CVS commit: src/sys/rump

["Antti Kantee" <pooka%netbsd.org@localhost>]

Module Name:    src
Committed By:   pooka
Date:           Sat Mar 15 15:15:27 UTC 2014

Modified Files:
        src/sys/rump: Makefile.rump
        src/sys/rump/include/machine: cpu.h
        src/sys/rump/include/rump: rumpuser.h
        src/sys/rump/librump/rumpkern: lwproc.c rump.c rump_private.h
            scheduler.c threads.c
        src/sys/rump/librump/rumpkern/arch/x86: rump_x86_cpu.c
Added Files:
        src/sys/rump/librump/rumpkern: rump_curlwp.h rump_curlwp___thread.h
            rump_curlwp_hypercall.h

Log Message:
Allow compile-time optimizations to curlwp.  This can have a pretty
staggering impact on performance.  When running sendto() in a loop,
the improvement is 200k more calls per second with an inlined __thread
curlwp as opposed to the default.  In other words, it shaves off hundreds
of CPU cycles per call (~20%).  Even just eliminating the x86_curlwp()
call in favor of an inline gives an improvement of 60k calls per second.


To generate a diff of this commit:
cvs rdiff -u -r1.91 -r1.92 src/sys/rump/Makefile.rump
cvs rdiff -u -r1.19 -r1.20 src/sys/rump/include/machine/cpu.h
cvs rdiff -u -r1.108 -r1.109 src/sys/rump/include/rump/rumpuser.h
cvs rdiff -u -r1.26 -r1.27 src/sys/rump/librump/rumpkern/lwproc.c
cvs rdiff -u -r1.289 -r1.290 src/sys/rump/librump/rumpkern/rump.c
cvs rdiff -u -r0 -r1.1 src/sys/rump/librump/rumpkern/rump_curlwp.h \
    src/sys/rump/librump/rumpkern/rump_curlwp___thread.h \
    src/sys/rump/librump/rumpkern/rump_curlwp_hypercall.h
cvs rdiff -u -r1.79 -r1.80 src/sys/rump/librump/rumpkern/rump_private.h
cvs rdiff -u -r1.35 -r1.36 src/sys/rump/librump/rumpkern/scheduler.c
cvs rdiff -u -r1.21 -r1.22 src/sys/rump/librump/rumpkern/threads.c
cvs rdiff -u -r1.1 -r1.2 \
    src/sys/rump/librump/rumpkern/arch/x86/rump_x86_cpu.c

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.