Re: [coreboot] Asus AM1I-A

[Elisenda Cuadros]

Hi Gergely,

Thank you for your config. Now I have coreboot + SeaBIOS perfectly working.

Console speed is now "normal" :-)

These are the changes between the two configs:

-------------

diff config_original.txt config_Gergely.txt
23c23
< # CONFIG_USE_BLOBS is not set
---
  CONFIG_USE_BLOBS=y
101c101
< CONFIG_UART_FOR_CONSOLE=0
---
  CONFIG_UART_FOR_CONSOLE=1
150c150
< CONFIG_POST_DEVICE=y
---
  # CONFIG_POST_DEVICE is not set
210c210
< CONFIG_TTYS0_BASE=0x3f8
---
  CONFIG_TTYS0_BASE=0x2f8
428,429d427
< # CONFIG_PCI_OPTION_ROM_RUN_REALMODE is not set
< # CONFIG_PCI_OPTION_ROM_RUN_YABEL is not set
431d428
< # CONFIG_VGA_TEXT_FRAMEBUFFER is not set
556c553
< # Serial port base address = 0x3f8
---
  # Serial port base address = 0x2f8
581,583d577
< CONFIG_POST_DEVICE_NONE=y
< # CONFIG_POST_DEVICE_LPC is not set
< # CONFIG_POST_DEVICE_PCI_PCIE is not set

--------------

I made some tweaks to the configuration, added VGA output before 
payload, a nice bootsplash, etc..

Other considerations:

VGA Device PCI IDs are 1002,9830 (The default 1002,9836 does not get 
output).

Memory has to be located firstly on second slot (yellow DIMM_A2). If not 
you get AGESA_FATAL_ERROR. I only have one module.

Please, find attached a working config.txt and console.log. Coreboot 
revision is 4.7-294-g2db6fbc47b.

If you verify these changes are working for you is it possible to change 
the defaults in Kconfig?

Thank you very much.

Best regards,

-- Eli



On 11/02/18 00:07, Gergely Kiss wrote:
  Strange, the log you shared looked less verbose to me
than expected 
 but seems like I was wrong. Anyway, you might find an example on a 
 full debug log here [1].

 I have a small build script I use to build coreboot for my board, 
 please find it attached.

 The version currently running on my box is 4.6-2554-ga1b4c94.

 [1] https://www.coreboot.org/Board:asus/am1i-a#Bootlog

 On 10 February 2018 at 23:18, Elisenda Cuadros &lt;lists(a)e4l.es 
 <mailto:[e-mail]>> wrote:

     Thank you for your kind reply Gergely.

     I think the cable is not the problem, I have been using it for
     years and it's intact.

     I am using a usb dongle but not a cheap one. Tomorrow I will try
     with another cable and a real COM port. Just to be 100% sure.

     Console is fast at the very first time (just the "normal") but
     after two seconds it becomes extremely slow.

     Please, can you say me which is the git revision you are using?

     The console log I attached is from a SPEW debug level, not
     corresponding with the config.. I tried several configs today :-) .

     Is it possible to share your config with me?

     Best regards,

     Eli


     On 10/02/18 21:44, Gergely Kiss wrote:
      Hi Eli,

     I've been using Coreboot on my board for several weeks, it is
     serving as an HTPC running 24/7 and it's working perfectly stable
     which suggests the firmware should be free of bugs. It is likely
     that you are facing some configuration or hardware issue here.

     I didn't see any issues with the serial output while working with
     the board once the SuperIO chip started to work. Make sure the
     cable you use is intact and try to attach it to another machine.
     At the time I was working with my board, I used a Dell Port
     Replicator with a native COM port so I could use a standard
     null-modem cable and it was working flawlessly. In case you use a
     USB serial adapter, try replacing it or attach the serial cable
     directly to a COM port if you have one available.

     Also, please enable debug_level=Spew as it seems the console log
     you attached comes from a less verbose setting and therefore it's
     not as useful as it should be.

     Note that the VBIOS image is executed by SeaBIOS which means you
     won't see anything on your display until the payload is executed.

     If all else fails, you can still attach a POST debug module to
     the LPC header of the board [1] which can help a lot to find out
     where the boot process hangs.

     Feel free to contact me if you need some more help or
     information, I'm happy to assist!

     Regards,
     Gergely

     [1] https://www.youtube.com/watch?v=aGGqsWx3-1c
     <https://www.youtube.com/watch?v=aGGqsWx3-1c>

     On 10 February 2018 at 18:17, Elisenda Cuadros &lt;lists(a)e4l.es
     <mailto:[e-mail]>> wrote:

         Hello,

         I'm trying to use Coreboot in an Asus AM1I-A board recently
         ported by Gergely Kiss (thank you!).

         I am using the default config settings and added vga rom
         extracted with UEFITool from the vendor bios.

         After flashing and booting I get no output from vga.

         Serial console is extremely slow too (30 minutes to write the
         log)

         I attach the coreboot console log, config and cbfs.txt.

         Any hints?

         Thanks in advance.

         Best regards,

         -- Eli

Please upload board status Asus AM1I-A

[Paul Menzel]

Dear Gergely, dear Elisenda,


Welcome to coreboot and thank you for making and using the port for the
Asus AM1I-A.

Could one of you, please upload the current status to the board status
repository? See the file `README` in `util/board_status/`.

Please make sure, that you build from a non-dirty coreboot commit, that
means, the output from `git describe --dirty` should not contain the
word *dirty*.


Thanks,

Paul

Please upload board status Asus AM1I-A

[Elisenda Cuadros]

Dear Paul,

Sure, I will be happy to do it :-)

Best Regards,

Eli


On 12/02/2018 9:47, Paul Menzel wrote:
  Dear Gergely, dear Elisenda,


 Welcome to coreboot and thank you for making and using the port for the
 Asus AM1I-A.

 Could one of you, please upload the current status to the board status
 repository? See the file `README` in `util/board_status/`.

 Please make sure, that you build from a non-dirty coreboot commit, that
 means, the output from `git describe --dirty` should not contain the
 word *dirty*.


 Thanks,

 Paul

Please upload board status Asus AM1I-A

[Gergely Kiss]

Hi Paul, hi Eli,

looks like I missed this thread, sorry about the late answer.

Thanks, Eli for taking care of Paul's request!

Regards,
Gergely

On 12 February 2018 at 11:55, Elisenda Cuadros <lists(a)e4l.es> wrote:

  Dear Paul,

 Sure, I will be happy to do it :-)

 Best Regards,

 Eli

 On 12/02/2018 9:47, Paul Menzel wrote:

 Dear Gergely, dear Elisenda,


 Welcome to coreboot and thank you for making and using the port for the
 Asus AM1I-A.

 Could one of you, please upload the current status to the board status
 repository? See the file `README` in `util/board_status/`.

 Please make sure, that you build from a non-dirty coreboot commit, that
 means, the output from `git describe --dirty` should not contain the
 word *dirty*.


 Thanks,

 Paul




 --
 coreboot mailing list: coreboot(a)coreboot.org
 https://mail.coreboot.org/mailman/listinfo/coreboot

Please upload board status Asus AM1I-A

[Paul Menzel]

Dear Eli,


Am Montag, den 12.02.2018, 11:55 +0100 schrieb Elisenda Cuadros:

  Sure, I will be happy to do it :-)
Thank you very much for uploading the data to the board status
repository [1].

I noticed, that the serial console is enabled in your run, which causes
the boot time to be quite slow as all messages have to be sent over the
serial cable.

Could you please do another upload with the serial console disabled?
You can just use latest commit from the master branch. That data would
be nice to have to keep track of possible boot time regressions.

Also, RAM initialization took 1.7 seconds according to the time stamps.

```
   2:before ram initialization                         178,042 (174,580)
   3:after ram initialization                          1,878,897 (1,700,855)
```

If it is really that long – I didn’t see any log messages, so it’s
shouldn’t be the serial console –, could you do at least two reboots to
see if it’s due to an empty cache of the RAM data?


Kind regards,

Paul


[1] asus/am1i-a/4.7-294-gac05cbdea4/2018-02-12T15_31_32Z

Asus AM1I-A

[Gergely Kiss]

Hi,

taking a look at the diffs, I can't see any major difference.
CONFIG_USE_BLOBS is y by default so setting it to y is not really useful.
The only notable difference is that you used COM1 for console output
instead of COM2 (which is the port soldered on-board). However, this does
not seem to explain why you got "slow" output on your console. If the COM
port number and the base address is specified incorrectly then you'd see no
output, at all, so I'm unsure what the real issue was in your case and how
you managed to fix it.

As for the PCI IDs of the VGA adapter, it can be different depending on
what APU is installed. My installation uses the one that is set in Kconfig
by default.

The requirement to place the DIMM to slot DIMM_A2 (the slot closest to the
edge of the board) when only one module is present is noted in the vendor
manual on page 1-7.

About changing the defaults in the board's Kconfig, I'm not sure what you
mean, could you please clarify? I think changing the default VGA PCI ID
does not make sense as the VGA controller is integrated to the APU so each
APU model and revision might use different PIDs and I believe there's no
point preferring one ID over the others. The one that I have included is
for reference only which might need to be changed to match the APU
installed (my installation uses a Sempron 3850 APU).

Anyway, I'm happy to see that you could finally make coreboot work on your
board, it's good to known I'm not the only user of my port! :)

Regards,
Gergely

On 11 February 2018 at 12:26, Elisenda Cuadros <lists(a)e4l.es> wrote:

  Hi Gergely,

 Thank you for your config. Now I have coreboot + SeaBIOS perfectly working.

 Console speed is now "normal" :-)

 These are the changes between the two configs:

 -------------

 diff config_original.txt config_Gergely.txt
 23c23
 < # CONFIG_USE_BLOBS is not set
 ---
  CONFIG_USE_BLOBS=y
 101c101
 < CONFIG_UART_FOR_CONSOLE=0
 ---
  CONFIG_UART_FOR_CONSOLE=1
 150c150
 < CONFIG_POST_DEVICE=y
 ---
  # CONFIG_POST_DEVICE is not set
 210c210
 < CONFIG_TTYS0_BASE=0x3f8
 ---
  CONFIG_TTYS0_BASE=0x2f8
 428,429d427
 < # CONFIG_PCI_OPTION_ROM_RUN_REALMODE is not set
 < # CONFIG_PCI_OPTION_ROM_RUN_YABEL is not set
 431d428
 < # CONFIG_VGA_TEXT_FRAMEBUFFER is not set
 556c553
 < # Serial port base address = 0x3f8
 ---
  # Serial port base address = 0x2f8
581,583d577
 < CONFIG_POST_DEVICE_NONE=y
 < # CONFIG_POST_DEVICE_LPC is not set
 < # CONFIG_POST_DEVICE_PCI_PCIE is not set

 --------------

 I made some tweaks to the configuration, added VGA output before payload,
 a nice bootsplash, etc..

 Other considerations:

 VGA Device PCI IDs are 1002,9830 (The default 1002,9836 does not get
 output).

 Memory has to be located firstly on second slot (yellow DIMM_A2). If not
 you get AGESA_FATAL_ERROR. I only have one module.

 Please, find attached a working config.txt and console.log. Coreboot
 revision is 4.7-294-g2db6fbc47b.

 If you verify these changes are working for you is it possible to change
 the defaults in Kconfig?

 Thank you very much.

 Best regards,

 -- Eli



 On 11/02/18 00:07, Gergely Kiss wrote:

 Strange, the log you shared looked less verbose to me than expected but
 seems like I was wrong. Anyway, you might find an example on a full debug
 log here [1].

 I have a small build script I use to build coreboot for my board, please
 find it attached.

 The version currently running on my box is 4.6-2554-ga1b4c94.

 [1] https://www.coreboot.org/Board:asus/am1i-a#Bootlog

 On 10 February 2018 at 23:18, Elisenda Cuadros &lt;lists(a)e4l.es&gt; wrote:

  Thank you for your kind reply Gergely.

 I think the cable is not the problem, I have been using it for years and
 it's intact.

 I am using a usb dongle but not a cheap one. Tomorrow I will try with
 another cable and a real COM port. Just to be 100% sure.

 Console is fast at the very first time (just the "normal") but after two
 seconds it becomes extremely slow.

 Please, can you say me which is the git revision you are using?

 The console log I attached is from a SPEW debug level, not corresponding
 with the config.. I tried several configs today :-) .

 Is it possible to share your config with me?

 Best regards,

 Eli

 On 10/02/18 21:44, Gergely Kiss wrote:

 Hi Eli,

 I've been using Coreboot on my board for several weeks, it is serving as
 an HTPC running 24/7 and it's working perfectly stable which suggests the
 firmware should be free of bugs. It is likely that you are facing some
 configuration or hardware issue here.

 I didn't see any issues with the serial output while working with the
 board once the SuperIO chip started to work. Make sure the cable you use is
 intact and try to attach it to another machine. At the time I was working
 with my board, I used a Dell Port Replicator with a native COM port so I
 could use a standard null-modem cable and it was working flawlessly. In
 case you use a USB serial adapter, try replacing it or attach the serial
 cable directly to a COM port if you have one available.

 Also, please enable debug_level=Spew as it seems the console log you
 attached comes from a less verbose setting and therefore it's not as useful
 as it should be.

 Note that the VBIOS image is executed by SeaBIOS which means you won't
 see anything on your display until the payload is executed.

 If all else fails, you can still attach a POST debug module to the LPC
 header of the board [1] which can help a lot to find out where the boot
 process hangs.

 Feel free to contact me if you need some more help or information, I'm
 happy to assist!

 Regards,
 Gergely

 [1] https://www.youtube.com/watch?v=aGGqsWx3-1c

 On 10 February 2018 at 18:17, Elisenda Cuadros &lt;lists(a)e4l.es&gt; wrote:

  Hello,

 I'm trying to use Coreboot in an Asus AM1I-A board recently ported by
 Gergely Kiss (thank you!).

 I am using the default config settings and added vga rom extracted with
 UEFITool from the vendor bios.

 After flashing and booting I get no output from vga.

 Serial console is extremely slow too (30 minutes to write the log)

 I attach the coreboot console log, config and cbfs.txt.

 Any hints?

 Thanks in advance.

 Best regards,

 -- Eli