| Age | Commit message (Collapse) | Author |
|---|
|
introduced in the previous step, and use atomic instructions to make the
tx completion path mpsafe as well.
ok claudio@, mpi@
|
|
fail. This codepath doesn't exist upstream, so it's fine to change.
ok jsg@
|
|
and return error. From Alexey Suslikov <alexey.suslikov at gmail.com>
|
|
Adds support for HT MCS 0-7 (theoretical limit 65 Mbit/s) and the
reception of A-MSDU and A-MPDU aggregated frames.
None of the optional 11n features are supported for now.
MIMO, 40Mhz channels, short guard interval, etc. are left for future work.
And we're not sending A-MSDU or A-MPDU frames yet either.
Tested with various 11a/b/g/n access points. With some APs I'm seeing
a noticable increase in throughput, especially on 5Ghz.
Also, fix automatic rate selection by using the current Tx rate selected
by AMRR as the upper bound for the firmware's rate table and updating the
firmware's table whenever AMRR switches Tx rate, rather than setting the
table just once after association and ignoring AMRR updates.
ok mpi@ krw@ (earlier version), ok jasper@
|
|
|
|
|
|
similar to config_defer(9).
ok mikeb@, deraadt@
|
|
accelerator drivers. No longer used by anything. ok sthen@ mikeb@
|
|
|
|
tweak it to use ifq_restart so ifq_clr_oactive is serialised with start.
ok jmatthew@
|
|
this is handled in audio(4) layer. From
Alexey Suslikov <alexey.suslikov at gmail.com>
|
|
(HBR2) for Broadwell and non-ULX Haswell DisplayPort. This enables
support for 3840x2160 60Hz SST.
Initial patch from and tested by Scot Doyle.
drm/i915: Enable 5.4Ghz (HBR2) link rate for Displayport 1.2-capable devices
from Todd Previte
06ea66b6bb445043dc25a9626254d5c130093199
drm/i915: don't try DP_LINK_BW_5_4 on HSW ULX
from Paulo Zanoni
9bbfd20abe5025adbb0ac75160bd2e41158a9e83
drm/i915/dp: add missing \n in the TPS3 debug message
from Jani Nikula
f8d8a672f9370278ae2c9752ad3021662dbc42fd
drm/i915/dp: only use training pattern 3 on platforms that support it
from Jani Nikula
7809a61176b385ebb3299ea43c58b1bb31ffb8c0
|
|
work is represented by struct task.
the start routine is now wrapped by a task which is serialised by the
infrastructure. if_start_barrier has been renamed to ifq_barrier and
is now implemented as a task that gets serialised with the start
routine.
this also adds an ifq_restart() function. it serialises a call to
ifq_clr_oactive and calls the start routine again. it exists to
avoid a race that kettenis@ identified in between when a start
routine discovers theres no space left on a ring, and when it calls
ifq_set_oactive. if the txeof side of the driver empties the ring
and calls ifq_clr_oactive in between the above calls in start, the
queue will be marked oactive and the stack will never call the start
routine again.
by serialising the ifq_set_oactive call in the start routine and
ifq_clr_oactive calls we avoid that race.
tested on various nics
ok mpi@
|
|
This is a simple driver for the virtual PCI device with only one
purpose: to allocate the interrupt vector and communicate that
to the Hypervisor in case the IDT callback method was not available.
With input from and OK deraadt, mlarkin, reyk, mpi
|
|
prevent it from moving off-channel during association. The firmware issues
interrupts at beginning and end of the time event. The driver tried detecting
the beginning with a tsleep() in the newstate task followed by a wakeup()
from the interrupt handler. However, sometimes the newstate task did not get
scheduled until the time event had already passed, and association was aborted.
In rare cases the newstate task would even sleep forever and the iwm(4)
interface would stop working until reboot.
Fix these issues by issuing the time event and continuing association without
checking for a "go" from the firmware. Our kernel does not provide the
scheduling guarantees required for such precise synchronization so
association is more likely to fail with the additional check than without.
ok mpi@ tedu@
|
|
drivers still set IFF_NOTRAILERS while others do not. Remove all usage of
the flag from the drivers which in ancient times used it (and the modern
drivers which blindly copied it from those drivers of yore).
suggested by guenther. ok mpi
|
|
|
|
from and tested by daniel@
|
|
from and tested by daniel@
|
|
problem in rev 1.47 spotted by clang's -Wswitch-bool
ok jmatthew@ kettenis@
|
|
If we haven't negotiated the flush feature, don't send
VIRTIO_BLK_T_FLUSH to the device.
noticed by mlarkin@
|
|
Declare all variables at the start of the function and get rid of the
strangely-indented block around the second half of the function.
|
|
The return value was unused
|
|
outside the kernel lock.
Remove tx descriptor lists (essentially backing out if_bnx.c r1.77),
add an interrupt barrier in bnx_stop, check the rx ring state before receiving
packets, adjust the tx counter with atomic operations, and rework bnx_start
to check for ring space before dequeueing and drop the packet if bnx_encap
fails.
tested on BCM5708 by me and on BCM5709 by Hrvoje Popovski
ok dlg@
|
|
as per the stack commit, the driver changes are:
1. setting ifp->if_xflags = IFXF_MPSAFE
2. only calling if_start() instead of its own start routine
3. clearing IFF_RUNNING before calling if_start_barrier() on its way down
4. only using IFQ_DEQUEUE (not ifq_deq_begin/commit/rollback)
|
|
the packet if encap fails, use m_defrag if the mbuf is too fragmented,
adjust sc_tx_cnt using atomic operations, hold the kernel lock while calling
cas_init and cas_start from the interrupt handler, use an interrupt barrier
in cas_stop, and finally mark the interupt handler mpsafe.
encouragement from dlg@ and kettenis@
|
|
This is required to get speaker audio on the HP Elitebook 2540p.
Reported and fix tested by Mike Cond.
|
|
condigure bridges left unconfigured by the system firmware. Not perfect yet,
but good enough to make the Apple Thunderbolt Giabit Ethernet adapter work
when inserted at boot time.
ok deraadt@
|
|
the MacBookPro12,1, 3rd gen Lenovo X1 Carbon (when hacked to use the standard
acpi mechanism for this purpose) and probably many others. Seems this
hasn't been figured out over in Linux land yet.
|
|
|
|
ok deraadt@
|
|
|
|
|
|
dequeueing a packet, then drop the packet if it can't be sent, rather than
using two phase dequeue then commit.
Adjust sc_tx_cnt using atomic operations, take the kernel lock before
calling gem_start or gem_init from the interrupt handler, and use an
interrupt barrier when taking the interface down.
With all this done, we can mark the interrupt handler for pci (but not sbus)
attached gem(4) as mpsafe.
mpi@ wrote this initially, I fixed bugs to keep sparc64 happy and kept it up
to date with ifq changes.
ok mpi@ dlg@
|
|
the system firmware.
|
|
|
|
there are two things shared between the network stack and drivers
in the send path: the send queue and the IFF_OACTIVE flag. the send
queue is now protected by a mutex. this diff makes the oactive
functionality mpsafe too.
IFF_OACTIVE is part of if_flags. there are two problems with that.
firstly, if_flags is a short and we dont have any MI atomic operations
to manipulate a short. secondly, while we could make the IFF_OACTIVE
operates mpsafe, all changes to other flags would have to be made
safe at the same time, otherwise a read-modify-write cycle on their
updates could clobber the oactive change.
instead, this moves the oactive mark into struct ifqueue and provides
an API for changing it. there's ifq_set_oactive, ifq_clr_oactive,
and ifq_is_oactive. these are modelled on ifsq_set_oactive,
ifsq_clr_oactive, and ifsq_is_oactive in dragonflybsd.
this diff includes changes to all the drivers manipulating IFF_OACTIVE
to now use the ifsq_{set,clr_is}_oactive API too.
ok kettenis@ mpi@ jmatthew@ deraadt@
|
|
|
|
From Bryan Vyhmeister.
|
|
Problem reported by Mikolaj Kucharski.
|
|
|
|
You never need <netinet/ip_var.h> nor <netinet/in_systm.h>.
|
|
|
|
This header is only needed because <netinet/if_ether.h> declares a
structure that needs it. But it turns out that <net/if.h> already
includes it as workaround.
A proper solution would be to stop declarting "struct ether_arp"
there. But no driver should need this header.
|
|
|
|
|
|
ServerWorks SATA.
Allow to reliably found disks on xserve G5.
from miod@
|
|
turns out i was calculating the number of packets (not descriptors)
on the tx ring, and then using that as the free space for descriptors.
|
|
|
|
ok jsg@
|
