The goal of this series is to add some support to easily report allocations statistics and thus be able to do some experiments about allocation. For example, playing a bit with this showed that most lists are quite short (can be as short as and avarage of 2.5 elements on some code) and that decreasing LIST_NODE_NR from 29 to 13 allow to win 5% when using make C=2 on the kernel (when all source files have already been compile by GCC), which is already quite interesting.. This series is available at: git://github.com/lucvoo/sparse.git alloc-stats based on commits (a merge): 263f8ce449097da4caacfac4e00144f626c6e6a7 (pre-llvm-fixes-v6) fc981fe285c37ee297e93ef1cc8725caac75f9b3 (fix-bitfield-init-v2 up to commit: bdd16ed276569a16b2362adb266048b91e617f32 Luc Van Oostenryck (5): add get_<allocator>_stats() add show_allocation_stats() add helper handle_simple_switch() teach sparse how to handle '-fmem-report' use -fmem-report to report allocation stats Makefile | 1 + allocate.c | 8 ++++++++ allocate.h | 13 ++++++++++++ lib.c | 26 ++++++++++++++++++++---- lib.h | 2 ++ sparse-llvm.c | 1 + sparse.1 | 4 ++++ sparse.c | 2 ++ stats.c | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ test-linearize.c | 2 ++ test-unssa.c | 1 + 11 files changed, 118 insertions(+), 4 deletions(-) create mode 100644 stats.c -- 2.12.0 -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
