Bug 10110

Summary: parselogs/intel-core-i7-64 on cherryhill /irq stat interface fatal error/failed to enable keyboard
Product: [Build System, Metadata & Runtime] BSPs Reporter: David Rodriguez <david.israelx.rodriguez.castellanos>
Component: bsps-meta-intelAssignee: California Sullivan <california.l.sullivan>
Status: RESOLVED FIXED QA Contact: Francisco Pedraza <francisco.j.pedraza.gonzalez>
Severity: normal    
Priority: Medium CC: davidx.lopez.barriba, sgw
Version: 2.2   
Target Milestone: 2.3   
Hardware: x86   
OS: Multiple   
Whiteboard:
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: Don't know
Attachments:
Description Flags
RESULTS-LSB-INTELI7-64-CHERRYHILL
none
LOG-SCREEN-TEST
none
sato-sdk-corei7_64_install
none
Logfile none

Description David Rodriguez 2016-08-10 20:35:37 UTC
Created attachment 3346 [details]
RESULTS-LSB-INTELI7-64-CHERRYHILL

STATUS: FAILED
BUILD: meta-intel-2.2_M2: a3cad8f0e0db9b2ad62ce0f2f1b0234cb9192f56
ENVIRONMENT: corei7 on Cherryhill

STEPS TO REPRODUCE: 
run auto-test intel-corei7-64
 
EXPECTED OUTCOME: 
  
No errors or fails should be present in logs


Attached are logs, this is present in both sato and lsb images.
Comment 1 David Rodriguez 2016-08-10 20:37:15 UTC
Created attachment 3347 [details]
LOG-SCREEN-TEST
Comment 2 California Sullivan 2016-08-11 20:00:19 UTC
I am unable to reproduce these warnings on my Cherry Hill board.

Looking at your logs, it appears that both the keyboard and boot media is having issues. Are you using a USB hub? If so, could you retest without it?

If not, could I get more information on the hardware and peripherals setup?

Thanks.
Comment 3 David Rodriguez 2016-08-12 14:42:37 UTC
(In reply to comment #2)
> I am unable to reproduce these warnings on my Cherry Hill board.
> 
> Looking at your logs, it appears that both the keyboard and boot media is
> having issues. Are you using a USB hub? If so, could you retest without it?
> 
> If not, could I get more information on the hardware and peripherals setup?
> 
> Thanks.

Hardware spec.

system serial: GPEMDP4506003
PCB: BRASSWELL RVP CRB FAB2
      [MB/DT] 25X27 TYPE3 SoC
MB SERIAL: SPBS43900476
           8006254 O-01 Rev. A

Other Tags:
SAS GK POLAND
OTC BSW eNB C0 SOC 25X27-F6XF
2R4382279M00730

SAS GK POLAND
OTC BSW eNB RVP MPI FAB2 BRD
SPB43900476

Tested on two connections set up:
1)Keboard and mouse connected to USB2 P0 and USB2 P1 respectively.
2)Keboard and mouse connected to USB3 P1 and USB2 P0 respectively.

Boot device (when is used) is connected to USB3 P1.

(note: USB3 P0 is disable, nothing connected to that port)

Tested with both booting from a installed system on a local 80G SSD Hard drive
and booting from a USB live image. Same results. Mouse an keyboard also works well on desktop.
Comment 4 California Sullivan 2016-09-29 21:07:37 UTC
[    6.860846] ata2.00: irq_stat 0x08000000, interface fatal error
[    6.878175] ata2: SError: { UnrecovData 10B8B BadCRC }

From the research I have done, it appears that these errors are usually caused by bad connections to the storage device.

I think the next steps would be using different cables to the boot device and if that does not work, using a SATA cable instead of USB.
Comment 5 California Sullivan 2016-10-06 21:24:19 UTC
Let me know if the results change with the above setup changes.
Comment 6 Stephen K Jolley 2016-10-24 18:03:59 UTC
Benjamin,

Since David is no longer around, can you assign this bug to be check by someone in your group.

Thanks,

Stephen
Comment 7 Juan Ramos 2016-10-28 22:09:02 UTC
Created attachment 3515 [details]
sato-sdk-corei7_64_install
Comment 8 Juan Ramos 2016-10-28 22:12:55 UTC
(In reply to comment #7)
> Created attachment 3515 [details]
> sato-sdk-corei7_64_install

Tested in:
dc8508f609974cc99606b9042bfa7f870ce80228

Tested booting from a installed system, no hub connected, mouse and keyboard present and functional in desktop.
Comment 9 Stephen K Jolley 2016-11-14 18:00:07 UTC
Cal,

It appears this has been reviewed.  Can you see if you still need this in NEEDINFO?  If so, as Juan for the information.
Comment 10 California Sullivan 2017-02-08 21:07:51 UTC
Closing this. I was able to get my hands on the same board and could not reproduce this error. There is a good chance it is related to the issues reported on bug 10614 as well. For future issues with the Braswell RVP, lets contain the discussion there.
Comment 11 David Lopez Barriba 2017-04-11 20:56:39 UTC
I found this error again in

eff56e4f0d59b1d965a68e4f009b7f07717b7edd

Build Configuration:
BB_VERSION        = "1.33.3"
BUILD_SYS         = "x86_64-linux"
NATIVELSBSTRING   = "universal-4.9"
TARGET_SYS        = "x86_64-poky-linux"
MACHINE           = "intel-corei7-64"
DISTRO            = "poky"
DISTRO_VERSION    = "2.2+snapshot-20170411"
TUNE_FEATURES     = "m64 corei7"
TARGET_FPU        = ""
meta              
meta-poky         
meta-yocto-bsp    = "HEAD:eff56e4f0d59b1d965a68e4f009b7f07717b7edd"
meta-intel        = "master:60c270394755c36d3cf2d1df61b3795521e1caa0"

Check the logfile
Comment 12 David Lopez Barriba 2017-04-11 20:58:06 UTC
Created attachment 3713 [details]
Logfile
Comment 13 Stephen K Jolley 2017-04-13 15:44:49 UTC
Cal,

Please work with David and insure they have a reliable board.

Thanks,

Stephen
Comment 14 California Sullivan 2017-04-13 23:48:58 UTC
I was able to reproduce this error message:

[    2.388186] atkbd serio0: Failed to enable keyboard on isa0060/serio0

I will take a look and see if I can find an obvious cause.
Comment 15 California Sullivan 2017-04-14 00:06:45 UTC
The error message is caused by the kernel driver at drivers/keyboard/atkbd.c. This driver is only for AT and PS/2 keyboards. Without **any** keyboard plugged in, I still got the error message, and I noticed this in the dmesg output:

AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input5

For this reason I believe that the error occurs because the board does in fact have a PS/2 keyboard connector, but we have nothing plugged in, so it cannot enable a keyboard on that port.

I believe this can be safely whitelisted, but I will try to find a PS/2 keyboard to test before submitting a patch.
Comment 16 California Sullivan 2017-04-21 22:20:32 UTC
Looks like the error persists even with a PS/2 keyboard plugged in. The keyboard worked fine, and the board does not have a separate AT keyboard connector, so it can be safely whitelisted.

Patch sent to mailing list.
Comment 18 David Lopez Barriba 2017-06-26 22:39:39 UTC
I found this in 2.4 M1: rc1-54d18a7efe307b9fb45d64156d53b886abc3792b
corei7-64_WIC_CherryHill