elf: align_up usage susceptible to data loss

[iulianbarbu]

In response to #66.

• While using aling_up we're converting the function result
  to other data types. The conversions were double checked and
  comments were added to clarify why they are safe.

[alxiord]

I'm not sure we should discard n_align here. The man page linked doesn't specifically call out the alignment being 4 (it's 4 in the example snippet but not in the definition) -

       Note sections contain a series of notes (see the struct
       definitions below).  Each note is followed by the name field
       (whose length is defined in n_namesz) and then by the descriptor
       field (whose length is defined in n_descsz) and whose starting
       address has a 4 byte alignment. 

However, the alignment used throughout the file (and which should be consistent with the one in memory) is in this case n_align, read from the program header -

       p_align
              This member holds the value to which the segments are
              aligned in memory and in the file. 

[iulianbarbu]

I've done some research and my understanding is that p_align refers to segments alignment rather than name or desc notes fields alignment. There are also other resources that point to 4-byte alignment [1], [2], [3], [4]. [1] mentions about an alignment dependent on the arch reg size (4 bytes vs 8 bytes), so I think we lack this sort of logic as well. [3] is an old mail thread regarding how should notes be aligned on 64bit systems and the overall idea was that most applications (gdb, gcc and other binutils tools) and the linux kernel expect 4 bytes alignment even on 64bit systems.

I suggest having only the alignment adjustment in this PR, letting arch dependent implementation arise as a requirement, if needed (it is unlikely, low priority at the moment and it is also dependent on the gABI implementation of the build toolchain used).

[1] https://refspecs.linuxbase.org/elf/gabi4+/ch5.pheader.html#note_section - LSB details on ELF notes
[2] http://www.sco.com/developers/devspecs/gabi41.pdf - Generic ABI, page 81
[3] http://lkml.iu.edu/hypermail/linux/kernel/0611.1/0538.html - discussion around modifying note alignment on 64 bit platforms.
[4] linux kernel elf core dumper notesize - https://elixir.bootlin.com/linux/latest/source/fs/binfmt_elf.c#L1429

[alxiord]

nit:typo: s/loose/lose. Similar typos in the commit title (s/loos/loss) and in a few comments below

[iulianbarbu]

Done.

[andreeaflorescu]

The commnet also needs to be updated as it is not 4-bytes.

[iulianbarbu]

Done.