vere64

[matthew-levan]

Cribbed from #794.

This PR introduces 64-bit Vere. c3_w is now either 32 or 64 bit, depending on if it was compiled with -DVERE64 (on by default, or off with zig build -Dvere32).

The following work still needs to be done:

• define final typedefs, update usages accordingly
• every appearance of _tmp in pkg/noun (excluding pkg/noun/jets) (~800 locations)
  • trace.{c,h}
  • ship.{c,h}
  • nock.h
  • version.h
  • imprison.{c,h}
  • jets.{c,h}
  • allocate.{c,h}
  • retrieve.{c,h}
  • *_tests.c
• every appearance of _tmp in pkg/noun/jets
• look at diff of pre and post VERE64 compilation with -Wconversion flag. Investigate every instance.
• run with llvm addr sanitizer, ubsan
  • found issues with asan and ubsan as of 61fc9de (lmdb addressed in vere: fix some undefined behavior #820)
• audit every jet
• migrations: bidirectional (use MSB of version tag to mark if vere64 or not? maybe also add metadata for pointer compression)
• integrate Joe's page based allocator
• new rsignal_setjmp/longjmp may have volatility problems (talk to Joe. maybe use atomics?)
• pkg/vere/noun-tests is completely broken
  • ensure noun tests make sense/are worth having
• every appearance of _tmp in any other subdirectory (~1000 locations)
• c3_w_clz, ctz, pc occurrences (we should change current occurrences to _tmp as we did with u3r_word, etc and go through them
• c3_etch_word, c3_sift_word occurrences (also change to _tmp)
• make sure everything works (C tests and hoon tests included) still for vere32
• revendor murmur3 with a size_t len
  • hoon: murmur3 revendor urbit#7186
  • mug: murmur3 revendor with size_t to support 64-bit #842
• test really big nouns in ames/fine/mesa
• the event log and other lmdb state should not need migration. we should double check lmdb.c, disk.c
• there's a lot of bloq size related things in the jets that could be changed to greater than 32 for the 32 bit case
• (mug .(eny 0, now 0)) equality test for fakezods booted in 32-bit and 64-bit modes
• mug-verified automated migration tests in ci

[mopfel-winrux]

can we remove this?

[dozreg-toplud]

When testing on Windows, I was getting strange results because mpz_init_set_ui takes unsigned long int, which is smaller than c3_w on Windows.

So you want to replace this with something more general, like:

void
mpz_init_set_word(mpz_t a_mp, c3_w b_w)
{
  if ( sizeof(unsigned long int) >= sizeof(b_w) ) {
    return mpz_init_set_ui(a_mp, b_w);
  }
  mpz_init2(a_mp, 1 << u3a_word_bits_log);
  mpz_import(a_mp, 1, -1, sizeof(c3_w), 0, 0, &b_w);
}

[dozreg-toplud]

Or maybe don't define a wrapper function for words and just compare to ULONG_MAX. Either way you should check interactions with external API like that.