| Age | Commit message (Collapse) | Author |
|---|
|
deiline drm_core_findmap(), and merge drm_core_ioremap{,free}() into
drm_ioremap{,free}() instead of having an inline that just calls another
function.
|
|
|
|
loading the raw memory into.
similair to the change made to src/sys/arch/amd64/amd6/bus_dma.c 1.18
|
|
loading the raw memory into, particularly the segment size constraint.
written in june/july last year, but my studies held me back from handling
it.
|
|
affected by adjusting the clock.
|
|
That way, you can edit the new domain Makefile before using it,
in particular to change variables like DIR and UNSECURE.
from ajacoutot@ with message tweaks and documentation updates by myself
"I like this" otto@
|
|
|
|
|
|
this option is needed so gnu cvs clients can play along
properly when running checkout <module> in a remote setup.
if we do not support this, gnu cvs refuses to send us any
existing files already on disk and opencvs will consider everything
as NEEDS_CHECKOUT, instead of running an update on existing stuff.
problem noticed by fgsch@
commited with opencvs, but please everybody else
do not use opencvs to commit yet.
|
|
|
|
Reported by Kamil Mielcarek
|
|
theyve been used, in particular the mfi header flags which has a bit
that specifies if a command should be completed via the interrupt path.
if we use a ccb during boot we set that bit, but it isnt necessarily
cleared by things that use it later on. this means a ccb we expected to
complete via an interrupt never actually generates an interrupt or appears
in the reply queue. this obviously stalls the io.
|
|
- don't mix use of const and non-const pointers
problem noted by Jordi Beltran Creix on tech@
|
|
new labels.
|
|
sthen likes it
|
|
code that wasnt suppose to dissapear:
for uptodate files, we want to compare rev1 against
file_rcsrev to see if we wish to diff or not. saves a bit off logic.
commited with opencvs,
permission by deraadt@
|
|
|
|
|
|
|
|
dying as soon as I can make it doesn't mean I shouldn't be able to look
at it in the meantime.
|
|
kill the compat define. Should have done this ages ago.
|
|
|
|
|
|
|
|
|
|
|
|
ago.
|
|
handler anyway. replace it with ${OpenBSD_driver_name}_intr like most of
the rest of the tree. No functional change.
|
|
defines. 0 and 1 are fine for returning from irq handlers so why bother
with the indirection?
|
|
the unmacroed/typedefed version.
|
|
the dev->driver irq_handler member since we now don't need to know.
Shaves a few bytes, and makes me happy.
|
|
We already have a lock protecting the irq registers, so extend it
trivially so it protects user interrupts too. Switch irq handler to
purely local handler. Finally remove an extra test in i915_wait_irq()
since we won't sleep even without that if the test will pass.
irq_lock and handler_wrap may now die. Coming soon.
|
|
handler directly, no fences to worry about right now.
|
|
interrupt we provide is vblank, it's just switching the handler. knf
while here.
|
|
races when waiting on fences, and use that in preference to irq_lock.
Don't use drm_irq_handler_wrap() anymore for irq handler, just the
mgadrm main irq handler. knf while here.
|
|
(vblank is already protected by that subsystem). Also, do all checks for
the correct software interrupt having fired inside DRM_WAIT_ON(), since
it won't sleep if it's already passed it doesn't gain us much (skips one
lock grab). Finally, establish the interrupt with only the main radeon
irq handler, don't bother with drm_irq_wrap. bunch of knf while i'm in
here.
|
|
expanding the function in line here.
|
|
and the lock to manipulate.
first step of removing drm_irq_handler_wrap which just grabs the
irq_lock, and eventually irq_lock. drivers should manage their own.
|
|
needed, but some machines seem to work much better with it.
|
|
forever.
|
|
noticed the hard way here at work.
commiting with opencvs,
permission by deraadt@
|
|
imap, mostly) to provide "username1\0real_username\0password" as your
base64 encoded string for authentication. We currently don't handle
that, instead expecting the first byte to be a NUL. So fix that up by
scanning for the first string, and ignoring it if it's there. The string
is also stupid in that the last bit (password) may not be NUL
terminated, so pay attention to that in our decoding and make sure that
it's always terminated correctly.
It's been discussed, and this decoding really should happen in the
unauthenticated process, not in the privileged one, but that is another
diff.
Problem found by todd@, who kindly helped me debug this and confirmed
that it now works with kmail, mutt and thunderbird.
"if it makes more stuff work, please commit" jacekm@.
-0- - not a smtpd hacker and I resent your implication.
|
|
of the struct proc* as the identifier for SEM_UNDO tracking and only
call semexit() from the original thread, once the process as a whole
is exiting
ok tedu@
|
|
member from the lock structure.
|
|
have a file on disk we'll need to load the revision into
a temporary file to do our comparison.
do not even bother trying diffing 2 off the same revisions
either. causes less noise... way less.
you guessed it, commited using opencvs entirely.
with permission from deraadt@
|
|
to the server for verification.
commited with opencvs,
permission by deraadt@
|
|
hardware virtual address offset, not a physical one. From upstream.
|
|
unobfuscate things a bit.
|
|
failure while i'm at it.
Been in snaps for a while. ok deraadt@
|
|
ok deraadt@
|
