[opal-prd] opal-prd is consuming 100% CPU

[bugproxy]

---Problem Description---
opal-prd fix for Ubuntu 18.04

 In some corner cases where opal-prd fails to access/read flash device (/dev/mtd0) we see opal-prd consuming 100% cpu. This mostly happens during boot time where opal-prd daemon tries to start even before mtd driver loaded. Its not easy to reproduce. It happens randomly.

---uname output---
Ubuntu 18.04

Machine Type = OpenPower System

Userspace tool common name: opal-prd

Below upstream patch will fix this issue.

commit cb16e55a234b91fd42112904cff15094fbae680d
Author: Vasant Hegde <email address hidden>
Date: Tue Apr 3 23:08:41 2018 +0530

    opal-prd: Insert powernv_flash module

    Explictly load powernv_flash module on BMC based system so that we are sure
    that flash device is created before starting opal-prd daemon.

-Vasant

[Launchpad Janitor (janitor)]

This bug was fixed in the package skiboot - 5.10~rc4-1ubuntu1

---------------
skiboot (5.10~rc4-1ubuntu1) bionic; urgency=medium

  * debian/patches/0001-opal-prd-Insert-powernv_flash-module.patch:
    cherry-pick from upstream to fix opal-prd spinning and consuming 100%
    CPU if it starts on boot before the mtd module has been loaded.
    LP: #1765460.

 -- Steve Langasek <email address hidden> Mon, 23 Apr 2018 09:50:43 -0700

[bugproxy (bugproxy)]

------- Comment From <email address hidden> 2018-04-25 02:49 EDT-------
Tested it with 5.10~rc4-1ubuntu1 package level, working fine.

root@ltc-boston125:~# 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 Wed 2018-04-25 00:40:44 CDT; 1min 38s ago
Docs: man:opal-prd(8)
Main PID: 3750 (opal-prd)
Tasks: 1 (limit: 19660)
CGroup: /system.slice/opal-prd.service
??3750 /usr/sbin/opal-prd

Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: PRDF:<<PRDF::initialize()
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: ATTN_SLOW:I>Service::enableAttns() enter
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: TARG:[TARG] I> getNextTarget: Using next node 1
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: TARG:[TARG] E> getNextTarget: Node 0 targets: first 0x702ede5efc04, current 0x702ede5f4ae8, last 0x702ede5f4ae8
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: TARG:[TARG] E> getNextTarget: Target not found
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: ATTN_SLOW:I>Service::enableAttns() exit
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: ATTN_SLOW:I><<ATTN_RT::enableAttns rc: 0
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: calling get_ipoll_events
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: enabling IPOLL events 0x5b90000000000000
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: FW: writing init message
root@ltc-boston125:~# opal-prd --version
opal-prd opal-prd-5.10~rc4
root@ltc-boston125:~# dpkg -l | grep -i opal
ii opal-prd 5.10~rc4-1ubuntu1 ppc64el OPAL Processor Recovery Diagnostics daemon
ii opal-utils 5.10~rc4-1ubuntu1 ppc64el OPAL firmware utilities
root@ltc-boston125:~#

root@ltc-boston125:~# 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 Wed 2018-04-25 00:40:44 CDT; 19min ago
Docs: man:opal-prd(8)
Main PID: 3750 (opal-prd)
Tasks: 1 (limit: 19660)
CGroup: /system.slice/opal-prd.service
??3750 /usr/sbin/opal-prd

Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: PRDF:<<PRDF::initialize()
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: ATTN_SLOW:I>Service::enableAttns() enter
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: TARG:[TARG] I> getNextTarget: Using next node 1
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: TARG:[TARG] E> getNextTarget: Node 0 targets: first 0x702ede5efc04, current 0x702ede5f4ae8, last 0x702ede5f4ae8
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: TARG:[TARG] E> getNextTarget: Target not found
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: ATTN_SLOW:I>Service::enableAttns() exit
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: ATTN_SLOW:I><<ATTN_RT::enableAttns rc: 0
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: calling get_ipoll_events
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: HBRT: enabling IPOLL events 0x5b90000000000000
Apr 25 00:40:46 ltc-boston125 opal-prd[3750]: FW: writing init message
...