suspend/memory_after_suspend_auto give incorrect alert

HWE team treats the alarm from the current test case as a false alarm.
So, it needs to be revised.

The false alarm io log:
```
2c2
< total: 16537169920
---
> total: 16537174016
```

refer to https://bugs.launchpad.net/somerville/+bug/1895788/comments/4
comment from HWE

"
From kernel doc. https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/filesystems/proc.rst

MemTotal
              Total usable RAM (i.e. physical RAM minus a few reserved
              bits and the kernel binary code)

The MemTotal is the total usable memory to the kernel, not total physical RAM size, so it may change.
I can't find a real case about why suspend affects this value, but if this value does not keep changing every time(a memory leak?), we may ignore this failed.
"

"
If it's purpose is to check memory leak, then try suspend 3 times and check if the usable memory keeps decreasing is more reasonable.
"