ASRock.com Homepage
Forum Home Forum Home > Technical Support > AMD Motherboards
  New Posts New Posts RSS Feed - X370 Taichi, Goes Blank; Code '00'
  FAQ FAQ  Forum Search Search  Events   Register Register  Login Login

X370 Taichi, Goes Blank; Code '00'
 Post Reply Post Reply Page  <12345>
Author
Message Reverse Sort Order
  Topic Search Topic Search  Topic Options Topic Options
Ivan_83 View Drop Down
Newbie
Newbie
Avatar

Joined: 22 May 2017
Status: Offline
Points: 85 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Ivan_83 Quote  Post ReplyReply Direct Link To This Post Posted: 25 May 2017 at 7:53pm
IMHO it is boot loop.
MOBO BIOS do soft reset, '00' leds is off for a very short time, then run startup sequence and set '00', then check some thing (I suspect MCE status), see that some thing wrong and do soft reset. Again and again.
'00' flashes = BIOS stick/fail at first boot step.

My first guess was that watchdog and may be some other devices without drivers not initialized as it should and may be watchdog fire reboot random time later.

But then I find MCE, and look like more reasonable that CPU reports about some error to OS/BIOS, OS do panic and software reboot. OS do not clean MCE status registers, I found comment on kernel code:
    /*
     * Clear machine check.  Don't do this for uncorrectable
     * errors so that the BIOS can see them.
     */
Then BIOS try init but don't clean MCE too, but check it.

I have no experience in UEFI + asm hacking (to many reverse engineering for me) to check this guess and have no time to start experiments.
All what I can is playing around MCE code in FreeBSD kernel, try to catch MCE event and decode it.
Back to Top
Atan87 View Drop Down
Newbie
Newbie


Joined: 01 May 2017
Status: Offline
Points: 25 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Atan87 Quote  Post ReplyReply Direct Link To This Post Posted: 25 May 2017 at 7:14pm
Mine does the same. That is not a bootloop, it's not even trying to boot.
Back to Top
Ivan_83 View Drop Down
Newbie
Newbie
Avatar

Joined: 22 May 2017
Status: Offline
Points: 85 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Ivan_83 Quote  Post ReplyReply Direct Link To This Post Posted: 25 May 2017 at 3:33pm
Yes, reset/power cycle is enough to reset CPU internal registers (MCE, watchdog and others) and fix '00' POST code and failure.

This is not error messages.
This is output from pciconf utility.
I dont understand about internal UEFI devices.

As you can see ryzen mostly supported by FreeBSD. CPU topology detect correctly, all devices that have drivers works OK.
I think that other OSes users from my links have same issue - not OS specific.

No, I dont change UEFI/BIOS file.


Edited by Ivan_83 - 25 May 2017 at 3:34pm
Back to Top
parsec View Drop Down
Moderator Group
Moderator Group
Avatar

Joined: 04 May 2015
Location: USA
Status: Offline
Points: 4996 		Post Options Post Options   Thanks (0) Thanks(0)   Quote parsec Quote  Post ReplyReply Direct Link To This Post Posted: 25 May 2017 at 10:48am
Originally posted by Ivan_83 Ivan_83 wrote:

Screen goes black and I get "00" on the Dr Debug display, that happen when the PC is running after a while.
System can work up to 2-6 day before that happen.
I unable to reproduce this by any actions in OS/software.

As I remember, '00' happen most times on video playback.
Last time I was working with xfreerdp with remote computer. No sound, no music, no video playback.

May this not connected, but 1-2 days before boot loop code '00', video playback with tearing.
Last time was no tearing on video playback, may be because I set Vsoc = 1.0V and SOC load line calibration = 3.

On my first link another taichi user get '00' on windows setup, music playback and playing game.

It may be quit difficult to RMA: 14 days are left and local reseller may take mobo for tests and because '00' code difficult to reproduce it can deny RMA or I will wait for long time for approve.
 

========
I suggest that boot loop instead reboot happen because some internal CPU registers or devices (inside CPU) contain incorrect values.
(May be some one forget set msr registers to protect some internal/system registers/mem regions during bios init, then OS try to use it - some CPU critical data may be damaged, then system try start but bios initial code does not init this part at all or fail to init.)

May be something wrong with watchdog inside CPU.
Is it possible that BIOS programmers forget init watchdog on BIOS start, it initialized by some random crap and after 2-6 day it fire reboot?
(Me and man with linux have no driver that reinit watchdog. User from first link and topic starter may be not setup all drivers for all devices and windows dont re init watch dog to. One of my theory.)

I try load watch dog driver but get:
amdsbwd0: <AMD FCH Rev 41h+ Watchdog Timer> at iomem 0xfed80b00-0xfed80b03,0xfed80b04-0xfed80b07 on isa0
amdsbwd0: watchdog hardware is disabled
device_attach: amdsbwd0 attach returned 6

I cant patch driver because I have no BIOS and Kernel Developer's Guide for AMD 17h, I request it and waiting: https://community.amd.com/thread/213584

Also I have no drivers for:
none0@pci0:0:0:2:    class=0x080600 card=0x14511022 chip=0x14511022 rev=0x00 hdr=0x00
    vendor     = 'Advanced Micro Devices, Inc. [AMD]'
    class      = base peripheral
    subclass   = IOMMU
    cap 0f[40] = unknown
    cap 05[64] = MSI supports 4 messages, 64 bit
    cap 08[74] = HT MSI fixed address window enabled at 0xfee00000
none1@pci0:0:20:0:    class=0x0c0500 card=0xffff1849 chip=0x790b1022 rev=0x59 hdr=0x00
    vendor     = 'Advanced Micro Devices, Inc. [AMD]'
    device     = 'FCH SMBus Controller'
    class      = serial bus
    subclass   = SMBus
none2@pci0:17:0:0:    class=0x130000 card=0x145a1022 chip=0x145a1022 rev=0x00 hdr=0x00
    vendor     = 'Advanced Micro Devices, Inc. [AMD]'
    cap 09[48] = vendor (length 8)
    cap 01[50] = powerspec 3  supports D0 D3  current D0
    cap 10[64] = PCI-Express 2 endpoint max data 256(256) RO NS
                 link x16(x16) speed 8.0(8.0) ASPM disabled(L0s/L1)
    ecap 000b[100] = Vendor 1 ID 1
    ecap 0019[270] = PCIe Sec 1 lane errors 0
none3@pci0:17:0:2:    class=0x108000 card=0x14561022 chip=0x14561022 rev=0x00 hdr=0x00
    vendor     = 'Advanced Micro Devices, Inc. [AMD]'
    class      = encrypt/decrypt
    bar   [18] = type Memory, range 32, base 0xfe600000, size 1048576, enabled
    bar   [24] = type Memory, range 32, base 0xfe700000, size 8192, enabled
    cap 09[48] = vendor (length 8)
    cap 01[50] = powerspec 3  supports D0 D3  current D0
    cap 10[64] = PCI-Express 2 endpoint max data 256(256) RO NS
                 link x16(x16) speed 8.0(8.0) ASPM disabled(L0s/L1)
    cap 05[a0] = MSI supports 2 messages, 64 bit
    cap 11[c0] = MSI-X supports 2 messages
                 Table in map 0x24[0x0], PBA in map 0x24[0x1000]
    ecap 000b[100] = Vendor 1 ID 1
none4@pci0:18:0:0:    class=0x130000 card=0x14551022 chip=0x14551022 rev=0x00 hdr=0x00
    vendor     = 'Advanced Micro Devices, Inc. [AMD]'
    cap 09[48] = vendor (length 8)
    cap 01[50] = powerspec 3  supports D0 D3  current D0
    cap 10[64] = PCI-Express 2 endpoint max data 256(256) RO NS
                 link x16(x16) speed 8.0(8.0) ASPM disabled(L0s/L1)
    ecap 000b[100] = Vendor 1 ID 1
    ecap 0019[270] = PCIe Sec 1 lane errors 0




So a simple PSU power cycle or press of the Reset button is enough to reset internal CPU registers? Or that somehow resets/initializes the watchdog? You said both of those things would fix the "00" POST code and POST failure.

What is producing all the error messages for the "drivers" you don't have? They seem to be internal UEFI/BIOS code? Wait, you are using FreeBSD 11 x64.

You have dismissed using FreeBSD 11 x64 as irrelevant, simply because someone reported a similar issue. AMD states only Windows 10 officially supports Ryzen. ASRock states only Windows 10 is supported for the X370 Taichi. Have you verified Ryzen support on FreeBSD 11 x64? If that is possible.

Have you modified the UEFI/BIOS file yourself?
Back to Top
Ivan_83 View Drop Down
Newbie
Newbie
Avatar

Joined: 22 May 2017
Status: Offline
Points: 85 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Ivan_83 Quote  Post ReplyReply Direct Link To This Post Posted: 25 May 2017 at 5:57am
Video with boot loop: https://www.dropbox.com/sh/uyj88uj1k9wa8cg/AAClalavS1r8bp_12PB9f-Tka/Crash%20Loop.mp4?dl=0
not mine, but I have same '00' flashes.
Back to Top
Atan87 View Drop Down
Newbie
Newbie


Joined: 01 May 2017
Status: Offline
Points: 25 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Atan87 Quote  Post ReplyReply Direct Link To This Post Posted: 25 May 2017 at 5:46am
No bootloop, just stays there doing nothing but showing code 00. I have to power down manually and power on again, which works everytime. Never have to clear cmos at this point. 
Back to Top
Ivan_83 View Drop Down
Newbie
Newbie
Avatar

Joined: 22 May 2017
Status: Offline
Points: 85 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Ivan_83 Quote  Post ReplyReply Direct Link To This Post Posted: 24 May 2017 at 3:36pm
Ok, I will disconnect both on next boot loop reboot.
Card reader: SEMA SFD-321F / T81UB


PS: can you transfer information about mce and boot sequence to BIOS team in asrock for check it?
Back to Top
wardog View Drop Down
Moderator Group
Moderator Group


Joined: 15 Jul 2015
Status: Offline
Points: 6447 		Post Options Post Options   Thanks (0) Thanks(0)   Quote wardog Quote  Post ReplyReply Direct Link To This Post Posted: 24 May 2017 at 11:25am
Originally posted by Ivan_83 Ivan_83 wrote:


ODD: ASUS DRW-1814BLT 1.14, it connected but not used (sata1, never try without it)
3,5 card reader USB 3.0 SEMA (Samsung) (never try without it)


Ivan, disconnect the above two and see what gives.

MCE's are hardware problems bad enough to stop/halt a computer.



The ODD is real old. 9+ years or so.

The card reader? Internal or external? Irrelevant, as most are temperamental at best and a PITA at worst.



Disconnect both the ODD and card reader and see if these MCEs reoccur.
Back to Top
Ivan_83 View Drop Down
Newbie
Newbie
Avatar

Joined: 22 May 2017
Status: Offline
Points: 85 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Ivan_83 Quote  Post ReplyReply Direct Link To This Post Posted: 24 May 2017 at 6:08am
Originally posted by Atan87 Atan87 wrote:

My taichi show's 00 code after crash due too high overclock. U sure your system is not crashing due to instability?


You got boot loop with code '00'?
If yes then +1 for my theory that BIOS does not clear MCE on soft reboot, but check it and do soft reboot...
If it true - this mean that Asrock should fix startup sequence to reset MCE and I should RMA CPU or  report to AMD and wait for microcode update.

Few peoples from another forum and different vendors mobos report about reboots on video playback, they have no overclock too.
Back to Top
Ivan_83 View Drop Down
Newbie
Newbie
Avatar

Joined: 22 May 2017
Status: Offline
Points: 85 		Post Options Post Options   Thanks (0) Thanks(0)   Quote Ivan_83 Quote  Post ReplyReply Direct Link To This Post Posted: 23 May 2017 at 7:08am
I will try to turn on logging for "machine check exception" = MCE, get and decode log.

Is it possible that BIOS does not clear mce registers on software reset/init sequence but check mce state and if there some error try to restart init sequence?


Edited by Ivan_83 - 23 May 2017 at 6:41pm
Back to Top
 Post Reply Post Reply Page  <12345>
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.04
Copyright ©2001-2021 Web Wiz Ltd.

This page was generated in 0.984 seconds.