This change switches c++ exception handling from sjlj to standard dwarf.
It's needed because clang for aarch64 doesn't support sjlj. It turns out
that libunwind had a bare-metal configuration that made this easy to do.
This change gets the new experimental cosmocc -mclang flag in a state of
working so well that it can now be used to build all of llamafile and it
goes 3x faster in terms of build latency, without trading away any perf.
The int_fast16_t and int_fast32_t types are now always defined as 32-bit
in the interest of having more abi consistency between cosmocc -mgcc and
-mclang mode.
If pthread_create() is linked into the binary, then the cosmo runtime
will create an independent dlmalloc arena for each core. Whenever the
malloc() function is used it will index `g_heaps[sched_getcpu() / 2]`
to find the arena with the greatest hyperthread / numa locality. This
may be configured via an environment variable. For example if you say
`export COSMOPOLITAN_HEAP_COUNT=1` then you can restore the old ways.
Your process may be configured to have anywhere between 1 - 128 heaps
We need this revision because it makes multithreaded C++ applications
faster. For example, an HTTP server I'm working on that makes extreme
use of the STL went from 16k to 2000k requests per second, after this
change was made. To understand why, try out the malloc_test benchmark
which calls malloc() + realloc() in a loop across many threads, which
sees a a 250x improvement in process clock time and 200x on wall time
The tradeoff is this adds ~25ns of latency to individual malloc calls
compared to MODE=tiny, once the cosmo runtime has transitioned into a
fully multi-threaded state. If you don't need malloc() to be scalable
then cosmo provides many options for you. For starters the heap count
variable above can be set to put the process back in single heap mode
plus you can go even faster still, if you include tinymalloc.inc like
many of the programs in tool/build/.. are already doing since that'll
shave tens of kb off your binary footprint too. Theres also MODE=tiny
which is configured to use just 1 plain old dlmalloc arena by default
Another tradeoff is we need more memory now (except in MODE=tiny), to
track the provenance of memory allocation. This is so allocations can
be freely shared across threads, and because OSes can reschedule code
to different CPUs at any time.
- Let OpenMP be usable via cosmocc
- Let libunwind be usable via cosmocc
- Make X86_HAVE(AVXVNNI) work correctly
- Avoid using MAP_GROWSDOWN on qemu-aarch64
- Introduce in6addr_any and in6addr_loopback
- Have thread stacks use MAP_GROWSDOWN by default
- Ask OpenMP to not use filesystem to manage threads
- Make NI_MAXHOST and NI_MAXSERV available w/o _GNU_SOURCE
We recently broke MODE=dbg support when we added C++ exception support.
This change adds the missing UBSAN interfaces, needed to get it working
again. Some of the ASAN checking in the SJLJ guts needed to be disabled
since I doubt anyone's combined the two features until now.
* third_party: Add libcxxabi
Added libcxxabi from LLVM 17.0.6
The library implements the Itanium C++ exception handling ABI.
* third_party/libcxxabi: Enable __cxa_thread_atexit
Enable `__cxa_thread_atexit` from libcxxabi.
`__cxa_thread_atexit_impl` is still implemented by the cosmo libc.
The original `__cxa_thread_atexit` has been removed.
* third_party/libcxx: Build with exceptions
Build libcxx with exceptions enabled.
- Removed `_LIBCPP_NO_EXCEPTIONS` from `__config`.
- Switched the exception implementation to `libcxxabi`. These two files
are taken from the same `libcxx` version as mentioned in `README.cosmo`.
- Removed `new_handler_fallback` in favor of `libcxxabi` implementation.
- Enable `-fexceptions` and `-frtti` for `libcxx`.
- Removed `THIRD_PARTY_LIBCXX` dependency from `libcxxabi` and
`libunwind`. These libraries do not use any runtime `libcxx` functions,
just headers.
* libc: Remove remaining redundant cxa functions
- `__cxa_pure_virtual` in `libcxxabi` is also a stub similar to the
existing one.
- `__cxa_guard_*` from `libcxxabi` is used instead of the ones from
Android.
Now there should be no more duplicate implementations.
`__cxa_thread_atexit_impl`, `__cxa_atexit`, and related supporting
functions, are still left to other libraries as in `libcxxabi`.
`libcxxabi` is also now added to `cosmopolitan.a` to make up for the
removed functions.
Affected in-tree libraries (`third_party/double-conversion`) have been
updated.
