OCC Reset command just hangs on OpenPower habanero system and occ's are entered in Disabled state.

[pridhiviraj]

Install a pnor image which is built from op-build master-next branch and try to do a OCC Reset on a OpenPower habanero system and command just hangs and occ's are entered into Disabled state on first attempt itself.

service opal-prd status

• opal-prd.service - OPAL PRD daemon
  Loaded: loaded (/lib/systemd/system/opal-prd.service; enabled; vendor preset: enabled)
  Active: active (running) since Thu 2016-10-27 10:11:39 EDT; 1min 0s ago
  Docs: man:opal-prd(8)
  Main PID: 4325 (opal-prd)
  Tasks: 1 (limit: 11059)
  CGroup: /system.slice/opal-prd.service
  `-4325 /usr/sbin/opal-prd --pnor /dev/mtd0

ipmitool fru print 43
Product Name : OpenPOWER Firmware
Product Version : open-power-habanero-e60dd7a
Product Extra : buildroot-211bd05
Product Extra : skiboot-5.3.7
Product Extra : hostboot-5a94e29
Product Extra : linux-4.4.24-openpower1-bf3db29
Product Extra : petitboot-v1.3.1-828e2bd
Product Extra : habanero-xml-5565b8f-3d88639
Product Extra : occ-bb092e9
Product Extra : hostboot-binaries-56532f5
Product Extra : capp-uco

ipmitool sdr elist | grep -i occ
OCC Active | 08h | ok | 210.0 | Device Enabled
opal-prd occ reset

On another terminal just check the occ's status again.
ipmitool sdr elist | grep -i occ
OCC Active | 08h | ok | 210.0 | Device Disabled

OCC's are entered into disabled state.

The above all tested on a fresh reboot it self.
opal-prd --version
opal-prd opal-prd-5.3.3

Tried same issue with upstream opal-prd from skiboot
./opal-prd --version
opal-prd opal-prd-5.4.0-rc2

~/skiboot/external/opal-prd# ./opal-prd occ reset

ipmitool sdr elist | grep -i occ
OCC Active | 08h | ok | 210.0 | Device Disabled

occ reset command just hangs and occ's are in disabled state.

[pridhiviraj]

syslog.txt

[pridhiviraj]

opal_msglog.txt

[dcrowell77]

Oct 27 10:15:33 pkvmhab012 opal-prd: HBRT: PNOR:>>RtPnor::readFromDevice: i_offset=0x0, i_procId=0 sec=11 size=0x20000 ecc=1
Oct 27 10:15:33 pkvmhab012 opal-prd: HBRT: PNOR:RtPnor::readFromDevice: removing ECC...
Oct 27 10:15:33 pkvmhab012 opal-prd: HBRT: PNOR:RtPnor::readFromDevice> Uncorrectable ECC error : chip=0,offset=0x0

We get an error trying to access the HBERRL section of PNOR. I don't see any error messages from the OPAL code so I don't have any guesses. I would've thought we'd see any issues during initial boot, but maybe not. I suggest you reflash the pnor.

[jazurin]

@pridhiviraj This looks similar to: #696
Does this reproduce every time?

[ghost]

I'm 99% sure this is the same as what the issue was in pull req #696 and open-power/skiboot#44 which is resolved in skiboot-5.4.0-rc3.

Please re-open if reproducible on rc3

[pridhiviraj]

Tested with skiboot 5.4.0-rc3 , occ reset works fine.