bad syscall ?

[Patrick Welche <prlw1%cam.ac.uk@localhost>]

This morning, I updated the kernels of 2 -current/amd64 boxen to today's HEAD.

One no longer goes multiuser. ls, /rescue/ls, /rescue/mount, /rescue/ps
all claim "Bad syscall". cat is OK. Guessed libutil, but popping in new
modules and libutil didn't help. Isn't it the other way though, the binary
not finding a syscall in the new kernel?

The other did boot multiuser. Just tried an old rump script, and its
hijacked ls gave:

Core was generated by `ls'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x00007f7ff6cae8a5 in __compat_fstatvfs1 (fd=5, buf=0x7f7fffbff830, 
    flags=2) at /usr/src/lib/libc/compat/sys/compat_statvfs.c:99
99              int error = __fstatvfs190(fd, &sb, flags);


I can probably get the boxen going, but wonder how one can end up in
this state? (/usr/src/UPDATING?)


Cheers,

Patrick