This is a bug fix for some versions of g200se cards while doing
mode-setting.
Signed-off-by: Christopher Harvey <charvey@matrox.com>
Tested-by: Julia Lemire <jlemire@matrox.com>
Acked-by: Julia Lemire <jlemire@matrox.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@gmail.com>
ARM cannot handle udelay for more than 2 miliseconds, so we
should use mdelay instead for those.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Dave Airlie <airlied@gmail.com>
The dependecies for BACKLIGHT_CLASS_DEVICE are defined a bit
strange, but it seems one has to always select both BACKLIGHT_CLASS_DEVICE
and BACKLIGHT_LCD_SUPPORT to avoid this error:
drivers/gpu/drm/tilcdc/tilcdc_panel.c:396:
undefined reference to `of_find_backlight_by_node'
Cc: Rob Clark <robdclark@gmail.com>
Cc: dri-devel@lists.freedesktop.org
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Dave Airlie <airlied@gmail.com>
When GPU acceleration is disabled, drm_vblank_cleanup() will free the
vblank-related data, such as vblank_refcount, vblank_inmodeset, etc.
But we found that drm_vblank_post_modeset() may be called after the
cleanup, which use vblank_refcount and vblank_inmodeset. And this will
cause a kernel panic.
Fix this by return immediately if dev->num_crtcs is zero. This is the
same thing that drm_vblank_pre_modeset() does.
Call trace of a drm_vblank_post_modeset() after drm_vblank_cleanup():
[ 62.628906] [<ffffffff804868d0>] drm_vblank_post_modeset+0x34/0xb4
[ 62.628906] [<ffffffff804c7008>] atombios_crtc_dpms+0xb4/0x174
[ 62.628906] [<ffffffff804c70e0>] atombios_crtc_commit+0x18/0x38
[ 62.628906] [<ffffffff8047f038>] drm_crtc_helper_set_mode+0x304/0x3cc
[ 62.628906] [<ffffffff8047f92c>] drm_crtc_helper_set_config+0x6d8/0x988
[ 62.628906] [<ffffffff8047dd40>] drm_fb_helper_set_par+0x94/0x104
[ 62.628906] [<ffffffff80439d14>] fbcon_init+0x424/0x57c
[ 62.628906] [<ffffffff8046a638>] visual_init+0xb8/0x118
[ 62.628906] [<ffffffff8046b9f8>] take_over_console+0x238/0x384
[ 62.628906] [<ffffffff80436df8>] fbcon_takeover+0x7c/0xdc
[ 62.628906] [<ffffffff8024fa20>] notifier_call_chain+0x44/0x94
[ 62.628906] [<ffffffff8024fcbc>] __blocking_notifier_call_chain+0x48/0x68
[ 62.628906] [<ffffffff8042d990>] register_framebuffer+0x228/0x260
[ 62.628906] [<ffffffff8047e010>] drm_fb_helper_single_fb_probe+0x260/0x314
[ 62.628906] [<ffffffff8047e2c4>] drm_fb_helper_initial_config+0x200/0x234
[ 62.628906] [<ffffffff804e5560>] radeon_fbdev_init+0xd4/0xf4
[ 62.628906] [<ffffffff804e0e08>] radeon_modeset_init+0x9bc/0xa18
[ 62.628906] [<ffffffff804bfc14>] radeon_driver_load_kms+0xdc/0x12c
[ 62.628906] [<ffffffff8048b548>] drm_get_pci_dev+0x148/0x238
[ 62.628906] [<ffffffff80423564>] local_pci_probe+0x5c/0xd0
[ 62.628906] [<ffffffff80241ac4>] work_for_cpu_fn+0x1c/0x30
[ 62.628906] [<ffffffff802427c8>] process_one_work+0x274/0x3bc
[ 62.628906] [<ffffffff80242934>] process_scheduled_works+0x24/0x44
[ 62.628906] [<ffffffff8024515c>] worker_thread+0x31c/0x3f4
[ 62.628906] [<ffffffff802497a8>] kthread+0x88/0x90
[ 62.628906] [<ffffffff80206794>] kernel_thread_helper+0x10/0x18
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Binbin Zhou <zhoubb@lemote.com>
Cc: <stable@vger.kernel.org>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Acked-by: Paul Menzel <paulepanter@users.sourceforge.net>
Signed-off-by: Dave Airlie <airlied@gmail.com>
Pull m68k fix from Geert Uytterhoeven:
"A boot lock-up on Mac, also destined for stable"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
m68k/mac: Fix unexpected interrupt with CONFIG_EARLY_PRINTK
If an extension device isn't initialized properly, or during hardware
initialization, a device might send extension data which is all 0xff.
This is ambigious because this is also a valid normal data report. But
it is impossible, under normal conditions, to trigger valid reports with
all 0xff. Hence, we can safely ignore them.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
I finally got a "Classic Controller" and "Classic Controller Pro" in my
hands and noticed that all analog data was incorrectly parsed. Fix this
up so we report the data that we pretend we do.
I really doubt that this breaks any backwards compatibility, but if we
get any reports, we only need to revert this single patch.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
We normally get EXT hotplug events or poll for MP hotplugging so we
don't need to force extension port initialization during device setup.
But for gen20 devices, we disable MP polling because MP is always
present. However, this prevents MP initialization during device setup
and users need to plug another extension to trigger EXT/MP detection.
Therefore, we now trigger EXT/MP detection during device setup
automatically. This also avoids slightly delayed extension detection
and provides sysfs child-devices prior to the "changed"-uevent during
device setup.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
We need to correctly zero-terminate the input to parse it. Otherwise, we
always end up interpreting it as numbers.
Furthermore, we actually want hexadecimal numbers instead of decimal. As
it is a debugfs interface, we can change the API at any time.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Devices which have built-in motion plus ports don't need MP detection
logic. The new WIIMOD_BUILTIN_MP modules sets the WIIPROTO_FLAG_BUILTIN_MP
flag which disables polling for MP.
Some other devices erroneously report that they support motion-plus. For
these devices and all devices without extension ports, we load
WIIMOD_NO_MP which sets WIIPROTO_FLAG_NO_MP. This effectively disables all
MP detection logic.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
We now have dynamic hotplug support so the old static extensions are no
longer needed nor used. Remove it along CONFIG_HID_WIIMOTE_EXT.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Balance-Boards provide 3 16bit calibration values for each of the 4
sensors. We provide these now as 192bit value via a new "bboard_calib"
sysfs attribute.
We also re-read the calibration data from the device whenever user-space
attempts to read this file. On normal Nintendo boards, this always
produces the same results, however, on some 3rd party devices these values
change until the device is fully initialized. As I have currently no idea
how long to wait until it's ready (sometimes takes up to 10s?) we provide
a simple workaround for users by reading this file.
If we, at some point, figure out how it works, we can implement it in the
kernel and provide offline data via "bboard_calib". This won't break
user-space then.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Two new attributes, "extension" and "devtype" now allow user-space to read
the extension type and device type. As device detection is asynchronous,
we send a CHANGED event after it is done. This also allows user-space to
wait for a device to settle before opening its input event devices.
The "extension" device is compatible with the old "extension" sysfs field
(which was registered by the static extension support code).
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
If we write a DRM mode via debugfs, we shouldn't allow normal operations
to overwrite this DRM mode. This is important if we want to debug
3rd-party devices and we want to see what data is sent on each mode.
If we write NULL/0 as DRM, the lock is removed again so the best matching
DRM is chosen by wiimote core.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
single_open() stores the seq_file pointer in file->private_data. It stores
our ctx pointer in seq_file->private.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Add parsers for motion plus data so we can hotplug motion plus extensions
and make use of them. This is mostly the same as the old static motion
plus parser.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Add a new extension module for the classic controller so we get hotplug
support for this device. It is mostly the same as the old static classic
controller parser.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This moves the nunchuk parser over to an extension module. This allows to
make use of hotplugged Nunchuks instead of the old static parser.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This adds Nintendo Wii Balance Board support to the new HOTPLUG capable
wiimote core. It is mostly copied from the old extension.
This also adds Balance Board device detection. Whenever we find a device
that supports the balance-board extension, we assume that it is a real
balance board and disable unsupported hardward like accelerometer, IR,
rumble and more.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
The Wii Remote has several extension ports. The first port (EXT) provides
hotplug events whenever an extension is plugged. The second port (MP)
does not provide hotplug events by default. Instead, we have to map MP
into EXT to get events for it.
This patch introduces hotplug support for extensions. It is fairly
complicated to get this right because the Wii Remote sends a lot of
noise-hotplug events while activating extension ports. We need to filter
the events and only handle the events that are real hotplug events.
Mapping MP into EXT is easy. But if we want both, MP _and_ EXT at the same
time, we need to map MP into EXT and enable a passthrough-mode. This will
then send real EXT events through the mapped MP interleaved with real MP
events. But once MP is mapped, we no longer have access to the real EXT
registers so we need to perform setup _before_ mapping MP. Furthermore, we
no longer can read EXT IDs so we cannot verify if EXT is still the same
extension that we expect it to be.
We deal with this by unmapping MP whenever we got into a situation where
EXT might have changed. We then re-read EXT and MP and remap everything.
The real Wii Console takes a fairly easy approach: It simply reconnects to
the device on hotplug events that it didn't expect. So if a program wants
MP events, but MP is disconnected, it fails and reconnects so it can wait
for MP hotplug events again.
This simplifies hotplugging a lot because we just react on PLUG events and
ignore UNPLUG events.
The more sophisticated Wii applications avoid reconnection (well, they
still reconnect during many weird events, but at least not during UNPLUG)
but they start polling the device. This allows them to disable the device,
poll for the extension ports to settle and then initialize them again.
Unfortunately, this approach fails whenever an extension is replugged
while it is initialized. We would loose UNPLUG events and polling the
device later will give unreliable results because the extension port might
be in some weird state, even though it's actually unplugged.
Our approach is a real HOTPLUG approch. We keep track of the EXT and
mapped MP hotplug events whenever they occur. We then re-evaluate the
device state and initialize any possible new extension or deinitialize any
gone extension. Only during initialization, we set an extension port
ACTIVE. However, during an unplug event we mark them as INACTIVE. This
guarantess that a fast UNPLUG -> PLUG event sequence doesn't keep them
marked as PLUGGED+ACTIVE but only PLUGGED.
To deal with annoying noise-hotplug events during extension mapping, we
simply rescan the device before performing any mapping. This allows us to
ignore all the noise events as long as the device is in the correct state.
Long story short: EXT and MP registers are sparsely known and we need to
jump through hoops to get reliable HOTPLUG working. But while Nintendo
needs *FOUR* Bluetooth reconnections for the shortest imaginable
boot->menu->game->menu->shutdown sequence, we now need *ZERO*.
As always, 3rd party devices tend to break whenever we behave differently
than the original Wii. So there are also devices which _expect_ a
disconnect after UNPLUG. Obviously, these devices won't benefit from this
patch. But all official devices were tested extensively and work great
during any hotplug sequence. Yay!
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
IR is the last piece that still is handled natively. This patch converts
it into a sub-device module like all other sub-devices. It mainly moves
code and doesn't change semantics.
We also implicitly sync IR data on ir_to_input3 now so the explicit
input_sync() calls are no longer needed.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Accelerometer data is very similar to KEYS handling. Therefore, convert
all ACCEL related handling into a sub-device module similar to KEYS.
This doesn't change any semantics but only moves code over to
wiimote-modules.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Each of the 4 LEDs may be supported individually by devices. Therefore,
we need one module for each device. To avoid code-duplication, we simply
pass the LED ID as "arg" argument to the module loading code.
This just moves the code over to wiimote-module. The semantics stay the
same as before.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This introduces a new sub-device module for the BATTERY handlers. It
moves the whole power_supply battery handling over to wiimote-modules.
This doesn't change any semantics or ABI but only converts the battery
handling into a sub-device module.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This introduces the first sub-device modules by converting the KEYS and
RUMBLE sub-devices into wiimote modules. Both must be converted at once
because they depend on the built-in shared input device.
This mostly moves code from wiimote-core to wiimote-modules and doesn't
change any semantics or ABI.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
To avoid loading all sub-device drivers for every Wii Remote, even though
the required hardware might not be available, we introduce a module layer.
The module layer specifies which sub-devices are available on each
device-type. After device detection, we only load the modules for the
detected device. If module loading fails, we unload everything and mark
the device as WIIMOTE_DEV_UNKNOWN. As long as a device is marked as
"unknown", no sub-devices will be used and the device is considered
unsupported.
All the different sub-devices, including KEYS, RUMBLE, BATTERY, LEDS,
ACCELEROMETER, IR and more will be ported in follow-up patches to the new
module layer.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Our output queue is asynchronous but synchronous reports may wait for a
response to their request. Therefore, wake them up unconditionally if an
output report couldn't be sent. But keep the report ID intact so we don't
incorrectly assume our request succeeded.
Note that the underlying connection is required to be reliable and does
retransmission itself. So it is safe to assume that if the transmission
fails, the device is in inconsistent state. Hence, we abort every request
if any output report fails. No need to verify which report failed.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Battery reports are sent along every status report of the Wii Remote.
So chances are pretty high that we have an up-to-date battery
cache at any time. Therefore, initialize the battery-cache to 100% and
then return battery values from the cache if the query fails.
This works around a power_supply limitation in that it requires us to be
able to query the device during power_supply registration and
removal. Otherwise, "add" or "remove" udev events are not sent. If
we answer these requests from our cache instead, we avoid dropping these
events and no longer cause warnings printed.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Nintendo produced many different devices that are internally based on the
Wii Remote protocol but provide different peripherals. To support these
devices, we need to schedule a device detection during initialization.
Device detection includes requesting a status report, reading extension
information and then evaluating which device we may be dealing with.
We currently detect gen1 and gen2 Wii Remote devices. All other devices
are marked as generic devices. More detections will be added later.
In followup patches we will be using these device IDs to control which
peripherals to initialize. For instance if a device is known to have no IR
camera, there is no need to provide the IR input device nor trying to
access IR registers. In fact, there are 3rd party devices that break if we
try things like this (hurray!).
The init_worker will be scheduled whenever we get hotplug events. This
isn't implemented, yet and will be added later. However, we need to make
sure that this worker can be called multiple times.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
We need constant I/O to keep the state up-to-date and not miss any
packets. Hence, call hid_hw_open() during setup and hid_hw_close() during
destruction.
These are no-ops for Bluetooth HIDP, but lets be safe.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
The output queue is independent of the other wiimote modules and can run
on its own. Therefore, move its members into a separate struct so we don't
run into name collisions with other modules.
This is only a syntactic change that renames all queue members to queue.*.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
The hid-wiimote driver supports more than the Wii Remote. Nintendo
produced many devices based on the Wii Remote, which have extension
devices built-in. It is not clear to many users, that these devices have
anything in common with the Wii Remote, so fix the driver description.
This also updates the copyright information for the coming hotplugging
rework.
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Pull s390 fixes from Martin Schwidefsky:
"Recent bug fixes, one of them touches a common code file.
It adds two #ifndef/#endif pairs to asm-generic/io.h to be able to
override xlate_dev_kmem_ptr and xlate_dev_mem_ptr."
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
s390/pgtable: Fix gmap notifier address
s390/dasd: fix handling of gone paths
s390/pgtable: Fix check for pgste/storage key handling
arch: s390: appldata: using strncpy() and strnlen() instead of sprintf()
s390/smp: lost IPIs on cpu hotplug
kernel: Fix s390 absolute memory access for /dev/mem
s390/dma: do not call debug_dma after free
Rework of per ring hangcheck made this obsolete.
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Keep track of ring seqno progress and if there are no
progress detected, declare hang. Use actual head (acthd)
to distinguish between ring stuck and batchbuffer looping
situation. Stuck ring will be kicked to trigger progress.
This commit adds a hard limit for batchbuffer completion time.
If batchbuffer completion time is more than 4.5 seconds,
the gpu will be declared hung.
Review comment from Ben which nicely clarifies the semantic change:
"Maybe I'm just stating the functional changes of the patch, but in case
they were unintended here is what I see as potential issues:
1. "If ring B is waiting on ring A via semaphore, and ring A is making
progress, albeit slowly - the hangcheck will fire. The check will
determine that A is moving, however ring B will appear hung because
the ACTHD doesn't move. I honestly can't say if that's actually a
realistic problem to hit it probably implies the timeout value is too
low.
2. "There's also another corner case on the kick. If the seqno = 2
(though not stuck), and on the 3rd hangcheck, the ring is stuck, and
we try to kick it... we don't actually try to find out if the kick
helped"
v2: use atchd to detect stuck ring from loop (Ben Widawsky)
v3: Use acthd to check when ring needs kicking.
Declare hang on third time in order to give time for
kick_ring to take effect.
v4: Update commit msg
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Paste in Ben's review comment.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Pull cgroup fixes from Tejun Heo:
- Fix for yet another xattr bug which may lead to NULL deref.
- A subtle bug in for_each_descendant_pre(). This bug requires quite
specific conditions to trigger and isn't too likely to actually
happen in the wild, but maybe that just makes it that much more
nastier.
- A warning message added for silly cgroup re-mount (not -o remount,
but unmount followed by mount) behavior.
* 'for-3.10-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: warn about mismatching options of a new mount of an existing hierarchy
cgroup: fix a subtle bug in descendant pre-order walk
cgroup: initialize xattr before calling d_instantiate()
Pull libata changes from Tejun Heo:
"Nothing too interesting. PCI ID additions, some sata_rcar fixes and a
fringe bug fix for DMADIR handling which shouldn't affect any device
remotely modern."
* 'for-3.10-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata:
sata_rcar: fix interrupt handling
ahci: add an observed PCI ID for Marvell 88se9172 SATA controller
sata_rcar: clear STOP bit in bmdma_start() method
libata: make ata_exec_internal_sg honor DMADIR
ata_piix: add PCI IDs for Intel BayTail
libata: update "Maintained by:" tags
When declaring playback and capture capabilities check for both CODEC
side and CPU side support rather than only checking for CODEC side
support. While it is unusual some CPUs do have unidirectional DAIs.
Reported-by: Fabio Estevam <fabio.estevam@freescale.com>
Tested-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Since it will be used for the global bound/unbound list with full PPGTT,
this helps clarify things for upcoming code rework.
Recommended-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
If we properly keep track of the pages_pin_count, then when we later add
multiple address spaces, the put_pages doesn't need any special checks
to be able to perform it's job.
CC: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Rebased on top of the fix for stolen memory pinning.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The way the stolen handling works is we take a pin on the backing pages,
but we never actually get a reference to the bo. On freeing objects
allocated with stolen memory, the final unref will end up freeing the
object with pinned pages count left. To enable an assertion to catch
bugs in this code path, this patch cleans up that remaining pin.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
arch/microblaze/include/asm/uaccess.h:101:3:
warning: cast removes address space of expression
arch/microblaze/include/asm/uaccess.h:107:2:
warning: cast removes address space of expression
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
It's not terribly interesting to know that a parameter doesn't exist,
and it can get in the way of interesting messages, especially with the
staggered VECS merging as we've done.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
As the device_node pointer is not changed in of_get_display_timing and
parse_timing_property it can be a const pointer.
Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
apic->pending_events processing has a race that may cause INIT and
SIPI
processing to be reordered:
vpu0: vcpu1:
set INIT
test_and_clear_bit(KVM_APIC_INIT)
process INIT
set INIT
set SIPI
test_and_clear_bit(KVM_APIC_SIPI)
process SIPI
At the end INIT is left pending in pending_events. The following patch
fixes this by latching pending event before processing them.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
The x86-64 extended low-byte registers were fetched correctly from reg,
but not from mod/rm.
This fixes another bug in the boot of RHEL5.9 64-bit, but it is still
not enough.
Cc: <stable@vger.kernel.org> # 3.9
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
