This finishes making the kernel compile without -O.
The "optimized" asm version of the function being inlined
(translate_bytes()) uses slow instructions. On a 486, assuming
everything is in the cache (unlikely), it is 21/15 times slower
than the dumb C version and 21/3 times slower than the best
possible bytewise method.
- use daemon() to daemonify ourselves
- the 'Usage' printf() was missing an argument
- remove declaration of rindex and #include <string.h> instead
bootparam.c:
- get rid of local declarations of YP functions and include headers
from /usr/include/rpcsvc instead.
for the kernel, but gcc provides an inline version of it if the
kernel is compiled with -O.
The inline memcmp() is OK for small compares and is better than
the dumb kernel bcmp() in all cases, but it has been hiding the
library memcmp() which is 4 times faster for large compares.
don't go away when the kernel is compiled with -O.
The functions are backed up by extern versions in cd9660_util.c,
but these versions are disabled by `#ifdef __notanymore__'. They
could have been enabled by using `#if defined(__notanymore__) ||
!defined(__OPTIMIZE__)' but then I would have had to check that
they still work. The correct way to handle all this is to replace
`extern inline' by `EXTERN_INLINE' and define `EXTERN_INLINE' as
`extern inline' in most modules and as empty in one module.
didn't work are somewhat bogusly optimized away before the constraint
is checked. We still expect constants passed to inline functions to
remain constant, but if the compiler ever decides that they aren't
constant then it will just generate slightly slower code instead of
an error.
Declare `cheat' as static. It was bogusly shared between the aha1742 and
ultrastor drivers.
Even static variables should have unique names so that they can be
debugged, but fixing them can wait.
different types of panics/inconsistencies with NFS clients.
Cleared PG_WANTED where appropriate.
Added checks for buffer busy in allocbuf and biodone.
Reviewed by: John Dyson
at install time. This will allow `install -C' to avoid replacing the
installed copy if the new copy is the same. `install -C' would actually
work right if `strip' is already installed, but the Makefile doesn't
know that and has to work for plain `install'.
Stripping should be done by default at link time, but complications
would still be required for installing `strip' in case the default is
overridden.
when it is moved to an NFS filesystem from from another filesystem and /bin/mv
failed to set the file ownership during the move.
I believe that this bug is present in STABLE but I have not tested it. The fix
would be the same in STABLE even though the code has changed quite considerably
in CURRENT.
of replacing it. This way you can point it to a site close to you
that carries many distfiles, and still let it go fetch from the
original site if the distfile is not there.
Original idea by: mmead@Glock.COM
level ourself. We failed for unreadable directories. E.g.,
`mtree -d -f /etc/mtree/BSD.usr.dist -p /usr' run by `nobody' was
confused after it couldn't descend into /usr/games/hide. It looked
for /usr/include and subsequent directories in /usr/games.
Don't search for `extra' files when the spec depth is less than the
fts level. The spec depth isn't incremented for leaf nodes because
that would give a NULL level pointer and make it inconvenient to go
back to the parent level. Leaf nodes are built for directories that
are empty in the spec. Since they are empty in the spec, all files
in them are extra. The search looked for files one spec level
too high, so for `mtree -d -f /etc/mtree/BSD.usr.dist -p /usr',
obj/sbin matched ./sbin and wasn't considered extra, so it was
descended into and lots of bogus extra things in it were found.
This was harmful for `mtree -U' (as reported in pr623) and worse
for `mtree -r'.
Use rmdir(), not unlink(), to remove `extra' directories. unlink()
succeeds for root but unlinking directories normally damages the
file system.
Report `fts_errno' instead of `errno' when the former applies.
ended that fork() uses to determine the time that the process
started when calculating the elapsed time. This prevents the
ac_etime field in the accounting record from getting set to -1
if the process exists for a VERY short period of time.
