[all-commits] [llvm/llvm-project] 112013: Flush bitcode incrementally for LTO output

[Jianzhou Zhao via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 11201315d5881a135faa5aa87f415ce03f99eb96
      https://github.com/llvm/llvm-project/commit/11201315d5881a135faa5aa87f415ce03f99eb96
  Author: Jianzhou Zhao <jianzhouzh at google.com>
  Date:   2020-09-17 (Thu, 17 Sep 2020)

  Changed paths:
    M lld/ELF/LTO.cpp
    M llvm/include/llvm/Bitcode/BitcodeWriter.h
    M llvm/include/llvm/Bitstream/BitstreamWriter.h
    M llvm/lib/Bitcode/Writer/BitcodeWriter.cpp

  Log Message:
  -----------
  Flush bitcode incrementally for LTO output

Bitcode writer does not flush buffer until the end by default. This is
fine to small bitcode files. When -flto,--plugin-opt=emit-llvm,-gmlt are
used, the final bitcode file is large, for example, >8G. Keeping all
data in memory consumes a lot of memory.

This change allows bitcode writer flush data to disk early when buffered
data size is above some threshold. This is only enabled when lld emits
LLVM bitcode.

One issue to address is backpatching bitcode: subblock length, function
body indexes, meta data indexes need to backfill. If buffer can be
flushed partially, we introduced raw_fd_stream that supports
read/seek/write, and enables backpatching bitcode flushed in disk.

Reviewed-by: tejohnson, MaskRay

Differential Revision: https://reviews.llvm.org/D86905