[Qemu-devel] [PATCH 0/2] Fainal TCG clean-up patches

[Evgeny Voevodin]

This set of patches moves rest global variables to tcg_ctx.
Also second patch introduces new TBContext for translation blocks
ans moves translation block globals there. We place tb_ctx inside
tcg_ctx and get noticable speed-up.


After this patchset was aplied,
I noticed ~4-5% speed-up of code generation.

Here is the test procedure:
1. Boot Linux Kernel 5 times.
2. For each iteration wait while "JIT cycles" is stable for ~10 seconds
3. Write down the "cycles/op"

Here are the results (tested on gcc-4.6):

Before clean-up:
min: 662.4
max: 696
avg: 672.28
standard deviation: ~17 ~= 3.5%

Average cycles/op = 672 +- 17


After clean-up:
min: 635
max: 650.5
avg: 640.14
standard deviation: ~8 ~= 1.6%

Average cycles/op = 640 +- 8

Evgeny Voevodin (2):
  TCG: Final globals clean-up
  TCG: Move translation block variables to new context inside tcg_ctx:
    tb_ctx

 cpu-exec.c              |   18 +++--
 include/exec/exec-all.h |   27 +++++---
 linux-user/main.c       |    6 +-
 tcg/tcg.c               |    2 +-
 tcg/tcg.h               |   16 ++++-
 translate-all.c         |  173 +++++++++++++++++++++++------------------------
 6 files changed, 130 insertions(+), 112 deletions(-)

-- 
1.7.9.5