| Age | Commit message (Collapse) | Author |
|---|
|
disk drivers are now smart enough to handle this correctly, as they
need to also handle pread(2)/pwrite(2).
ok deraadt@, krw@; ok marco@ on an similar, earlier diff
|
|
|
|
link->bus->sc_dev.dv_unit.
|
|
|
|
due to sizeof pointer instead of sizeof of type
with tedu; some quibbles spotted by miod, ok tedu miod jsg
jakob -- please push this upstream
|
|
isn't it awesome that 5 out of 6 cases of this crap were in the various ppp
implementations? ok claudio ryan
the 6th to be fixed when we return from MEC
|
|
make sppp_flush use IFQ_PURGE instead of handrolling the same making
assumptions about ifqueue internals. ok ryan claudio
|
|
leaves an empty page in curpage, and this inconsistency slowly spreads
until finally one of the other pool checks freaks out.
ok art deraadt
|
|
ones that make assumptions about ifqueue internals... ok ryan claudio
|
|
why bother with APIs when you can muck with internals directly, it's obvious
there'll NEVER be changes, right. ok claudio ryan
|
|
|
|
|
|
the extraction loop should stop.
No more 298 pages in 42 segments when asking for only 32 pages in 1 segment.
ok oga@
|
|
found by jsg, "looks right" tedu jsg
|
|
Instead of allocating a crypto op and the optional dma buffer on each
and every io, preallocate a list of softraid crypto wus that contain a
buffer of the max size we will use (MAXPHYS). since we know the number
of openings we have in advance this means that on each io we just pick
one, shorten the list of crypto descs, init any values then do the io.
ok jsing (who provided many useful comments. he also provided a smarter way of
handling the cryptop lists which is not in this diff but will be implemented
soonish), marco@.
dlg@ pointed out that this should probably use iopools but letting
disciplines allocate their own iopool involves more softraid rejigging
that will have to be done first. For now this is sufficient.
|
|
|
|
|
|
|
|
devices must have an OpenBSD MBR partition to install /boot into.
But search anything except floppy devices (e.g. vnd) for such a
partition.
Feedback & ok deraadt@
|
|
|
|
|
|
a ffs frag size can be less than the d_secsize of the disk. Make
sure amd64 writedisklabel() puts the disklabel where readdoslabel()
will read it. Tweak i386/amd64 installboot/biosboot so sectors are
indeed used where sectors are claimed.
Lets me fdisk, newfs, mount and installboot onto 2048 and 4096
byte sector devices. Other filesystem utilites will still hold
surprises.
Note that actually booting from such devices will await BIOSen that
acknowledge such devices as bootable.
ok guenther@
|
|
i386 disobeys the Nth commandment. Fix this. While here, make i386 and amd64
definitions of iplclock and statclock match.
ok art@, kettenis@
|
|
the MSI enabled flag there such that the driver actually pays attention to it.
Found out the hard way by Chris Smith on an 82540EM, which defenitely does
not like MSI.
ok deraadt@
|
|
and a pool_init flag to aggressively run pool_chk. ok art deraadt
|
|
|
|
- State that ftell() and ftello() return -1 on error rather than just a
non-zero value. The latter is not specific enough, because these
functions can also return non-zero on success.
- For clarity and search-friendliness, don't refer to fgetpos() and
fsetpos() as "the others", but mention them by name.
ok millert otto
|
|
|
|
|
|
after_unlock code is needed if this happens, and running it was even
wrong because we weren't setting b_resid, so uio_offset (and higher
up, f_offset) would be incorrectly adjusted.
Discussed with deraadt@.
|
|
physio().
ok deraadt@
|
|
msync(size == 0) did strange things based on the original mapping
segments and trying to manipulate same. This code was copied from the
original vm when we moved to uvm.
posix says nothing about this behaviour and anything that depends on it is
systemically broken, so rip it out and make sys_msync about 30% smaller.
ok deraadt@, tedu@, guenther@.
|
|
|
|
|
|
|
|
|
|
are available in the same tree, only show the newer of the two.
So far, this only handles the case without -a;
i need to look at the case with -a next.
|
|
|
|
ok guenther
|
|
become empty. In that case, the hme driver could not recover as
the ring was only filled after receiving data. Check and potentially
fill an empty receive ring every second in hme_tick().
ok kettenis@
|
|
ok mikeb@
|
|
YES!! miod@
|
|
rtm_hdrlen instead. Also move the ifindex check way further up.
OK dlg@ bluhm@
|
|
even though -interface is the same as -iface.
OK jmc@
|
|
|
|
by using one mandoc process and one timestamp file per man page.
This might be slightly slower on single-processor machines,
but it's considerably faster with make -j.
"I think so" deraadt@
|
|
problem has been tracked down. This fixes the sharing of the signal
handling state: shared bits go in sigacts, per-rthread bits goes in
struct proc.
ok deraadt@
|
|
- mdoc(7): fix an assertion if the first line after .Bd -column
starts with a blank, and some simplifications in mdoc_argv.c
- man(7): literal mode ends at .SH and .SS (bug reported by naddy@)
- allow .RS/.RE blocks to nest (bug reported by dcoppa@ and gsoares@)
- improve vertical spacing of man(7) blocks
- roff(7): clear user-defined strings when starting a new file
- correct ID tags in -T[x]html
|
|
bounds_check_with_label() and generally cleanup the code while here.
ok krw@ ("I like this a *lot*"); "i like it" tedu@
|
|
ok oga@
|
