This patch set does alter the format of the trace files, and thus means we'll have to re-generate these. However, the space saved for sve trace files is significant, so I consider it worthwhile.
In addition, the new --dump option allows one to inspect the contents of the trace file. Version 2 adds a number of patches that clean up the whole reporting of errors thing. This allows us to give sensible output when things go awry. If using old trace files with the new risu, you get mismatch header after 1 checkpoints header mismatch detail (master : apprentice): magic: 000000fc vs 52695375 If for somehow you use different risu for master and apprentice on sockets, you get i/o error after 1 checkpoints Which is less helpful, but since the master has not written enough data to the socket, the apprentice notices and reports the i/o error before having a look at the header contents. r~ Richard Henderson (17): Use bool for tracing variables Unify master_fd and apprentice_fd to comm_fd Hoist trace file opening Adjust tracefile open for write Use EXIT_FAILURE, EXIT_SUCCESS Make some risu.c symbols static Add enum RisuOp Add enum RisuResult Unify i/o functions and use RisuResult Pass non-OK result back through siglongjmp Always write for --master Simplify syncing with master Split RES_MISMATCH for registers and memory Add magic and size to the trace header Compute reginfo_size based on the reginfo aarch64: Reorg sve reginfo to save space Add --dump option to inspect trace files risu.h | 88 +++++---- risu_reginfo_aarch64.h | 16 +- comms.c | 34 ++-- reginfo.c | 274 ++++++++++++++++----------- risu.c | 409 +++++++++++++++++++++++++---------------- risu_aarch64.c | 6 +- risu_arm.c | 6 +- risu_i386.c | 4 +- risu_m68k.c | 4 +- risu_ppc64.c | 4 +- risu_reginfo_aarch64.c | 192 ++++++++++--------- risu_reginfo_arm.c | 6 +- risu_reginfo_i386.c | 8 +- risu_reginfo_m68k.c | 6 +- risu_reginfo_ppc64.c | 6 +- 15 files changed, 635 insertions(+), 428 deletions(-) -- 2.20.1