I updated my patches after fixing a sever bug that caused fallout in lots of (apparently unrelated) atf tests by fooling ld.elf_so into insecure mode, thus not allowing LD_PRELOAD and breaking all the rump clients...
ftp.netbsd.org:/pub/NetBSD/misc/martin/posix_spawn/newfiles_20120125.tar.gz ftp.netbsd.org:/pub/NetBSD/misc/martin/posix_spawn/posix_spawn_20120125.diff.gz I have been unable to find any statistically valid performance differences, either between posix_spawn and vfork, nor between execv* on a patched vs. an unpatched kernel - it is all in the sub-prommille range within a few percent noise. A working alternative aproach to this patch is to use a vfork based emulation in userland only (FreeBSD is doing that). It is less kernel code, but to me sounds a bit hackish and/or fragile - maybe a matter of personal taste. I will ask core which of the two variants we want to go with. Martin