| Age | Commit message (Collapse) | Author |
|---|
|
is created.
|
|
|
|
The inheritace is implemented by setting the default vnodeop to a
bypass op that repeats the operation on the spec/fifo vnodeop vector.
The overhead of one extra indirect function call is worth the cleanup
and improved correctness.
This actually solves a few bugs where some vnode ops were missing from
some vectors (like kqfilter or revoke). (and even more on the ubc
branch).
Inspired by the same thing done in FreeBSD.
|
|
|
|
|
|
the user addresses directly from the kernel. copyin is faster and can
correctly deal properly with mappings that uvm_useracc thinks are
correct but will fault anyway (to figure out how to generate such
mappings is left as en excercise for the reader).
|
|
the user addresses directly from the kernel. copyin is faster and can
correctly deal properly with mappings that uvm_useracc thinks are
correct but will fault anyway (to figure out how to generate such
mappings is left as en excercise for the reader).
Blatantly stolen from art@'s similar fix to sparc.
|
|
newsyslog.conf to reflect that. More people seem to prefer that,
and it can be easily changed if you like it another way.
millert, fgs and others verbally agree
|
|
|
|
utility; form DJ Gregor via pr2851
|
|
|
|
|
|
there is something other than a dir in the way. Closes PR 2935.
|
|
because HostbasedAuthentication might be enabled based on the
target host and ssh-keysign(8) does not know the remote hostname
and not trust ssh(1) about the hostname, so we add a new option
EnableSSHKeysign; ok djm@, report from zierke@informatik.uni-hamburg.de
|
|
|
|
pr2962
|
|
the short write cases.
|
|
PR 2902, Francesco Toscan, Thanks!
|
|
samba wants to create usernames that end in '$'. From Paul Chakravarti
|
|
PR 2891, jlouis at diku dk, Thanks!
|
|
|
|
|
|
|
|
|
|
add a missing pte flush in page_remove(),
removed a few dead debugs and ifdefs.
use local vars inestead of long -> chains./
miod@ ok
|
|
|
|
We don't support the bridge chip on this yet...
|
|
|
|
|
|
inspired by mpech@ pointing out a missing free(buf). so handle the
it-is-an-interface case where we don't need the buf before allocating memory.
|
|
|
|
|
|
rotating a file 512KB instead of 512 bytes, as was intended. Change
things to deal in bytes internally and drop the cutoff to 256 bytes.
|
|
|
|
when altqd is not running.
report and fix by Chad Loder <cloder@loder.us>
|
|
the user addresses directly from the kernel. copyin is faster and can
correctly deal properly with mappings that uvm_useracc thinks are
correct but will fault anyway (to figure out how to generate such
mappings is left as en excercise for the reader).
|
|
In sys_sigreturn - don't uvm_useracc the context address and then copyin
it afterwards. The copyin deals better with finding if the user address is
correct and it's faster.
|
|
in libc so static linking works with -lpthreads. There is a
(linker?) problem that still shows up with static -pthread, though.
Solution to problem from freebsd.
|
|
|
|
logic. As result, MDSETS gone from everything but mac68k.
MDSETS now treated as a list of complete file names to add to THESETS,
allowing more MD flexibility, i.e. may or may not be .tgz files.
All traces of tar.gz removed. Not used for sets since 2.7 or so. As a
result of code simplification from not having to consider tar.gz, the
functions get_sets(), make_lastset(), and set_not_done() are
eliminated, along with the global SETS.
Build THESETS and DEFAULTSETS to contain full file names.
Always install sets in the order they appear in THESETS, keeping
siteXX last but without needing special logic. More predictable for
user if a lot of set additions and removals done before being
satisfied, as the sets are now installed in the same order they appear
in the selection list.
Remove ability to look for non-set .tgz/.tar.gz files. The job of the
install scripts is to install the standard sets in a simple and
rational way. Not to install general tar balls. The user can do that
post-install or simply '!' out to do it during install. On their head
be it.
If no sets are found in a specified source, list the filenames in
THESETS to show what is being looked for.
|
|
|
|
helps if the ramdac does not get initialized (idea from jason@)
- only register a shutdown hook for the frame buffers which need it, if this
is the console frame buffer. Otherwise this is just a waste of time.
|
|
after. This avoids a bunch of reloc warnings by working around a
deficiency in the a.out loader. miod@ OK
|
|
last commit
|
|
- mount point may be specified with trailing slash in all cases
(including fstab).
- mount doesn't look in fstab at "-o update". So update doesn't fails
for fs which isn't listed in fstab.
- 'mount -o update' only changes options which supplied via command line
and preserves options which currently in use.
- fix argv[] buffer overflow in mountfs/mangle.
- correct usage.
- Other improvements and fixes which I can't remember.
Thanks to everyone who test this patch.
miod@ ok.
|
|
|
|
|
|
sigreturn. sigblock has not been a syscall since 4.3BSD and using
SYSCALL directly seems like a bad idea in general.
Fixes a build problem on vax. OK miod@
|
|
this part of falling back to non-ecn on timeout slipped during
merge from KAME.
ok @art
report to bugs@ by Han Boetes and Otto Moerbeek
|
|
|
