sourcediving

To learn you really you need to do it a while.  But we'll discuss some techniques that are helpful.

Starting:
  clone (always, if you want to study)
  git ls-files, skim, to see where code is

Finding where some feature is implemented, bottom up:
  usually the easiest way to find anything specific
  demo example: CPython, __add__
  grep, preferably git grep, or grep foo -- $(find . -name \*.c) etc. 
  string literals are great to grep for
    (from UI, or external APIs: commands, SQL queries, etc)
  for log messages etc., search for constant-looking substrings
  for short words, use grep -w, or \b, or quotes (for entire literal), etc.
  get more context with grep -10, -A10

finding stuff, top down:
  demo example: luajit
  find 'main', or appropriate for language
  have to guess to ignore boring stuff, drill down on interesting
  have to skim code; gets easier as you read more
    (and more in this language, or this codebase)

once you have a toehold in the code, navigating:
  grep -w
  grep 'def foo', or '^foo', '^\S.*foo', etc, to find just definitions
  better: ctags, other tools below (highly recommended if spending much time on one codebase)

once you're looking at code, understanding:
  VCS history
  git log --stat -p -- $file
  /^c in pager to go up/down a commit at a time
  git log -G
  git log --full-diff (I didn't mention this, but very handy; check 'git help log')

fancier tools for finding things:
 ctags, etags
 OpenGrok http://opengrok.github.io/OpenGrok/
 linux: livegrep, lxr (sadly lxr seems only half-working)
 chromium: https://code.google.com/p/chromium/codesearch
 github search (not very good; handy mainly when haven't cloned)