| Age | Commit message (Collapse) | Author |
|---|
|
notably fix a regression in parsing DTD parameter entities.
ok miod@ nicm@
|
|
pmap_{copy,zero}_page, but use direct mapped pages instead.
|
|
fix OpenBSD/sparc operation in QEMU, which only provides the frame buffer
memory and none of the accelerated stipple and blit spaces.
|
|
This option was not useful, you never want mandoc to die
just because there is an invalid character in the input file,
neither in production nor when linting: a warning is sufficient.
This was particularly annoying because it was part of -fstrict
and could not be switched off.
"less is more" kristaps@
|
|
more to come later)
Specfically, there is no reason to reserve a special virtual address just so we
can do boot dump, we have a direct map of every page anyway.
since pmap_map is deprecated and MD only anyway, this means we can remove that
interface too. If anything this should increase reliability since pmap_enter
won't fail under memory pressure during dump (unlikely but possible). It is also
simpler and smaller ;)
Tested by myself and ckuethe, no regressions.
ok miod@
|
|
do not attach to the Gen2b devices yet as it is not clear wether or
not they have their own hardware revision type.
|
|
|
|
turns out the Rx gain tables are the same as 2.0 (and the Tx gain
registers too), which saves us a few bytes.
|
|
kristaps will fix this soon;
from Christian Brueffer
|
|
|
|
|
|
|
|
it's queued in the Tx FIFO.
otherwise we might end up losing the reference to the Tx buffer
if we return on error in ar{5008,9003}_tx().
cleanup some dma sync calls while i'm here.
|
|
Tested (and initial tweaked diff) from Erik Mugele; thanks!
|
|
|
|
Tested (and inital patch that I tweaked) from Erik Mugele; thanks!
|
|
from being left hanging around when, for example, a SSH session is
disconnected.
ok nicm@
|
|
comments and okay kettenis@
|
|
okay kettenis@
|
|
notifications.
|
|
|
|
C application.
OK yasuoka@, reyk@
|
|
This is needed because pf_state_peer_hton() skips some fields in certain
situations which could result in garbage beeing sent to the other peer.
This seems to fix the pfsync storms seen by stephan@ and so dlg owes me
a whiskey.
OK dlg@, stephan@
|
|
opposite) and on AR9380 2.0.
tested on my AR9280 2.1 with a NETGEAR WNHDE111 AP.
|
|
|
|
#ifndef IEEE80211_NO_HT for consistency.
|
|
- the ROM deviceCap field is now 8 bits (instead of 16), so the number
of entries in the key cache is always 0; just use the default value.
- AR_CR_RXE is equal to 0 on AR9003.
- do not use ``m'' unititialized in ar9003_rx_process.
- if an Rx descriptor is not valid (bad signature), skip it instead of
leaving it at the head of the FIFO.
- update the Rx software descriptor with new virtual and physical address
of Rx descritor when mapping a new buffer.
|
|
see http://marc.info/?l=linux-wireless&m=127354213505700&q=raw
for the list of changes.
|
|
and this lets gcc emit code with xadd/cmpxchg instructions.
ok kettenis@
|
|
as 32bit values instead of 64bit values, synchronizing it with the
libc stub. gcc4 apparently packs the stack in _dl_bind() more
tightly, so this was resulting in the saved rbx register being
overwriten.
ok drahn@, marco@, NicM@
|
|
|
|
|
|
|
|
|
|
OK sthen@
|
|
where it is not necessary to guess protocols by looking at the first nibble.
in_gif_output() will encapsulate the packet but not send it. Because of
etherip support and the way the bridge works a minimal hack is needed in
gif_start() to ensure that the bridged packets are encapsulated as well.
This actually started with the idea to add MPLS support but that turned out
to be not as simple as in the gre(4) case.
Tested by myself (IP, IPv6, etherip, MPLS), sthen@ (IP, IPv6), naddy (IPv6)
OK sthen@
|
|
OK sthen@
|
|
is defined.
|
|
|
|
This enabled GEM for the intel driver unconditionally. The legacy
codepaths will be removed in approximately one week since they are now
completely unused.
After discussion with matthieu@, drahn@, kettenis@ and marco@ (well,
mostly nagging from marco ;).
|
|
(reading back the relocation).
It doesn't add any real security and when we actually need to map the
buffer on demand to read/write it makes things cripplingly slow. The
correct way to make this utterly incorruptible is a radeon-kms-like
command checker to the command streams. This is on my todo list.
Thanks to drahn@ for additional testing.
|
|
Makes gcc4 happy.
ok miod@
|
|
this driver to work on machine with low kva and large apertures.
tested by myself and drahn@
|
|
things that there really isn't a decent api for elsewhere.
Since on recent intel IGPs the gtt aperture is too big (256meg is not
uncommon) to be mapped on a kva-constrained arch like i386, introduce an agp
mapping api that does things depending on arch.
On amd64 which can afford the space (and will use the direct mapping
again soon)just do bus_space_map() on init, then parcels things out
using bus_space_subregion(), thus avoiding map/unmap overhead on every
call (this is how inteldrm does things right now).
On i386, we do bus_space_map() and bus_space_unmap as appropriate. Linux
has some tricks here involving ``atomic'' maps that are on only one cpu
and that you may not sleep with to avoid the ipi overhead for tlb
flushing. For now we don't go down that route but it is being
considered.
I am also considering if it is worth abstracting this a little more,
improving the api and making it a general MD interface.
Tested by myself on i386 and amd64 and by drahn@ (who has one of the
machines with an aperture that is too big) on i386.
|
|
ok millert@
|
|
|
|
the types used by GCC match those in machine/_types.h.
|
|
|
|
|
|
ok guenther@
|
