The gmu driver is initialized and cleaned up with calls from the gpu driver. As
such, the platform device stays valid after a6xx_gmu_remove is called and the
device managed resources are not freed. In the case of gpu probe failures or
unbind, these resources will remain managed.
If the gpu bind is run again (eg: if there's a probe defer somewhere in msm),
these resources will be initialized again for the same device, creating multiple
references. In the case of irqs, this causes failures since the irqs are
not shared (nor should they be).
This patch removes all devm_* calls and manually cleans things up in
gmu_remove.
Changes in v2:
- Add iounmap and free_irq to gmu_probe error paths
Cc: Jordan Crouse <jcrouse@codeaurora.org>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20190523171653.138678-4-sean@poorly.run
KVM has helpers to handle the condition codes of trapped aarch32
instructions. These are marked __hyp_text and used from HYP, but they
aren't built by the 'hyp' Makefile, which has all the runes to avoid ASAN
and KCOV instrumentation.
Move this code to a new hyp/aarch32.c to avoid a hyp-panic when starting
an aarch32 guest on a host built with the ASAN/KCOV debug options.
Fixes: 021234ef37 ("KVM: arm64: Make kvm_condition_valid32() accessible from EL2")
Fixes: 8cebe750c4 ("arm64: KVM: Make kvm_skip_instr32 available to HYP")
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
KVM's pmu.c contains the __hyp_text needed to switch the pmu registers
between host and guest. Because this isn't covered by the 'hyp' Makefile,
it can be built with kasan and friends when these are enabled in Kconfig.
When starting a guest, this results in:
| Kernel panic - not syncing: HYP panic:
| PS:a00003c9 PC:000083000028ada0 ESR:86000007
| FAR:000083000028ada0 HPFAR:0000000029df5300 PAR:0000000000000000
| VCPU:000000004e10b7d6
| CPU: 0 PID: 3088 Comm: qemu-system-aar Not tainted 5.2.0-rc1 #11026
| Hardware name: ARM LTD ARM Juno Development Platform/ARM Juno Development Plat
| Call trace:
| dump_backtrace+0x0/0x200
| show_stack+0x20/0x30
| dump_stack+0xec/0x158
| panic+0x1ec/0x420
| panic+0x0/0x420
| SMP: stopping secondary CPUs
| Kernel Offset: disabled
| CPU features: 0x002,25006082
| Memory Limit: none
| ---[ end Kernel panic - not syncing: HYP panic:
This is caused by functions in pmu.c calling the instrumented
code, which isn't mapped to hyp. From objdump -r:
| RELOCATION RECORDS FOR [.hyp.text]:
| OFFSET TYPE VALUE
| 0000000000000010 R_AARCH64_CALL26 __sanitizer_cov_trace_pc
| 0000000000000018 R_AARCH64_CALL26 __asan_load4_noabort
| 0000000000000024 R_AARCH64_CALL26 __asan_load4_noabort
Move the affected code to a new file under 'hyp's Makefile.
Fixes: 3d91befbb3 ("arm64: KVM: Enable !VHE support for :G/:H perf event modifiers")
Cc: Andrew Murray <Andrew.Murray@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
I no longer have time to actively review patches and manage the tree and
it's time to make that official.
Huge thanks to the incredible Linux community and all the contributors
who have put up with me over the past years.
I also take this opportunity to remove the website link to the Columbia
web page, as that information is no longer up to date and I don't know
who manages that anymore.
Signed-off-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
The driver checks for gmu->mmio as a sign that the device has been
initialized, however there are failures in probe below the mmio init.
If one of those is hit, mmio will be non-null but freed.
In that case, a6xx_gmu_probe will return an error to a6xx_gpu_init which
will in turn call a6xx_gmu_remove which checks gmu->mmio and tries to free
resources for a second time. This causes a great boom.
Fix this by adding an initialized member to gmu which is set on
successful probe and cleared on removal.
Changes in v2:
- None
Cc: Jordan Crouse <jcrouse@codeaurora.org>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20190523171653.138678-1-sean@poorly.run
Variable 'dmtimer_ops' was declared const static instead of static const.
../drivers/clocksource/timer-ti-dm.c:899:1: warning: ‘static’ is not at beginning of declaration [-Wold-style-declaration]
const static struct omap_dm_timer_ops dmtimer_ops = {
^~~~~
Signed-off-by: Philippe Mazenauer <philippe.mazenauer@outlook.de>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
The SUNXI_WATCHDOG option is required to make the
watchdog available on Allwinner H6.
Enable this option as a module.
Signed-off-by: Clément Péron <peron.clem@gmail.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Allwinner H6 has a r_watchdog similar to A64.
Declare it in the device-tree.
Signed-off-by: Clément Péron <peron.clem@gmail.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Allwinner H6 has a watchog node which seems broken
on some boards.
Test has been performed on several boards.
Chen-Yu Tsai boards:
Pine H64 - H6448BA 7782 => OK
OrangePi Lite 2 - H8068BA 61C2 => KO
Martin Ayotte boards:
Pine H64 - H8069BA 6892 => OK
OrangePi 3 - HA047BA 69W2 => KO
OrangePi One Plus - H7310BA 6842 => KO
OrangePi Lite2 - H6448BA 6662 => KO
Clément Péron board:
Beelink GS1 - H7309BA 6842 => KO
As it seems not fixable for now, declare the node
but leave it disable with a comment.
Signed-off-by: Clément Péron <peron.clem@gmail.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Allwinner H6 has a similar watchdog as the A64 which is already
a compatible of the A31.
This commit add the H6 compatible.
Signed-off-by: Clément Péron <peron.clem@gmail.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The call to of_parse_phandle returns a node pointer with refcount
incremented thus it must be explicitly decremented after the last
usage.
Detected by coccinelle with the following warnings:
drivers/media/platform/qcom/venus/firmware.c:90:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 82, but without a corresponding object release within this function.
drivers/media/platform/qcom/venus/firmware.c:94:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 82, but without a corresponding object release within this function.
drivers/media/platform/qcom/venus/firmware.c:128:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 82, but without a corresponding object release within this function.
Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
Acked-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Drop the reference to "parallel-ports" and remote endpoint's parent
nodes obtained previously with of_get_child_by_name() and
of_get_parent() respectively.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
'default n' is the default value for any bool or tristate Kconfig
setting so there is no need to write it explicitly.
Also since commit f467c5640c ("kconfig: only write '# CONFIG_FOO
is not set' for visible symbols") the Kconfig behavior is the same
regardless of 'default n' being present or not:
...
One side effect of (and the main motivation for) this change is making
the following two definitions behave exactly the same:
config FOO
bool
config FOO
bool
default n
With this change, neither of these will generate a
'# CONFIG_FOO is not set' line (assuming FOO isn't selected/implied).
That might make it clearer to people that a bare 'default n' is
redundant.
...
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Removing link components results in topology unloading. So,
acquire the client_mutex before removing components in
soc_remove_link_components. This will prevent the lockdep warning
seen when dai links are removed during topology removal.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Revert commit 069d037aea ("ASoC: simple-card: Fix configuration of
DAI format"). During further review, it turns out that the actual issue
was caused by an incorrectly formatted device-tree node describing the
soundcard.
The following is incorrect because the simple-audio-card
'bitclock-master' and 'frame-master' properties should not reference the
actual codec phandle ...
sound {
compatible = "simple-audio-card";
...
=> simple-audio-card,bitclock-master = <&codec>;
=> simple-audio-card,frame-master = <&codec>;
...
simple-audio-card,cpu {
sound-dai = <&xxx>;
};
simple-audio-card,codec {
=> sound-dai = <&codec>;
};
};
Rather, these properties should reference the phandle to the
'simple-audio-card,codec' property as shown below ...
sound {
compatible = "simple-audio-card";
...
=> simple-audio-card,bitclock-master = <&codec>;
=> simple-audio-card,frame-master = <&codec>;
...
simple-audio-card,cpu {
sound-dai = <&xxx>;
};
=> codec: simple-audio-card,codec { /* simple-card wants here */
sound-dai = <&xxx>; /* not here */
};
};
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Just use a local variable *ena_gpiod in slg51000_of_parse_cb instead.
With this change, the struct slg51000_pdata can be removed.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
These regulator_ops variables never need to be modified, make them const so
compiler can put them to .rodata.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Make use of the struct_size() helper instead of an open-coded version
in order to avoid any potential type mistakes, in particular in the
context in which this code is being used.
So, replace code of the following form:
sizeof(*fwk) + fwk->num_services * sizeof(fwk->svc_api_info[0]);
with:
struct_size(fwk, svc_api_info, fwk->num_services)
and so on...
Notice that variables bytes and len are unnecessary, hence they are
removed.
This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Make use of the struct_size() helper instead of an open-coded version
in order to avoid any potential type mistakes, in particular in the
context in which this code is being used.
So, replace the following form:
uuid_params->num_modules * sizeof(struct skl_mod_inst_map) + sizeof(uuid_params->num_modules)
with:
struct_size(params, u.map, uuid_params->num_modules)
and so on...
This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
One of the more common cases of allocation size calculations is finding the
size of a structure that has a zero-sized array at the end, along with
memory for some number of elements for that array. For example:
struct foo {
int stuff;
struct boo entry[];
};
instance = krealloc(instance, sizeof(struct foo) + count * sizeof(struct boo), GFP_KERNEL);
Instead of leaving these open-coded and prone to type mistakes, use the new
struct_size() helper:
instance = krealloc(instance, struct_size(instance, entry, count), GFP_KERNEL);
This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
We've used a standard interface for machine drivers for some time now,
there is no need for this dependency on a Skylake-specific header
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Add pinconf support for Bitmain BM1880 SoC. Pinconf support includes
pin bias, slew rate and schmitt trigger. Drive strength support will
be added later.
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Rework the BM1880 SoC pinmux handling by removing the
BM1880_PINMUX_FUNCTION_MUX define and merging it with the
BM1880_PINMUX_FUNCTION definition. Since the PWM muxing is handled by
generic pin controller in the SoC itself, there is no need to have a
dedicated code to do the muxing in PWM registers. So, lets club all
pinmux handling in the same per pin mux handling code.
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Earlier, the PWM registers were included as part of the pinctrl memory
map, but this turned to be useless as the muxing is being handled by the
SoC pin controller itself. So, lets modify the pinctrl memory map to
reflect the same.
Fixes: 07b734fbde ("dt-bindings: pinctrl: Add BM1880 pinctrl binding")
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Add pinconf support for PIN_CONFIG_OUTPUT_ENABLE and PIN_CONFIG_OUTPUT
in the meson pinctrl driver.
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
add support for the pinconf DT property output-enable, output-disable,
output-low and output-high in the meson pinctrl driver.
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
GPIOs are split between several banks (A, B, ...) and each bank can have
up to 16 lines. Those GPIOs could be used as interrupt lines thanks to
exti lines. As there are only 16 exti lines, a mux is used to select which
gpio line is connected to which exti line. Mapping is done as follow:
-A0, B0, C0.. -->exti_line_0 (X0 selected by mux_0)
-A1, B1, C1.. -->exti_line_1 (X1 selected by mux_1)
...
This patch adds a protection to avoid overriding on mux_n for exti_line_n.
Signed-off-by: Alexandre Torgue <alexandre.torgue@st.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
During power sequence, GPIO hardware registers could be lost if the power
supply is switched off. Each device using pinctrl API is in charge of
managing pins during suspend/resume sequences. But for pins used as gpio or
irq stm32 pinctrl driver has to save the hardware configuration.
Signed-off-by: Alexandre Torgue <alexandre.torgue@st.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This reverts commit 55143dc23c.
This causes build breakags with some Kconfigs so revert for now.
Fixes: 55143dc23c ("drm/amd/display: Don't load DMCU for Raven 1")
Signed-off-by: Dave Airlie <airlied@redhat.com>
All currently known ECs in the wild are very sensitive to timing.
Specifically the ECs are known to drop a transfer if more than 8 ms
passes from the assertion of the chip select until the transfer
finishes.
Let's use the new feature introduced in the patch (spi: Allow SPI
devices to request the pumping thread be realtime") to request the SPI
pumping thread be realtime. This means that if we get shunted off to
the SPI thread for whatever reason we won't get downgraded to low
priority.
NOTES:
- We still need to keep ourselves as high priority since the SPI core
doesn't guarantee that all transfers end up on the pumping thread
(in fact, it tries pretty hard to do them in the calling context).
- If future Chrome OS ECs ever fix themselves to be less sensitive
then we could consider adding a property (or compatible string) to
not set this property. For now we need it across the board.
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Guenter Roeck <groeck@chromium.org>
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Immutable branch between spi and chrome-platform for driver changes to
allow setting spi pump to RT priority.
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
As per the current design, in the case of sw crypto controlled devices,
it is the device which advertises the support for AP/VLAN iftype based
on it's ability to tranmsit packets encrypted in software
(In VLAN functionality, group traffic generated for a specific
VLAN group is always encrypted in software). Commit db3bdcb9c3
("mac80211: allow AP_VLAN operation on crypto controlled devices")
has introduced this change.
Since 4addr AP operation also uses AP/VLAN iftype, this conditional
way of advertising AP/VLAN support has broken 4addr AP mode operation on
crypto controlled devices which do not support VLAN functionality.
In the case of ath10k driver, not all firmwares have support for VLAN
functionality but all can support 4addr AP operation. Because AP/VLAN
support is not advertised for these devices, 4addr AP operations are
also blocked.
Fix this by allowing 4addr operation on devices which do not support
AP/VLAN iftype but can support 4addr AP operation (decision is based on
the wiphy flag WIPHY_FLAG_4ADDR_AP).
Cc: stable@vger.kernel.org
Fixes: db3bdcb9c3 ("mac80211: allow AP_VLAN operation on crypto controlled devices")
Signed-off-by: Manikanta Pubbisetty <mpubbise@codeaurora.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.
This patch fixes the following warning:
drivers/net/wireless/mac80211_hwsim.c: In function ‘init_mac80211_hwsim’:
drivers/net/wireless/mac80211_hwsim.c:3853:21: warning: this statement may fall through [-Wimplicit-fallthrough=]
param.reg_strict = true;
~~~~~~~~~~~~~~~~~^~~~~~
drivers/net/wireless/mac80211_hwsim.c:3854:3: note: here
case HWSIM_REGTEST_DRIVER_REG_ALL:
^~~~
Warning level 3 was used: -Wimplicit-fallthrough=3
This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The reported rate is not scaled down correctly. After applying this patch,
the function will behave just like the v/ht equivalents.
Signed-off-by: Shashidhar Lakkavalli <slakkavalli@datto.com>
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Fixes gcc '-Wunused-but-set-variable' warning:
net/mac80211/key.c: In function 'ieee80211_set_tx_key':
net/mac80211/key.c:271:24: warning:
variable 'old' set but not used [-Wunused-but-set-variable]
It is not used since introduction in
commit 96fc6efb9a ("mac80211: IEEE 802.11 Extended Key ID support")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
When receiving a deauthentication/disassociation frame from a TDLS
peer, a station should not disconnect the current AP, but only
disable the current TDLS link if it's enabled.
Without this change, a TDLS issue can be reproduced by following the
steps as below:
1. STA-1 and STA-2 are connected to AP, bidirection traffic is running
between STA-1 and STA-2.
2. Set up TDLS link between STA-1 and STA-2, stay for a while, then
teardown TDLS link.
3. Repeat step #2 and monitor the connection between STA and AP.
During the test, one STA may send a deauthentication/disassociation
frame to another, after TDLS teardown, with reason code 6/7, which
means: Class 2/3 frame received from nonassociated STA.
On receive this frame, the receiver STA will disconnect the current
AP and then reconnect. It's not a expected behavior, purpose of this
frame should be disabling the TDLS link, not the link with AP.
Cc: stable@vger.kernel.org
Signed-off-by: Yu Wang <yyuwang@codeaurora.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
As a result of heavy refactoring based on IR context header, the packet
handler becomes simpler.
This commit merges the packet handler into function for IR context
callback. The logic to parse IR context header and tracepoints event is
split to a function.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
As a result of heavy refactoring based on IT packet header, the packet
handler becomes simpler.
This commit merges the packet handler into function for IT context
callback. The logic to build IT packet header and tracepoints event is
split to a function.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The value of data block counter is not calculated for incoming packet
without CIP header. This commit fixes the bug.
Fixes: 947b437e12 ("ALSA: firewire-lib: unify packet handler for IR context")
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The gpio-adp5588 driver uses interfaces that are provided by
GPIOLIB_IRQCHIP, so select that symbol in its Kconfig entry.
Fixes these build errors:
../drivers/gpio/gpio-adp5588.c: In function ‘adp5588_irq_handler’:
../drivers/gpio/gpio-adp5588.c:266:26: error: ‘struct gpio_chip’ has no member named ‘irq’
dev->gpio_chip.irq.domain, gpio));
^
../drivers/gpio/gpio-adp5588.c: In function ‘adp5588_irq_setup’:
../drivers/gpio/gpio-adp5588.c:298:2: error: implicit declaration of function ‘gpiochip_irqchip_add_nested’ [-Werror=implicit-function-declaration]
ret = gpiochip_irqchip_add_nested(&dev->gpio_chip,
^
../drivers/gpio/gpio-adp5588.c:307:2: error: implicit declaration of function ‘gpiochip_set_nested_irqchip’ [-Werror=implicit-function-declaration]
gpiochip_set_nested_irqchip(&dev->gpio_chip,
^
Fixes: 459773ae8d ("gpio: adp5588-gpio: support interrupt controller")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: linux-gpio@vger.kernel.org
Reviewed-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Acked-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Setting bit5 (headerless msg for preemptible GPGPU context) of SAMPLER_MODE
register to enable support for the headless msgs on gen11. None of existing
use cases will be affected by this as this change makes both types of
message - headerless and w/ header supported at the same time. It also
complies with the new recommendation for the default bit value for the
next gen.
v2: rewrote commit message to include more information
v3: setting the bit in icl_ctx_workarounds_init()
Signed-off-by: Dongwon Kim <dongwon.kim@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190425055005.21790-1-chris@chris-wilson.co.uk
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Having deferred the vma destruction to a worker where we can acquire the
struct_mutex, we have to avoid chasing back into the now destroyed
ppgtt. The pd_vma is special in having a custom unbind function to scan
for unused pages despite the VMA itself being notionally part of the
GGTT. As such, we need to disable that callback to avoid a
use-after-free.
This unfortunately blew up so early during boot that CI declared the
machine unreachable as opposed to being the major failure it was. Oops.
Fixes: d3622099c7 ("drm/i915/gtt: Always acquire struct_mutex for gen6_ppgtt_cleanup")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Tomi Sarvela <tomi.p.sarvela@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190524064529.20514-1-chris@chris-wilson.co.uk
