This takes hw constraints specific to pixel formats into account,
e.g., the inability of older hw to scale fp16 format framebuffers.
It should now allow safely to enable fp16 formats also on DCE-8,
DCE-10, DCE-11.0
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
And drop it when we detach. If the shared buffer is in vram,
we need to make sure we don't put the device into runtime
suspend.
Acked-by: Shashank Sharma <shashank.sharma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We need these to support PSR on DCN302
Signed-off-by: Joshua Aberback <joshua.aberback@amd.com>
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
We can only use this feature when the display is idle. When active vblank
irq count is 0 we know all the displays are idle.
[How]
-Add a active vblank irq counter
-Update the counter when we enable/disable vblank irq
-if vblank irq count is 0 we can consider mall stutter
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Nick Kazlauskas <Nicholas.Kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[How]
- use dc interface instead of hwss interface in cursor functions, to keep
dc->idle_optimizations_allowed updated
- add dc interface to check if idle optimizations might apply to a plane
Signed-off-by: Joshua Aberback <joshua.aberback@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Some of the newly added code is hidden inside of #ifdef
blocks, but one variable is unused when debugfs is disabled:
drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm.c:8370:8: error: unused variable 'configure_crc' [-Werror,-Wunused-variable]
Change the #ifdef to an if(IS_ENABLED()) check to fix the warning
and avoid adding more #ifdefs.
Fixes: c920888c60 ("drm/amd/display: Expose new CRC window property")
Reviewed-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fix the following coccinelle report:
drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c:1357:6-8:
WARNING: possible condition with no effect (if == else)
Signed-off-by: Tian Tao <tiantao6@hisilicon.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Remove including <linux/version.h> that don't need it.
Signed-off-by: Tian Tao <tiantao6@hisilicon.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to improve the fine grain tuning function for RV/RV2/PCO.
The fine grain tuning function uses the sysfs node -- pp_od_clk_voltage
to config gfxclk. Meanwhile, another sysfs
node -- power_dpm_force_perfomance_level also affects the gfx clk.
It will cause confusion when these two sysfs nodes works
together. So this patch adds one flag to avoid this confusion, the flag
will make these two sysfs nodes work separately.
The flag is set as "disabled" by default, so the fine grain tuning function
will be disabled by default.
Only when power_dpm_force_perfomance_level is changed to
"manual" mode, the flag will be set as "enabled",
and the fine grain tuning function will be enabled.
In other profile modes, including "auto", "high", "low",
"profile_peak", "profile_standard", "profile_min_sclk",
"profile_min_mclk", the flag will be set as "disabled",
and the od range of fine grain tuning function will
be restored default value.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to enable dclk/vclk dpm function in "auto" power profile
for vangogh.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In the psp command submit logic,
the function msleep(1) delayed too long,
Changing it to usleep_range(10, 100) to
have a better performance.
Signed-off-by: Peng Ju Zhou <PengJu.Zhou@amd.com>
Reviewed-by: Emily.Deng <Emily.Deng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to fix the failure when change power profile to
"profile_peak" for renoir.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Remove hdp_flush function from amdgpu_nbio struct as it have been unified
into hdp struct.
Remove the include about hdp register which was not used.
V2: Remove hdp golden setting which is unnecessary.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Switch to use the HDP functions which unified on hdp structure instead of
the scattered hdp callback functions.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Switch to use the HDP functions which unified on hdp structure instead of
the scattered hdp callback functions.
V2: clean up hdp reset ras error count function.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Unify hdp related function into hdp structure for hdp version 5.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Unify hdp related function into hdp structure for hdp version 4.
V2: Add function to remap hdp registers for nbio version 6.1
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
amdgpu_hdp hold all the callbacks for hdp
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Asic funcs pointer needs being checked whether is NULL. Because when
check_atom_bios() fails in req_init_data handshake, asic funcs pointer
is not initialized yet.
Signed-off-by: Jiawei Gu <Jiawei.Gu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On modern gpus, GTT (system memory) works as well here, and this may
also be a workaround for platforms which cannot map vram correctly.
Signed-off-by: chenli <chenli@uniontech.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When GFXOFF is enabled and GPU is idle, driver will fail to access some
registers. Therefore change to disable power gating before all access
registers with MMIO.
Dmesg log is as following:
amdgpu 0000:03:00.0: amdgpu: amdgpu: finishing device.
amdgpu: cp queue pipe 4 queue 0 preemption failed
amdgpu 0000:03:00.0: amdgpu: failed to write reg 2890 wait reg 28a2
amdgpu 0000:03:00.0: amdgpu: failed to write reg 1a6f4 wait reg 1a706
amdgpu 0000:03:00.0: amdgpu: failed to write reg 2890 wait reg 28a2
amdgpu 0000:03:00.0: amdgpu: failed to write reg 1a6f4 wait reg 1a706
Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
global noretry setting is cached to gmc.noretry
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For sdma5.2, all sdma instances will share the same fw,
remove unnecessary asic check to be more generic.
Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to add support to umd P-state function for vangogh.
It enables the "set" function of 3 sysfs nodes: pp_dpm_mclk,
pp_dpm_fclk, pp_dpm_socclk, the functions is used to set the DPM
frequency level of memclk/fclk/socclk.
Due to only after enabling the "power_dpm_force_performance_level"
sysfs node, it is allowed to set these three nodes,
so this patch also enables the "powe_dpm_force_performance_level"
sysfs node, which is used to change power profile.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to add some basic functions to support
umd P-state function for vangogh.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to add some basic functions to support umd
P-state function for vangogh.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to enable the fine grain tuning function for renoir.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to add the fine grain tuning function for renoir.
This function uses the pp_od_clk_voltage sysfs file to configure the min
and max value of gfx clock frequency manually or restore the default value.
Command guide:
echo "s level value" > pp_od_clk_voltage
"s" - set the sclk frequency
"level" - 0 or 1, "0" represents the min value, "1" represents
the max value
"value" - the target value of sclk frequency, it should
be limited in the safe range
echo "r" > pp_od_clk_voltage
"r" - reset the sclk frequency, restore the default value instantly
echo "c" > pp_od_clk_voltage
"c" - commit the min and max value of sclk frequency to the system
only after the commit command, the target values set by "s" command
will take effect.
Example:
1)change power profile from "auto" to "standard"
$ cat power_dpm_force_performance_level
auto
$ echo "profile_standard" > power_dpm_force_performance_level
$ cat power_dpm_force_performance_level
profile_standard
2)check the default sclk frequency
$ cat pp_od_clk_voltage
OD_SCLK:
0: 200Mhz
1: 1400Mhz
OD_RANGE:
SCLK: 200MHz 1400MHz
3)use "s" -- set command to configure the min and max sclk frequency
$ echo "s 0 600" > pp_od_clk_voltage
$ echo "s 1 1000" > pp_od_clk_voltage
$ echo "c" > pp_od_clk_voltage
$ cat pp_od_clk_voltage
OD_SCLK:
0: 600Mhz
1: 1000Mhz
OD_RANGE:
SCLK: 200MHz 1400MHz
4)use "r" -- reset command to restore the min or max sclk frequency
$ echo "r" > pp_od_clk_voltage
$ cat pp_od_clk_voltage
OD_SCLK:
0: 200Mhz
1: 1400Mhz
OD_RANGE:
SCLK: 200MHz 1400MHz
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
sienna cichlid interfaces with RAS eeprom on I2C controller port 1
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to improve the fine grain tuning function for RV/RV2/PCO.
This patch adds two new commands: "restore" and "commit".
This function uses the pp_od_clk_voltage sysfs file to configure the min
and max value of gfx clock frequency manually or restore the default value.
Command guide:
echo "s level value" > pp_od_clk_voltage
"s" - set the sclk frequency
"level" - 0 or 1, "0" represents the min value, "1" represents
the max value
"value" - the target value of sclk frequency, it should be limited in the
safe range
echo "r" > pp_od_clk_voltage
"r" - reset the sclk frequency, restore the default value instantly
echo "c" > pp_od_clk_voltage
"c" - commit the min and max value of sclk frequency to the system
only after the commit command, the target values set by "s" command
will take effect.
Example:
1)change power profile from "auto" to "manual"
$ cat power_dpm_force_performance_level
auto
$ echo "manual" > power_dpm_force_performance_level
$ cat power_dpm_force_performance_level
manual
2)check the default sclk frequency
$ cat pp_od_clk_voltage
OD_SCLK:
0: 200Mhz
1: 1400Mhz
OD_RANGE:
SCLK: 200MHz 1400MHz
3)use "s" -- set command to configure the min and max sclk frequency
$ echo "s 0 600" > pp_od_clk_voltage
$ echo "s 1 1000" > pp_od_clk_voltage
$ echo "c" > pp_od_clk_voltage
$ cat pp_od_clk_voltage
OD_SCLK:
0: 600Mhz
1: 1000Mhz
OD_RANGE:
SCLK: 200MHz 1400MHz
4)use "r" -- reset command to restore the min or max sclk frequency
$ echo "r" > pp_od_clk_voltage
$ cat pp_od_clk_voltage
OD_SCLK:
0: 200Mhz
1: 1400Mhz
OD_RANGE:
SCLK: 200MHz 1400MHz
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to enable the "fetch" function of pp_dpm_vclk and
pp_dpm_dclk.
It allows to fetch the current frequency of vcn and dcn and their
DPM levels for vangogh.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to add two new sysfs nodes for vangogh:
pp_dpm_dclk and pp_dpm_vclk.
The two sysfs nodes are similar to pp_dpm_fclk/memclk/socclk.
pp_dpm_dclk represents the DPM frequency of dcn unit.
pp_dpm_vclk represents the DPM frequency of vcn unit.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to add new feature map macros to resolve duplicate name.
Vangogh uses one different format to name some feature bits of swSMU, it
causes some duplicate name in the existing feature map list.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to add supoort to umd P-state function for vangogh.
It enables the "fetch" function of 3 sysfs nodes: pp_dpm_mclk,
pp_dpm_fclk, pp_dpm_socclk,the function is used to fetch
the current frequency of memclk/fclk/socclk.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to correct the sensor value of power for vangogh.
Signed-off-by: Xiaojian Du <Xiaojian.Du@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
It programs Plane's calculated x, y, offset to Plane SF register.
It does the calculation of x and y offsets using
skl_calc_main_surface_offset().
v3: Update commit message
Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Tested-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210104205654.238928-4-jose.souza@intel.com
The calculation the offsets of the main surface will be needed by PSR2
selective fetch code so here splitting and exporting it.
No functional changes were done here.
v3: Rebased
Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Tested-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210104205654.238928-3-jose.souza@intel.com
Now using plane damage clips property to calcualte the damaged area.
Selective fetch only supports one region to be fetched so software
needs to calculate a bounding box around all damage clips.
Now that we are not complete fetching each plane, there is another
loop needed as all the plane areas that intersect with the pipe
damaged area needs to be fetched from memory so the complete blending
of all planes can happen.
v2:
- do not shifting new_plane_state->uapi.dst only src is in 16.16 format
v4:
- setting plane selective fetch area using the whole pipe damage area
- mark the whole plane area damaged if plane visibility or alpha
changed
v5:
- taking in consideration src.y1 in the damage coordinates
- adding to the pipe damaged area planes that were visible but are
invisible in the new state
v6:
- consider old state plane coordinates when visibility changes or it
moved to calculate damaged area
- remove from damaged area the portion not in src clip
v7:
- intersec every damage clip with src to minimize damaged area
v8:
- adjust pipe_damaged area to 4 lines grouping
- adjust calculation now that is understood that uapi.src is the
framebuffer coordinates that plane will start to fetch from
v9:
- Only add plane dst or src to damaged_area if visible
- Early skip plane damage calculation if it was not visible in old and
new state
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210104205654.238928-2-jose.souza@intel.com
In generating the reference LRC, we want a page-aligned address for
simplicity in computing the offsets within. This then shares the
computation for the HW LRC which is mapped and so page aligned, making
the comparison straightforward. It seems that kmalloc(4k) is not always
returning from a 4k-aligned slab cache (which would give us a page aligned
address) so force alignment by explicitly allocating a page.
Reported-by: "Gote, Nitin R" <nitin.r.gote@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: "Gote, Nitin R" <nitin.r.gote@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200923114156.17749-1-chris@chris-wilson.co.uk
Older platforms use rawclk to derive the CS clock. rawclk is being
determined during intel_device_info_init(), and so that needs to be
pushed slightly earlier.
Fixes: f170523a7b ("drm/i915/gt: Consolidate the CS timestamp clocks")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210104115145.24460-2-chris@chris-wilson.co.uk
Because clk_disable_unprepare() already checked NULL clock parameter,
so the additional check is unnecessary, just remove them.
Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Because clk_disable_unprepare() already checked NULL clock parameter,
so the additional check is unnecessary, just remove them.
Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
A non-const structure containing function pointers is a possible attack
vector. The drm_driver structure is already const in most drivers, but
there are a few exceptions. Constify the structure in the drivers that
don't need to modify at, as a low-hanging fruit. The rest of the drivers
will need a more complex fix.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Now that the legacy PCI support code doesn't need to write to the
drm_driver structure, it can be treated as const through the whole DRM
core, unconditionally. This allows declaring the structure as const in
all drivers, removing one possible attack vector.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The drm_driver structure contains a single field (legacy_dev_list) that
is modified by the DRM core, used to store a linked list of legacy DRM
devices associated with the driver. In order to make the structure
const, move the field out to a global variable. This requires locking
access to the global where the local field didn't require serialization,
but this only affects legacy drivers, and isn't in any hot path.
While at it, compile-out the legacy_dev_list field when DRM_LEGACY isn't
defined.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
The drm_gem_dumb_destroy() isn't used in drivers, don't export it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use the drm_plane_cleanup() function directly as the drm_plane_funcs
.destroy() handler without creating an unnecessary wrapper around it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use the drm_plane_cleanup() function directly as the drm_plane_funcs
.destroy() handler without creating an unnecessary wrapper around it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use the drm_plane_cleanup() function directly as the drm_plane_funcs
.destroy() handler without creating an unnecessary wrapper around it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use the drm_plane_cleanup() function directly as the drm_plane_funcs
.destroy() handler without creating an unnecessary wrapper around it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The device references acquired by of_find_device_by_node() are not
released by the driver. Fix this by registering a cleanup action.
Fixes: 8de707aeb4 ("drm: rcar-du: kms: Initialize CMM instances")
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
The local encoder variable is an alias for &renc->base, and is only use
twice. It doesn't help much, drop it, along with the
rcar_encoder_to_drm_encoder() macro that is then unused.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
The rcar-du driver skips registration of the encoder for the LVDS1
output when LVDS is used in dual-link mode, as the LVDS0 and LVDS1 links
are bundled and handled through the LVDS0 output. It however still
allocates the encoder and immediately destroys it, which is pointless.
Skip allocation of the encoder altogether in that case.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Now that drm_device is embedded in rcar_du_device, we can use
container_of to get the rcar_du_device pointer from the drm_device,
instead of using the drm_device.dev_private field.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Embedding drm_device in rcar_du_device allows usage of the DRM managed
API to allocate both structures in one go, simplifying error handling.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
devm_kzalloc() is the wrong API to allocate encoders, as the lifetime of
the encoders is tied to the DRM device, not the device to driver
binding. drmm_kzalloc() isn't a good option either, as it would result
in the encoder being freed before being unregistered during the managed
cleanup of the DRM objects. Use a plain kzalloc(), and register a drmm
action to cleanup the encoder.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
devm_kcalloc() is the wrong API to allocate planes, as the lifetime of
the planes is tied to the DRM device, not the device to driver
binding. drmm_kcalloc() isn't a good option either, as it would result
in the planes being freed before being unregistered during the managed
cleanup of the DRM objects. Use a plain kcalloc(), and cleanup the
planes and free the memory in the existing rcar_du_vsp_cleanup()
handler.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
The encoder->name field can never be non-null in the error path, as that
can only be possible after a successful call to
drm_simple_encoder_init(). Drop the cleanup.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Use drmm_add_action_or_reset() instead of drmm_add_action() to ensure
the vsp device reference is released in case the function call fails.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
On D3 and E3 platforms, the LVDS encoder includes a PLL that can
generate a clock for the corresponding CRTC, used even when the CRTC
output to a non-LVDS port. This mechanism is supported by the driver,
but the implementation is broken in dual-link LVDS mode. In that case,
the LVDS1 drm_encoder is skipped, which causes a crash when trying to
access its bridge later on.
Fix this by storing bridge pointers internally instead of retrieving
them from the encoder. The rcar_du_device encoders field isn't used
anymore and can be dropped.
Fixes: 8e8fddab0d ("drm: rcar-du: Skip LVDS1 output on Gen3 when using dual-link LVDS mode")
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
The DRM CRTC helpers add default modes to connectors in the connected
state if no mode can be retrieved from the connector. This behaviour is
useful for VGA or DVI outputs that have no connected DDC bus. However,
in such cases, the status of the output usually can't be retrieved and
is reported as connector_status_unknown.
Extend the addition of default modes to connectors in an unknown state
to support outputs that can retrieve neither the modes nor the
connection status.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
pm_runtime_get_sync will increment pm usage counter even it failed.
Forgetting to putting operation will result in a reference leak here.
A new function pm_runtime_resume_and_get is introduced in [0] to keep
usage counter balanced. So We fix the reference leak by replacing it
with new funtion.
[0] dd8088d5a8 ("PM: runtime: Add pm_runtime_resume_and_get to deal with usage counter")
Fixes: e08e934d6c ("drm: rcar-du: Add support for CMM")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Qinglang Miao <miaoqinglang@huawei.com>
Acked-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
The thc63lvd1024 driver requests a supply using regulator_get_optional()
but both the name of the supply and the usage pattern suggest that it is
being used for the main power for the device and is not at all optional
for the device for function, there is no handling at all for absent
supplies. Such regulators should use the vanilla regulator_get()
interface, it will ensure that even if a supply is not described in the
system integration one will be provided in software.
Signed-off-by: Mark Brown <broonie@kernel.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
In the near future, upstream will introduce a SZ_8G macro that is
slightly different to our own. Employ a temporary ifndef to avoid
compilation failure until we have backmerged.
References: 8b0fac44bd ("sizes.h: add SZ_8G/SZ_16G/SZ_32G macros")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210104171511.32684-1-chris@chris-wilson.co.uk
Use use drmm_crtc_alloc_with_planes() to align crtc memory lifetime
with the drm device. drm_crtc_cleanup() is called automatically before
the memory is freed.
Also use drmm_add_action_or_reset() to make sure IPU resources are
released automatically.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use drmm_universal_plane_alloc() to align plane memory lifetime with
the drm device. drm_plane_cleanup() is called automatically before the
memory is freed.
Also move the call to ipu_plane_get_resources() into ipu_plane_init()
and use drm managed resources to put IPU resources automatically when
required. Handle error return values of the plane property creation
functions.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use drmm_simple_encoder_alloc() to align encoder memory lifetime with
the drm device. drm_encoder_cleanup() is called automatically before
the memory is freed.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use drmm_simple_encoder_alloc() to align encoder memory lifetime with
the drm device. drm_encoder_cleanup() is called automatically before
the memory is freed.
Also fold imx_tve_register() into imx_tve_bind().
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use drmm_simple_encoder_alloc() to align encoder memory lifetime with
the drm device. drm_encoder_cleanup() is called automatically before
the memory is freed.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use drmm_simple_encoder_alloc() to align encoder memory lifetime with
the drm device. drm_encoder_cleanup() is called automatically.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Parts of the initialization that do not require the drm device can be
done once during probe instead of possibly multiple times during bind.
The bind function only creates the encoder.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use local variables for bridge and connector.
This simplifies the following commits.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Avoid leaking the clock provider when the driver is unbound.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Parts of the initialization that do not require the drm device can be
done once during probe instead of possibly multiple times during bind.
The bind function only creates the encoder.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Introduce local variables for encoder and connector.
This simplifies the following commits.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Parts of the initialization that do not require the drm device can be
done once during probe instead of possibly multiple times during bind.
The bind function only creates the encoders.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use a local variable for the connector.
This simplifies the following commits.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Parts of the initialization that do not require the drm device can be
done once during probe instead of possibly multiple times during bind.
The bind function only creates the encoder and attaches the bridge.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Add an alternative to drm_crtc_init_with_planes() that allocates
and initializes a crtc and registers drm_crtc_cleanup() with
drmm_add_action_or_reset().
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Add an alternative to drm_universal_plane_init() that allocates
and initializes a plane and registers drm_plane_cleanup() with
drmm_add_action_or_reset().
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Add an alternative to drm_simple_encoder_init() that allocates and
initializes a simple encoder and registers drm_encoder_cleanup() with
drmm_add_action_or_reset().
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Add an alternative to drm_encoder_init() that allocates and initializes
an encoder and registers drm_encoder_cleanup() with
drmm_add_action_or_reset().
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Simple managed encoders do not require the .destroy callback,
make the whole funcs structure optional.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Replace a comma between expression statements by a semicolon.
Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
The iMX DRM LVDS driver uses Common Clock Framework thus it cannot be
built on platforms without it (e.g. compile test on MIPS with RALINK and
SOC_RT305X):
/usr/bin/mips-linux-gnu-ld: drivers/gpu/drm/imx/imx-ldb.o: in function `imx_ldb_encoder_disable':
imx-ldb.c:(.text+0x400): undefined reference to `clk_set_parent'
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
They're taking up too much space on the stack.
Fixes the following W=1 kernel build warning(s):
drivers/gpu/ipu-v3/ipu-di.c: In function ‘ipu_di_sync_config_noninterlaced’:
drivers/gpu/ipu-v3/ipu-di.c:391:1: warning: the frame size of 1064 bytes is larger than 1024 bytes [-Wframe-larger-than=]
Cc: Philipp Zabel <p.zabel@pengutronix.de>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
A driver must not set drm_crtc.cursor and any of the legacy funcs at the
same time, otherwise it's not clear which one DRM core should use for
legacy cursor updates.
Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20201222134002.161613-1-contact@emersion.fr
Randconfig builds on 32-bit machines show lots of warnings for
the i915 driver for passing a 32-bit value into __const_hweight64():
drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c:2584:9: error: shift count >= width of type [-Werror,-Wshift-count-overflow]
return hweight64(VDBOX_MASK(&i915->gt));
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/asm-generic/bitops/const_hweight.h:29:49: note: expanded from macro 'hweight64'
#define hweight64(w) (__builtin_constant_p(w) ? __const_hweight64(w) : __arch_hweight64(w))
Change it to hweight_long() to avoid the warning.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210103135158.3591442-1-arnd@kernel.org
This allows us to remove pin_map from state allocation, which saves
us a few retry loops. We won't need this until first pin, anyway.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20201231170405.22843-1-chris@chris-wilson.co.uk
Let's modify the "workaround lost" error message slightly to make it
more clear what the various numbers represent. Also, the 'expected'
value needs to be &'d with wa->read so that it doesn't include the mask
bits for masked registers (those bits are write-only in the hardware and
will usually always read out as 0's).
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20201231191103.854519-1-matthew.d.roper@intel.com
Since we use a flag within i915_request.flags to indicate when we have
boosted the request (so that we only apply the boost) once, this can be
used as the serialisation with i915_request_retire() to avoid having to
explicitly take the i915_request.lock which is more heavily contended.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201231093149.19086-1-chris@chris-wilson.co.uk
We only need to evaluate the current status of the context when it is
scheduled in, we will force a reschedule when the context is closed
propagating the change to inflight contexts.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201231093946.11649-1-chris@chris-wilson.co.uk
Since we process schedule-in of a context after submitting the request,
if we decide to reset the context at that time, we also have to cancel
the requets we have marked for submission.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201230220028.17089-1-chris@chris-wilson.co.uk
Since multiple connectors may run intel_dp_aux_xfer conncurrently, a
single global pm_qos does not suffice. (One connector may disable the
dma-latency boost prematurely while the second is still depending on
it.) Instead of a single global pm_qos, track the pm_qos request for
each intel_dp.
v2: Move the pm_qos setup/teardown to intel_dp_aux_init/fini
Fixes: 9ee32fea5f ("drm/i915: irq-drive the dp aux communication")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201230202309.23982-1-chris@chris-wilson.co.uk
Declare that, under extreme circumstances, the shrinker may need to wait
upon a request, in which case reset must not itself deadlock in order to
ensure forward progress of the driver. That is since the shrinker may
depend upon a reset, any reset cannot touch the shrinker.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201229141626.4773-1-chris@chris-wilson.co.uk
The "max-brightness" is a standard backlight property that
we need to support for the Samsung GT-I8190 Golden because
the display will go black if we crank up the brightness
too high.
As the platform needs this ability to give picture this is
a regression fix along with the addition of the property
to the GT-I8190 device tree.
Cc: Stephan Gerhold <stephan@gerhold.net>
Fixes: 9c3f0a0dd6 ("drm/panel: s6e63m0: Implement 28 backlight levels")
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20201214222210.238081-1-linus.walleij@linaro.org
clang (quite rightly) complains fairly loudly about the newly added
mpc1_get_mpc_out_mux() function returning an uninitialized value if the
'opp_id' checks don't pass.
This may not happen in practice, but the code really shouldn't return
garbage if the sanity checks don't pass.
So just initialize 'val' to zero to avoid the issue.
Fixes: 110b055b28 ("drm/amd/display: add getter routine to retrieve mpcc mux")
Cc: Josip Pavic <Josip.Pavic@amd.com>
Cc: Bindu Ramamurthy <bindu.r@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
accesses, inefficient and disfunctional code. The goal is to remove the
export of irq_to_desc() to prevent these things from creeping up again.
-----BEGIN PGP SIGNATURE-----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=BXv+
-----END PGP SIGNATURE-----
Merge tag 'irq-core-2020-12-23' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq updates from Thomas Gleixner:
"This is the second attempt after the first one failed miserably and
got zapped to unblock the rest of the interrupt related patches.
A treewide cleanup of interrupt descriptor (ab)use with all sorts of
racy accesses, inefficient and disfunctional code. The goal is to
remove the export of irq_to_desc() to prevent these things from
creeping up again"
* tag 'irq-core-2020-12-23' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (30 commits)
genirq: Restrict export of irq_to_desc()
xen/events: Implement irq distribution
xen/events: Reduce irq_info:: Spurious_cnt storage size
xen/events: Only force affinity mask for percpu interrupts
xen/events: Use immediate affinity setting
xen/events: Remove disfunct affinity spreading
xen/events: Remove unused bind_evtchn_to_irq_lateeoi()
net/mlx5: Use effective interrupt affinity
net/mlx5: Replace irq_to_desc() abuse
net/mlx4: Use effective interrupt affinity
net/mlx4: Replace irq_to_desc() abuse
PCI: mobiveil: Use irq_data_get_irq_chip_data()
PCI: xilinx-nwl: Use irq_data_get_irq_chip_data()
NTB/msi: Use irq_has_action()
mfd: ab8500-debugfs: Remove the racy fiddling with irq_desc
pinctrl: nomadik: Use irq_has_action()
drm/i915/pmu: Replace open coded kstat_irqs() copy
drm/i915/lpe_audio: Remove pointless irq_to_desc() usage
s390/irq: Use irq_desc_kstat_cpu() in show_msi_interrupt()
parisc/irq: Use irq_desc_kstat_cpu() in show_interrupts()
...
Pull the individual strands of creating a custom heartbeat requests into
a pair of common functions. This will reduce the number of changes we
will need to make in future.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224160213.29521-1-chris@chris-wilson.co.uk
The reloc batch is short lived but can exist in the user visible ppGTT,
and since it's backed by an internal object, which lacks page clearing,
we should take care to clear it upfront.
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224151358.401345-2-matthew.auld@intel.com
Cc: stable@vger.kernel.org
The shadow batch is an internal object, which doesn't have any page
clearing, and since the batch_len can be smaller than the object, we
should take care to clear it.
Testcase: igt/gen9_exec_parse/shadow-peek
Fixes: 4f7af1948a ("drm/i915: Support ro ppgtt mapped cmdparser shadow buffers")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224151358.401345-1-matthew.auld@intel.com
Cc: stable@vger.kernel.org
Now that the tasklet completely controls scheduling of the requests, and
we postpone scheduling out the old requests, we can keep a hanging
virtual request bound to the engine on which it hung, and remove it from
te queue. On release, it will be returned to the same engine and remain
in its queue until it is scheduled; after which point it will become
eligible for transfer to a sibling. Instead, we could opt to resubmit the
request along the virtual engine on unhold, making it eligible for load
balancing immediately -- but that seems like a pointless optimisation
for a hanging context.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-8-chris@chris-wilson.co.uk
Having recognised that we do not change the sibling until we schedule
out, we can then defer the decision to resubmit the virtual engine from
the unwind of the active queue to scheduling out of the virtual context.
This improves our resilence in virtual engine scheduling, and should
eliminate the rare cases of gem_exec_balance failing.
By keeping the unwind order intact on the local engine, we can preserve
data dependency ordering while doing a preempt-to-busy pass until we
have determined the new ELSP. This means that if we try to timeslice
between a virtual engine and a data-dependent ordinary request, the pair
will maintain their relative ordering and we will avoid the
resubmission, cancelling the timeslicing until further change.
The dilemma though is that we then may end up in a situation where the
'demotion' of the virtual request to an ordinary request in the engine
queue results in filling the ELSP[] with virtual requests instead of
spreading the load across the engines. To compensate for this, we mark
each virtual request and refuse to resubmit a virtual request in the
secondary ELSP slots, thus forcing subsequent virtual requests to be
scheduled out after timeslicing. By delaying the decision until we
schedule out, we will avoid unnecessary resubmission.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2079
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2098
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-7-chris@chris-wilson.co.uk
Let's only wait for the list iterator when decoupling the virtual
breadcrumb, as the signaling of all the requests may take a long time,
during which we do not want to keep the tasklet spinning.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-6-chris@chris-wilson.co.uk
The issue with stale virtual breadcrumbs remain. Now we have the problem
that if the irq-signaler is still referencing the stale breadcrumb as we
transfer it to a new sibling, the list becomes spaghetti. This is a very
small window, but that doesn't stop it being hit infrequently. To
prevent the lists being tangled (the iterator starting on one engine's
b->signalers but walking onto another list), always decouple the virtual
breadcrumb on schedule-out and make sure that the walker has stepped out
of the lists.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-5-chris@chris-wilson.co.uk
Inside schedule_out, we do extra work upon idling the context, such as
updating the runtime, kicking off retires, kicking virtual engines.
However, if we are in a series of processing single requests per
contexts, we may find ourselves scheduling out the context, only to
immediately schedule it back in during dequeue. This is just extra work
that we can avoid if we keep the context marked as inflight across the
dequeue. This becomes more significant later on for minimising virtual
engine misses.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-4-chris@chris-wilson.co.uk
Once a virtual engine has been bound to a sibling, it will remain bound
until we finally schedule out the last active request. We can not rebind
the context to a new sibling while it is inflight as the context save
will conflict, hence we wait. As we cannot then use any other sibliing
while the context is inflight, only kick the bound sibling while it
inflight and upon scheduling out the kick the rest (so that we can swap
engines on timeslicing if the previously bound engine becomes
oversubscribed).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-3-chris@chris-wilson.co.uk
Rather than going back and forth between the rb_node entry and the
virtual_engine type, store the ve local and reuse it. As the
container_of conversion from rb_node to virtual_engine requires a
variable offset, performing that conversion just once shaves off a bit
of code.
v2: Keep a single virtual engine lookup, for typical use.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-2-chris@chris-wilson.co.uk
Rather than having special case code for opportunistically calling
process_csb() and performing a direct submit while holding the engine
spinlock for submitting the request, simply call the tasklet directly.
This allows us to retain the direct submission path, including the CS
draining to allow fast/immediate submissions, without requiring any
duplicated code paths, and most importantly greatly simplifying the
control flow by removing reentrancy. This will enable us to close a few
races in the virtual engines in the next few patches.
The trickiest part here is to ensure that paired operations (such as
schedule_in/schedule_out) remain under consistent locking domains,
e.g. when pulled outside of the engine->active.lock
v2: Use bh kicking, see commit 3c53776e29 ("Mark HI and TASKLET
softirq synchronous").
v3: Update engine-reset to be tasklet aware
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201224135544.1713-1-chris@chris-wilson.co.uk
* dma-buf: Include <linux/vmalloc.h> for building on MIPS
* komeda: Fix order of operation in commit tail; Fix NULL-pointer and
out-of-bounds access; Cleanups
* ttm: Fix an unused-function warning
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEchf7rIzpz2NEoWjlaA3BHVMLeiMFAl/iRSAACgkQaA3BHVML
eiMrsQf/dlHwJykbsGbkE/lG8qgrnz4jFKkei+DnZjqNaAwvgbz9s1YvaMYIIdsE
wvYnBSTvdvhYzde08+E0Q4JLCvHh/0aHWqkZR3dGd22tAgNxGfslBITu07EyI+k8
Agy4AeBbYJUwYcfRAHqwDuKCDQ28/FXcBpLajSF9K9oG6et2GgftshuIwWij2xNR
kNQTddCsXdXXbFDK/N3X/Inb+WSbRU0GWDTMTUPC31rMXX6kIKASQmVvCQ8h6a5F
N9EqjUmGf7XGjdqWEkSM34HpCUv5c5N2MWc4OAiSaP3Q3QVJyp3RHg8zLOpeSIsy
hVAj+8SN8UrI+Fz0Tx+SFdU72t0eVA==
=hlDA
-----END PGP SIGNATURE-----
Merge tag 'drm-misc-next-fixes-2020-12-22' of git://anongit.freedesktop.org/drm/drm-misc into drm-next
Short summary of fixes pull:
* dma-buf: Include <linux/vmalloc.h> for building on MIPS
* komeda: Fix order of operation in commit tail; Fix NULL-pointer and
out-of-bounds access; Cleanups
* ttm: Fix an unused-function warning
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/X+JFYlW1SEZa6ShA@linux-uq9g
If we want to reuse a fence that is in active use by the GPU, we have to
wait an uncertain amount of time, but if we reuse an inactive fence, we
can change it right away. Loop through the list of available fences
twice, ignoring any active fences on the first pass.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201223122051.4624-1-chris@chris-wilson.co.uk
Pull the GT clock information [used to derive CS timestamps and PM
interval] under the GT so that is it local to the users. In doing so, we
consolidate the two references for the same information, of which the
runtime-info took note of a potential clock source override and scaling
factors.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201223122359.22562-2-chris@chris-wilson.co.uk
We assume that both timestamps are driven off the same clock [reported
to userspace as I915_PARAM_CS_TIMESTAMP_FREQUENCY]. Verify that this is
so by reading the timestamp registers around a busywait (on an otherwise
idle engine so there should be no preemptions).
v2: Icelake (not ehl, nor tgl) seems to be using a fixed 80ns interval
for, and only for, CTX_TIMESTAMP -- or it may be GPU frequency and the
test is always running at maximum frequency?. As far as I can tell, this
isolated change in behaviour is undocumented.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201223122359.22562-1-chris@chris-wilson.co.uk
the psp access ih path is not needed in navi
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jane Jian <Jane.Jian@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Adjust a jump target so that a bit of exception handling can be better
reused at the end of this function.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
* Return directly after a call of the function “kzalloc” failed
at the beginning.
* Delete a label which became unnecessary with this refactoring.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We can have the same src ids for different client ids so make sure to
check both the client id and the source id when handling interrupts.
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We need to take into account the client id otherwise we'll end
up sending generic events for any src id that is registered.
We only support irq domains on pre-soc15 parts so client is
always legacy.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
tear down software ih ring and its state.
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
software ih ring is enabled in vega10 and navi
ih block by default.
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
software ih ring will be used as a workaround
in case hardware ih ring 1 and ring 2 don't work
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This is supported by Sienna Cichlid, Navy Flounder and Dimgrey
Cavefish. For these ASICs, the target voltage calculation can be
illustrated by "voltage = voltage calculated from v/f curve +
overdrive vddgfx offset".
V2: limit the smu_version check for Sienna Cichlid only
Here are some sample usages about this new OD setting:
1. Check current vddgfx offset setting by
cat /sys/class/drm/card0/device/pp_od_clk_voltage
...
...
OD_VDDGFX_OFFSET:
0mV
...
...
2. Set new vddgfx offset by
echo "vo 10" > /sys/class/drm/card0/device/pp_od_clk_voltage
cat /sys/class/drm/card0/device/pp_od_clk_voltage
...
...
OD_VDDGFX_OFFSET:
10mV
...
...
3. Commit the new setting by
echo "c" > /sys/class/drm/card0/device/pp_od_clk_voltage
Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Populate the bootup overdrive table settings.
Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When user specifies a reduced WGP(CU) config via disalbe_cu module
parameter, this does not disable the clocks which uses additional
power. This interface send active WGP number to SMU and SMU will
cooperate with RLC to power off relative WGPs.
v2: Add request active WGPs in Vangogh smu post init.
Signed-off-by: Jinzhou.Su <Jinzhou.Su@amd.com>
Reviewed-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Documents the hooks in struct pptable_funcs.
v2: Improved documentation accuracy.
v3: Improved set_default_od_settings() definition.
Signed-off-by: Ryan Taylor <Ryan.Taylor@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
APUs don't support ih ring 1 and ring 2.
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For all the ASICs that integrate psp v11, vega20
doesn't support ih reroute. arcturus and later will
allow kernel driver to program ih_cfg_index/data
through mmio directly. navi1x and onwards will only
support grb_ih_set command in sriov configuration.
psp_v11_0_reroute_ih is not needed any more.
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
except for RENOIR, it is not correct to have
IH_CHICKEN programming in vega10 ih block.
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
