Hi TBOOT list,

I got a lot of help from Charles, but I forgot to CC the list.

Here is the short version:  ERRORCODE is set on the *next* "tboot" after the boot when tboot fails.  In my case, the machine reliably hangs on POST after tboot fails, but that's not a problem.  After filming a second tboot, I found the error code 0xc0027861 which I decoded like this:

ERRORCODE = 1 1 00000 000000010 0 11110 000110 0001

Ok, so from left to right:
1 = Valid error
1 = Induced by external software
0..0 = Reserved
0...10 = Minor error code  (This is undocumented, only valid number is 0)
0 = Error is generated by AC module
11110 = Major error code (0x1E = PS index not defined)
000110 = Class code  (6 = Class launch control policy)
0001 = SINIT

So my guess is that TBOOT tries to use the default LCP, but when it doesn't find a Platform Supplier, it doesn't work.  I think this behavior is new in the SINIT v67.

This is for 3rd_gen_i5_i7-SINIT_67

Maybe the minor error code is the "unknown" Platform Supplier? I will set up an LCP as "Platform Owner", but maybe someone can shed some light on what actually happens in my case?

Could my BIOS be at fault for not properly setting up a default Platform Supplier LCP?


On Fri, Jul 26, 2013 at 7:24 PM, Charles Bushong <bushong1@gmail.com> wrote:
When the system reboots after GETSEC halts, it should show an ERRORCODE of some kind.  Can you find that?

TBOOT: AC module error : acm_type=0x#, progress=0x##, error=0x#


On Fri, Jul 26, 2013 at 1:16 PM, Alexander Kjeldaas <alexander.kjeldaas@gmail.com> wrote:

Hi tboot developers,

The newest SINIT code required a bios upgrade on my Q77M-D2H motherboard.

With the new bios, I am getting a hang either after MTRR registers are set, or after GETSEC[SENTER].

I slowed down the boot using a busy-wait in printk, and if the boot is slow enough, GETSEC[SENTER] is never written to the console.

However, with a slightly faster boot, the hang is after GETSEC[SENTER].

Here are some videos from the boot (faster and slower).
Any ideas on how to proceed?


Btw, if tboot had a "fallthrough" mode where nothing is done, but the previous boot-log is preserved.  That the logs from tboot hangs could potentially be salvaged.

See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
tboot-devel mailing list