| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
activate it in the build.
|
|
pmap.h now includes pte.h.
|
|
tested by deraadt@ on a gcc2 arch
looks ok ray@ ok deraadt@
|
|
while there, put all function prototypes in header file.
ok kurt@
|
|
|
|
use a simpler way to allocated memory in allocset, mostly done by
replacing malloc/realloc dance, by only one realloc
add comments about variables that are not used uninitialized, even
if gcc told the contrary
another malloc/realloc -> realloc change
Work initially started by otto@, and then I joined him
ok otto@ ray@
|
|
SSL_get_shared_ciphers() with strlcat(3).
ok deraadt@ markus@
|
|
- Move the functionality of choosing a process from cpu_switch into
a much simpler function: cpu_switchto. Instead of having the locore
code walk the run queues, let the MI code choose the process we
want to run and only implement the context switching itself in MD
code.
- Let MD context switching run without worrying about spls or locks.
- Instead of having the idle loop implemented with special contexts
in MD code, implement one idle proc for each cpu. make the idle
loop MI with MD hooks.
- Change the proc lists from the old style vax queues to TAILQs.
- Change the sleep queue from vax queues to TAILQs. This makes
wakeup() go from O(n^2) to O(n)
there will be some MD fallout, but it will be fixed shortly.
There's also a few cleanups to be done after this.
deraadt@, kettenis@ ok
|
|
ok jmc@
|
|
OK millert
|
|
instead of AUDIO_GETINFO
- make SNDCTL_DSP_GETISPACE return correct values by using AUDIO_GETRRINFO
instead of AUDIO_GETINFO
- remove calls to setblocksize() from SNDCTL_DSP_GET[IO]SPACE. changing
the blocksize inside play or record loops, where these ioctls are most
often used, can have serious negative consequences. the blocksize
should be "fixed" to a power of 2 before these ioctls are called by
using SNDCTL_DSP_SETFRAGMENT or SNDCTL_DSP_GETBLKSIZE.
heavily tested by myself. no negative feedback from anyone else.
ok ratchov
|
|
|
|
remove the recursive lock in __seekdir() and ensure all callers properly
lock dd_lock prior to calling. reported and tested by bernd@
okay marc@ tedu@
|
|
SNDCTL_DSP_SETTRIGGER is used to pause and unpause recording and
playback data transfers. SNDCTL_DSP_GETTRIGGER returns the paused
states.
this implementation of SNDCTL_DSP_SETTRIGGER also sets the audio(4)
mode (AUMODE_[PLAY|RECORD]), depending on which direction (play, record
or both) is to be triggered.
this can be used allow OSS audio programs to work in full-duplex
mode without changing the audio(4) layer itself. according to
OSS docs, SNDCTL_DSP_SETTRIGGER is a fairly important ioctl, and
should be used "when even the first read() cannot block."
tested to not break any ports currently using SNDCTL_DSP_SETTRIGGER
which has been a no-op for quite some time.
ok ratchov
|
|
From OpenSSL_0_9_8-stable branch.
ok djm@
|
|
assign values to struct members directly
remove now-unused variable
ok otto@
ok & input millert@
|
|
|
|
From: Pierre Riteau <pierre.riteau@free.fr>
|
|
ok ray@
|
|
ok millert
|
|
|
|
variable TMPDIR or the argument `dir' is an empty string.
With and ok millert@ ray@
|
|
Mostly path construction, where truncation could be bad.
ok and input from deraadt@ millert@ ray@
|
|
|
|
triggered and are off-by-one anyways. Instead replace them
with correct checks and add all the missing cases.
ok deraadt@
|
|
OK joris@, otto@
|
|
assembly on all platforms. ok deraadt@
|
|
providing a stub. ok deraadt@
|
|
|
|
|
|
|
|
from Robert Elz, netbsd-bugs pr#36933
ok otto deraadt
|
|
|
|
case and 0 if the locale was found. So checking against -1 makes
no sense.
ok millert@
|
|
ok millert@ ray@
|
|
using the return value of _ng_print(), which returns void.
ok millert@
|
|
|
|
Based on a diff from Mike Belopuhov. OK jmc@
|
|
|
|
|
|
object destructors called at dlclose() time. Inspired by similar changes in FreeBSD and NetBSD.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
