Android distribution of pm-qa test assets are in a read-only filesystem folder that causes tests to fail

[Basil Eljuse]

pm-qa test assets are by default distributed in android configuration under /system/xbin/pm-qa

Based on results we obtain by running this suite it is evident that /system/xbin/pm-qa is read-only and this causes a number of test steps to fail during execution and moreover the execution logs of the test runs are lost.

Also there are failures thrown by the scripts when it queries for user id and uses that info in the script
(./cpuidle_02.sh[35]: [: gid=1007(log): unexpected operator/operand)

on an ubuntu shell:
root@localhost:~# id -u
root@localhost:~# 0

on an android shell:
root@android:/root # id -u
uid=0(root) gid=1007(log)

Need to ensure that the scripts do handle this info appropriately

[Basil Eljuse (basil-eljuse)]

An example failure log can be seen below:
cpuidle_01.10/cpu4: checking 'time' exists...
     pass
cpuidle_01.11/cpu4: checking 'usage' exists...
     pass
./cpuidle_02.sh
./cpuidle_02.sh[35]: [: gid=1007(log): unexpected operator/operand
cpuidle_02.0: checking cpuidle program runs successfully (120 secs)...
     jiffies are : 10000 usecs
found 5 cpu(s)
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us

Test duration exceeded
fail
./cpuidle_03.sh
./cpuidle_03.sh[35]: [: gid=1007(log): unexpected operator/operand
./cpuidle_03.sh[55]: [: gid=1007(log): unexpected operator/operand
cpuidle_03.0/cpu0: skipping cpu0...
     skip
[ 515.150257] [sched_delayed] process 3454 (cpuidle_killer) no longer
affine to cpu1
[ 515.157914] CPU1: shutdown
cpuidle_03.0/cpu1: checking cpuidle program runs successfully (120
secs)... jiffies are : 10000 usecs
found 5 cpu(s)
cpu1 is offline, ignoring
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
Test duration exceeded
fail
[ 640.264465] arm_big_little: CPU 3 initialized
[ 640.327904] CPU2: shutdown
cpuidle_03.0/cpu2: checking cpuidle program runs successfully (120
secs)... jiffies are : 10000 usecs
found 5 cpu(s)
cpu1 is offline, ignoring
cpu2 is offline, ignoring
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
Test duration exceeded
fail

[sanjay singh rawat (sanjay-rawat)]

Now the logs will be like this:

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
### cpufreq_01:
### test the cpufreq framework is available for frequency
### https://wiki.linaro.org/WorkingGroups/PowerManagement/Doc/QA/Scripts#cpufreq_01
###
cpufreq_01.0/cpu0: checking 'scaling_available_frequencies' exists... Ok
cpufreq_01.1/cpu0: checking 'scaling_cur_freq' exists... Ok
cpufreq_01.2/cpu0: checking 'scaling_setspeed' exists... Ok
cpufreq_01.0/cpu1: checking 'scaling_available_frequencies' exists... Ok
cpufreq_01.1/cpu1: checking 'scaling_cur_freq' exists... Ok
cpufreq_01.2/cpu1: checking 'scaling_setspeed' exists... Ok
cpufreq_01.0/cpu2: checking 'scaling_available_frequencies' exists... Ok
cpufreq_01.1/cpu2: checking 'scaling_cur_freq' exists... Ok
cpufreq_01.2/cpu2: checking 'scaling_setspeed' exists... Ok
cpufreq_01.0/cpu3: checking 'scaling_available_frequencies' exists... Ok
cpufreq_01.1/cpu3: checking 'scaling_cur_freq' exists... Ok
cpufreq_01.2/cpu3: checking 'scaling_setspeed' exists... Ok
-------- total = 12
-------- pass = 12

###
### cpuhotplug_01:
### test the sysfs files are present
### https://wiki.linaro.org/WorkingGroups/PowerManagement/Doc/QA/Scripts#cpuhotplug_01
###
cpuhotplug_01.0: checking 'online' exists... Ok
cpuhotplug_01.1: checking 'offline' exists... Ok
cpuhotplug_01.2: checking 'possible' exists... Ok
cpuhotplug_01.3: checking 'present' exists... Ok
cpuhotplug_01.0/cpu0: checking 'online' exists... Err
cpuhotplug_01.0/cpu1: checking 'online' exists... Ok
cpuhotplug_01.0/cpu2: checking 'online' exists... Ok
cpuhotplug_01.0/cpu3: checking 'online' exists... Ok
-------- total = 8
-------- pass = 7
-------- fail = 1
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

- As you noticed, "fail" status case will appear only when there is a failure

[sanjay singh rawat (sanjay-rawat)]

fyi. Dependency on change in lava script(change posted, in merge progress)

file: pm-qa.sh
change: /system/xbin/pm-qa -> /data/benchmark/pm-qa