procfs: add parsing of sys/vm

[TheMeier]

Read data from /proc/sys/vm entries
Closes: #157

Theres still work to be done for lowmem_reserve_ratio, actually i don't understand the format yet fully especially when dealing with NUMA zones.

[pgier]

Looks good, just a couple minor nits.
@discordianfish @mdlayher What do you think regarding the use of reflect here?

[pgier]

@TheMeier PR #178 has been merged, so you can take a look at that one for avoiding the use of reflect.

[TheMeier]

@pgier i changed the MR using the same pattern. I used your util ValueParser.Pint64 although I don't use pointers here, I wonder if that should be changed also.

Also simply unpacking/re-packing the fixtures.ttar with the provided makefile leads to re-ordering.

[pgier]

The fixtures file ordering should be fixed by #185
Edit: if you temporarily change the base branch of the PR and then change it back to master, it should update the diff to remove the bad fixtures stuff.

[pgier]

LGTM. Although after thinking about possible nil pointers similar to #187, I'm wondering if we should be using pointer types for the int64 values. @TheMeier what are the chances of some of the sys/vm files containing non-integer values and causing nil pointer dereference? @mdlayher wdyt?

[TheMeier]

@pgier hm fair question the documentation at https://www.kernel.org/doc/Documentation/sysctl/vm.txt doesn't really say anything about data format.
Looking around a bit in the kernel source i found a bunch of the settings defined as CTL_INT https://github.com/torvalds/linux/blob/f1c2f8857c5aa6c92aa903bc06437503422e5925/kernel/sysctl_binary.c#L147-L174 which is a define to bin_intvec which returns ssize_t so those should be fine.
Some others i randomly checked are defined as int ot u64.
My intuition is it should be quite safe, but to really answer the question someone (me?) should check the definition of each setting in the kernel sources.

[TheMeier]

LGTM. Although after thinking about possible nil pointers similar to #187, I'm wondering if we should be using pointer types for the int64 values. @TheMeier what are the chances of some of the sys/vm files containing non-integer values and causing nil pointer dereference? @mdlayher wdyt?

Actually I just realized I should have used pointers here. In case the valueparser comes back with nil the current implementation will lead to:

panic: runtime error: invalid memory address or nil pointer dereference [recovered]
        panic: runtime error: invalid memory address or nil pointer dereference

Will open an issue and patch...