ae3dec8f27aa8234d2216/trunks/trunksd-seccomp-arm64.policy
[modify] https://crrev.com/b5f6248c030f2653d79ae3dec8f27aa8234d2216/trunks/trunksd-seccomp-amd64.policy
[modify] https://crrev.com/b5f6248c030f2653d79ae3dec8f27aa8234d2216/trunks/trunksd-seccomp-x86.policy
[modify] https://crrev.com/b5f6248c030f2653d79ae3dec8f27aa8234d2216/trunks/trunksd-seccomp-arm.policy
[modify] https://crrev.com/b5f6248c030f2653d79ae3dec8f27aa8234d2216/trunks/trunksd-seccomp-x86_64.policy

Actually, the CL below also belongs to this issue: https://chromium-review.googlesource.com/#/c/415597/