Re: [coreboot] AMD Family 15h and 16h: ASSERTION ERROR

[Mike Banon]

Regarding Lenovo G505S - also AMD family 15h : the very recent (April
08/09) build of coreboot
(for which I have submitted a board status) - gave me the assertion
errors as well:
...
BS: BS_DEV_INIT times (us): entry 0 run 718163 exit 0
Finalize devices...
Devices finalized

APIC 00: ** Enter AmdInitLate [00020004]
[1] ASSERTION ERROR: file
'src/vendorcode/amd/agesa/f15tn/Proc/Common/CommonReturns.c', line 187
[2] ASSERTION ERROR: file
'src/vendorcode/amd/agesa/f15tn/Proc/CPU/cpuGeneralServices.c', line
776
^^^ these two lines repeat 9 times, one after another.

FILE at [1] :
Line 181 - * ASSERT if this routine is called.
Line 182 - *
Line 183 - */
Line 184 - VOID
Line 185 - CommonAssert ( VOID )
Line 186 - {
Line 187 - ASSERT (FALSE);
Line 188 - }

FILE at [2] :
Line 775: InitApicIdCpuIdLo =
FamilyServices->CoreIdPositionInInitialApicId (FamilyServices,
StdHeader);
Line 776: ASSERT (InitApicIdCpuIdLo < CoreIdPositionMax);

^^^ What does it mean, and why is it failing?

After 18 assertion errors described above, there is a small log with
one more error at the end:

AmdInitLate() returned AGESA_SUCCESS
APIC 00: Heap in SystemMem (4) at 0x10000014
APIC 00: ** Exit  AmdInitLate [00020004]

APIC 00: ** Enter AmdS3Save [0002000b]
Manufacturer: 1c
SF: Unsupported EON ID 0x7016
SF: Unsupported manufacturer!
Manufacturer: 1c
SF: Unsupported EON ID 0x7016
SF: Unsupported manufacturer!
AmdS3Save() returned AGESA_SUCCESS
[3] ASSERTION ERROR: file 'src/drivers/amd/agesa/state_machine.c', line 309

FILE at [3] :
Line 308: status = amd_release_struct(&aip);
Line 309: ASSERT(status == AGESA_SUCCESS);

Could you please tell, why these errors are happening? Just in case,
my full coreboot console log is at
https://review.coreboot.org/cgit/board-status.git/tree/lenovo/g505s/4.7-662-...

On Mon, Apr 16, 2018 at 6:30 PM, Marshall Dawson
 wrote:
> Hey Paul,
>
> I don't have either of these two systems, and have only quickly scanned the
> Kabini AGESA source.
>
> But if I had to make a guess, I believe the GetDmiInfoMain() likely has a
> problem with its return value.  Look at how it returns Flag, but the
> function starts by setting Flag=TRUE.  Since Flag is an AGESA_STATUS and
> gets downgraded as necessary, it should've been initialized with
> AGESA_SUCCESS instead.  The same idea applies to GetType4Type7Info() too.
>
> It's not immediately apparent to me where the apu2 board picks up its
> buildOpts, but most systems are configured for BUILDOPT_REMOVE_CDIT TRUE.  I
> seem to recall CDIT causing an error to be returned if the option wasn't
> removed (as in imb-a180/buildOpts.c).
>
> Thanks,
> Marshall
>
> On Mon, Apr 16, 2018 at 1:50 AM, Paul Menzel
>  wrote:
> > Dear coreboot folks,
> >
> > Looking through the board status uploads, the AMD Family 15h and 16h
> > uploads contain lines with *ASSERTION ERROR*. For example:
> >
> > asrock/imb-a180/4.6-812-g4a1d450/2017-07-22T05_25_01Z/coreboot_console.txt:ASSERTION
> > ERROR: file 'src/vendorcode/amd/agesa/f16kb/Proc/CPU/cpuGeneralServices.c',
> > line 779
> >
> > pcengines/apu2/4.6-1904-ga547e40/2017-10-26T06_40_33Z/coreboot_console.txt:ASSERTION
> > ERROR: file 'src/northbridge/amd/pi/agesawrapper.c', line 279
> >
> > Has somebody analyzed those yet?
> >
> > Kind regards,
> >
> > Paul