| Age | Commit message (Collapse) | Author |
|---|
|
|
|
broadwell/skylake bug.
|
|
Include file with font data (courier11x18.h) was removed from NetBSD due
to licensing concerns before wsfont was imported into OpenBSD.
Cookie values have been renumbered for consistency.
|
|
device, and factor it into a single function. Without this fix, if the
device is open in full-duplex mode, it could start with empty play
buffer.
|
|
set variables that will be later used as the size argument to
free(NULL calls. This should be harmless as free returns early if the
address is NULL without checking the size. Initialise these variables
before the call to ensure they are never passed to another function
uninitialised.
ok tedu@ millert@ deraadt@
|
|
longer needed, including the imx board IDs.
ok patrick@, jsg@
|
|
ok sthen tb deraadt
|
|
|
|
assuming that the 'a' partition starts at the same location as the
volume header partition #0.
Diff from Miod Vallat
|
|
routing table.
Found the hardway by Chris Jackman.
|
|
launch with valid vmcs" issue some people have been seeing when launching
vmm(4) VMs.
tested by reyk@
|
|
means that even though the hardware in the end never leaves the 32-bit
address space, some addresses used solely in the device tree can be
bigger than 32-bit. As bus_space_map(9) takes addresses of size
bus_addr_t, which is 32-bit on ARMv7, we cannot pass those virtual
addresses to the parent bus, even though it will be mapped back into a
32-bit address in the end. To work around this, make bus_space_map(9)
take a 64-bit address. Since this is implemented as a macro and
function pointer we can safely do that without harming any other
architecture.
ok kettenis@ deraadt@
|
|
i386. As newer ARMs where we use device tree from the get go don't
necessarily have a 'platform', this will allow drivers to hook
themselves as a way to reset the CPU.
ok jsg@ kettenis@ tom@
|
|
Use a local variable struct process *pr to simplify expressions
ok deraadt@
|
|
ok mpi@ mikeb@
|
|
telling the audio subsystem about this later on in boot. (Mute state
is maintained by the BIOS over a suspend, hibernate, reboot, etc -- this
may be last missing piece of the puzzle we tore our hair out in Brisbane)
diagnosis and code by Anton Lindqvist, ok ratchov
|
|
response thus respecting what the OpenFlow 1.3.5 spec says.
ok reyk@
|
|
in HC drivers.
Only ehci(4) was performing pre & post synchronization completly. Now
pre-synchronization is done in usbd_transfer() and post-synchronization
in usbd_transfer_complete(). This will allow us to remove the partial
synchronization primitives from the pipe handlers.
This is similar to what FreeBSD did in r158998, r164519 and r176203.
From Marius Strobl.
ok kettenis@
|
|
There's no need to insert an RTF_LOCAL route if it is already there, not
if a route with the same destination exist.
This fixes a KASSERT() triggered by adding an alias for an address already
present in the ARP cache as reported by weerd@ and Peter J. Philipp.
This should also fix a KASSERT() triggered by a NDP change reported by
Sebastien Marie.
ok bluhm@
|
|
Found by Chris Jackman, thanks!
|
|
The current reason is that rtalloc_mpath(9) inside ip_output() might
end up inserting a RTF_CLONED route and that require a write lock.
ok kettenis@, bluhm@
|
|
string literal" warning from clang.
ok jca@
|
|
latter will always be non-NULL, and therefore the test will always be false.
Caught by clang.
ok jca@, mikeb@, jsg@
|
|
child_return() like on other archs, so cpu_fork() doesn't need to
touch them.
ok kettenis@ deraadt@
|
|
ok mpi@
|
|
Do not fold multiple DMA synchronizations into one when chaining TRBs
as the ring might wrap.
Add missing "READ" transfer direction from the HC to host when applicable.
From Marius Strobl.
|
|
time_second is unix time so it can be affected by clock changes.
time_uptime is monotonic so it isnt affected by clock changes. that
in turn means route expiries wont jump with clock changes if set
against time_uptime.
the expiry is translated into unix time for export to userland though.
Should fix mismatch between route timers that were already converted
and ND default routers that were still using time_second.
Tested by matthieu@ and sthen@
ok sthen@, dlg@
|
|
code to if.c.
ok mpi@
|
|
as noted by haesbaert, this is necessary to avoid deadlocks because
the scheduler can call back into the timeout subsystem while its
holding its own locks.
this happened in two places. firstly, in softclock() it would take
timeout_mutex to find pending work. if that pending work needs a
process context, it would queue the work for the thread and call
wakeup, which enters the scheduler locks. if another cpu is trying
to tsleep (or msleep) with a timeout specified, the sleep code would
be holding the sched lock and call timeout_add, which takes
timeout_mutex.
this is solved by deferring the wakeup to after timeout_mutex is
left. this also has the benefit of mitigating the number of wakeups
done per softclock tick.
secondly, the timeout worker thread takes timeout_mutex and calls
msleep when there's no work to do (ie, the queue is empty). msleep
will take the sched locks. again, if another cpu does a tsleep
with a timeout, you get a deadlock.
to solve this im using sleep_setup and sleep_finish to sleep on an
empty queue, which is safe to do outside the lock as it is comparisons
of the queue head pointers, not derefs of the contents of the queue.
as long as the sleeps and wakeups are ordered correctly with the
enqueue and dequeue operations under the mutex, this all works.
you can think of the queue as a single descriptor ring, and the
wakeup as an interrupt.
the second deadlock was identified by guenther@
ok tedu@ mpi@
|
|
|
|
configuration that attaches with the fdt and writes values to registers
based on the pinctrl data in the fdt. Handles 16 bit values
(OMAP3/OMAP4) and 32 bit values (AM335x/AM57xx).
Feedback from and ok kettenis@ tom@ on an earlier version.
|
|
"ifconfig: SIOCSETVLAN: No buffer space available" when creating vlan
interfaces without first lowering the mtu. Reported by Steven Chamberlain,
diagnosed by mikeb@, patch from Brad. ok bmercer@
|
|
from Sebastien Marie
|
|
ok deraadt@ tom@ patrick@
|
|
gets set. Therefore, ahci_port_softreset() restores the sate of the
former register once the device detection sequence is finished. The
device detection code in ahci_pmp_port_portreset() does not restore
AHCI_PREG_CMD afterwards, so let it catch up. Apparently, this part
was an oversight as ahci_pmp_port_portreset() did not otherwise use
"cmd".
From Marius Strobl
tested by awolk (amd64), bluhm (amd64, i386), myself (amd64, armv7)
ok jmatthew@
|
|
slot. Thus, extend the check whether NCQ actually should be enabled
accordingly.
From Marius Strobl
tested by awolk (amd64), bluhm (amd64, i386), myself (amd64, armv7)
ok jmatthew@
|
|
recovery from the CCB pool sized based on the NCS capability, i. e.
number of command slots reported by the controller, it is necessary
to pretend at least 2 slots in sc->sc_ncmds for devices without NCQ
support. That way, also at least 1 available slot is made available
for atascsi(4). Otherwise, controllers having only a single command
slot will trigger "no free xfers on a new port" in atascsi(4).
Note that pretending 2 command slots is also fine with the abuse of
the NCQ error recovery CCB in ahci_port_softreset().
From Marius Strobl
tested by awolk (amd64), bluhm (amd64, i386), myself (amd64, armv7)
ok jmatthew@
|
|
does not say anything about which type (snoopable/uncached etc.) of
bus transactions are issued by HBAs on accesses to the descriptors.
Thus, the right assumption would be "classical" concurrent accesses
by both CPU/driver as well as hardware to the descriptors and their
DMA memory backings respectively, which is also confirmed by actual
testing. Consequently, switch to BUS_DMA_COHERENT mappings for said
DMA memory as otherwise corruption of descriptors is seen.
From Marius Strobl
tested by awolk (amd64), bluhm (amd64, i386), myself (amd64, armv7)
ok jmatthew@
|
|
ok mpi@
|
|
controller. This register controls whether the controller is in device or
host mode on many dual role controllers and gets reset during a reset of
the controller, placing the controller in (non-functional) idle mode. By
saving and restoring it, we keep the controller in host mode. Since this is
a non-standard register, add a new EHCIF_USBMODE flag and only do the save and
restore if it has been set.
Makes the upper "OTG" port of the Cubox-i work.
ok mpi@, jsg@
|
|
|
|
encodings.
ok deraadt@, jca@
|
|
|
|
Years ago Theo made read(2) return 0 on directories, instead of dumping
the directory content. Another behavior is allowed as an extension by
POSIX, returning an EISDIR error, as used on a few other systems. This
behavior is deemed more useful as it helps spotting errors. This
implies that it might break some setups.
Ports bulk builds by ajacoutot@ and naddy@, ok millert@ bluhm@ naddy@
deraadt@
|
|
ok reyk
|
|
|
|
to previous behaviour of starting quick, exponentially backing off and
settling on every 60 seconds.
sthen@ noticed that this broke the backing off when we don't receive
an advertisment and so we would hammer the network every second which
is particularly bad on wifi networks.
OK sthen@
|
|
|
|
input goda, ok reyk goda
|
|
confusion about the tunnel endpoints when responding to the peer.
OK yasuoka@
|
