platform/kernel/linux-rpi.git
2 years agodrm/rockchip: vop2: Register Esmart0-win0 as primary plane
Sascha Hauer [Mon, 26 Sep 2022 08:16:43 +0000 (10:16 +0200)]
drm/rockchip: vop2: Register Esmart0-win0 as primary plane

Esmart0-win0 could serve as primary plane, so mark it as such. On
RK3568 this window will never be used as primary plane, because the
three windows at the beginning of the rk3568_vop_win_data[] array
will be used. On RK3566 however, two of the windows at the beginning
of the rk3568_vop_win_data[] array cannot not be used due to hardware
limitations, so without this patch we end up with CRTCs without primary
planes when multiple VPs are active.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Chris Morgan <macromorgan@hotmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220926081643.304759-1-s.hauer@pengutronix.de
2 years agodrm: rockchip: remove rockchip_drm_framebuffer_init() function
Johan Jonker [Wed, 19 Oct 2022 21:35:03 +0000 (23:35 +0200)]
drm: rockchip: remove rockchip_drm_framebuffer_init() function

The function rockchip_drm_framebuffer_init() was in use
in the rockchip_drm_fbdev.c file, but that is now replaced
by a generic fbdev setup. Reduce the image size by
removing the rockchip_drm_framebuffer_init() and sub function
rockchip_fb_alloc() and cleanup the rockchip_drm_fb.h header file.

Signed-off-by: Johan Jonker <jbx6244@gmail.com>
Reviewed-by: John Keeping <john@metanate.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/ebe91504-c5df-99e4-635f-832218584051@gmail.com
2 years agodrm: bridge: adv7511: use dev_err_probe in probe function
Ahmad Fatoum [Wed, 26 Oct 2022 12:52:46 +0000 (14:52 +0200)]
drm: bridge: adv7511: use dev_err_probe in probe function

adv7511 probe may need to be attempted multiple times before no
-EPROBE_DEFER is returned. Currently, every such probe results in
an error message:

[    4.534229] adv7511 1-003d: failed to find dsi host
[    4.580288] adv7511 1-003d: failed to find dsi host

This is misleading, as there is no error and probe deferral is normal
behavior. Fix this by using dev_err_probe that will suppress
-EPROBE_DEFER errors. While at it, we touch all dev_err in the probe
path. This makes the code more concise and included the error code
everywhere to aid user in debugging.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221026125246.3188260-1-a.fatoum@pengutronix.de
2 years agodrm/vc4: Make sure we don't end up with a core clock too high
Maxime Ripard [Thu, 27 Oct 2022 12:52:47 +0000 (14:52 +0200)]
drm/vc4: Make sure we don't end up with a core clock too high

Following the clock rate range improvements to the clock framework,
trying to set a disjoint range on a clock will now result in an error.

Thus, we can't set a minimum rate higher than the maximum reported by
the firmware, or clk_set_min_rate() will fail.

Thus we need to clamp the rate we are about to ask for to the maximum
rate possible on that clock.

Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Link: https://lore.kernel.org/r/20220815-rpi-fix-4k-60-v5-7-fe9e7ac8b111@cerno.tech
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agodrm/vc4: hdmi: Add more checks for 4k resolutions
Dom Cobley [Thu, 27 Oct 2022 12:52:46 +0000 (14:52 +0200)]
drm/vc4: hdmi: Add more checks for 4k resolutions

At least the 4096x2160@60Hz mode requires some overclocking that isn't
available by default, even if hdmi_enable_4kp60 is enabled.

Let's add some logic to detect whether we can satisfy the core clock
requirements for that mode, and prevent it from being used otherwise.

Signed-off-by: Dom Cobley <popcornmix@gmail.com>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Link: https://lore.kernel.org/r/20220815-rpi-fix-4k-60-v5-6-fe9e7ac8b111@cerno.tech
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agodrm/vc4: hdmi: Rework hdmi_enable_4kp60 detection code
Maxime Ripard [Thu, 27 Oct 2022 12:52:45 +0000 (14:52 +0200)]
drm/vc4: hdmi: Rework hdmi_enable_4kp60 detection code

In order to support higher HDMI frequencies, users have to set the
hdmi_enable_4kp60 parameter in their config.txt file.

This will have the side-effect of raising the maximum of the core clock,
tied to the HVS, and managed by the HVS driver.

However, we are querying this in the HDMI driver by poking into the HVS
structure to get our struct clk handle.

Let's make this part of the HVS bind implementation to have all the core
clock related setup in the same place.

Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Link: https://lore.kernel.org/r/20220815-rpi-fix-4k-60-v5-5-fe9e7ac8b111@cerno.tech
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agodrm/vc4: hdmi: Fix hdmi_enable_4kp60 detection
Maxime Ripard [Thu, 27 Oct 2022 12:52:44 +0000 (14:52 +0200)]
drm/vc4: hdmi: Fix hdmi_enable_4kp60 detection

In order to support higher HDMI frequencies, users have to set the
hdmi_enable_4kp60 parameter in their config.txt file.

We were detecting this so far by calling clk_round_rate() on the core
clock with the frequency we're supposed to run at when one of those
modes is enabled. Whether or not the parameter was enabled could then be
inferred by the returned rate since the maximum clock rate reported by
the firmware was one of the side effect of setting that parameter.

However, the recent clock rework we did changed what clk_round_rate()
was returning to always return the minimum allowed, and thus this test
wasn't reliable anymore.

Let's use the new clk_get_max_rate() function to reliably determine the
maximum rate allowed on that clock and fix the 4k@60Hz output.

Fixes: e9d6cea2af1c ("clk: bcm: rpi: Run some clocks at the minimum rate allowed")
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Link: https://lore.kernel.org/r/20220815-rpi-fix-4k-60-v5-4-fe9e7ac8b111@cerno.tech
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agofirmware: raspberrypi: Provide a helper to query a clock max rate
Maxime Ripard [Thu, 27 Oct 2022 12:52:43 +0000 (14:52 +0200)]
firmware: raspberrypi: Provide a helper to query a clock max rate

The firmware allows to query for its clocks the operating range of a
given clock. We'll need this for some drivers (KMS, in particular) to
infer the state of some configuration options, so let's create a
function to do so.

Acked-by: Stephen Boyd <sboyd@kernel.org>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20220815-rpi-fix-4k-60-v5-3-fe9e7ac8b111@cerno.tech
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agofirmware: raspberrypi: Move the clock IDs to the firmware header
Maxime Ripard [Thu, 27 Oct 2022 12:52:42 +0000 (14:52 +0200)]
firmware: raspberrypi: Move the clock IDs to the firmware header

We'll need the clock IDs in more drivers than just the clock driver from
now on, so let's move them in the firmware header.

Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Link: https://lore.kernel.org/r/20220815-rpi-fix-4k-60-v5-2-fe9e7ac8b111@cerno.tech
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agofirmware: raspberrypi: Introduce rpi_firmware_find_node()
Maxime Ripard [Thu, 27 Oct 2022 12:52:41 +0000 (14:52 +0200)]
firmware: raspberrypi: Introduce rpi_firmware_find_node()

A significant number of RaspberryPi drivers using the firmware don't
have a phandle to it, so end up scanning the device tree to find a node
with the firmware compatible.

That code is duplicated everywhere, so let's introduce a helper instead.

Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20220815-rpi-fix-4k-60-v5-1-fe9e7ac8b111@cerno.tech
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agodrm/bridge: it6505: Fix return value check for pm_runtime_get_sync
Pin-yen Lin [Thu, 27 Oct 2022 03:21:49 +0000 (11:21 +0800)]
drm/bridge: it6505: Fix return value check for pm_runtime_get_sync

`pm_runtime_get_sync` may return 1 on success. Fix the `if` statement
here to make the code less confusing, even though additional calls to
`it6505_poweron` doesn't break anything when it's already powered.

This was reported by Dan Carpenter <dan.carpenter@oracle.com> in
https://lore.kernel.org/all/Y1fMCs6VnxbDcB41@kili/

Fixes: 10517777d302 ("drm/bridge: it6505: Adapt runtime power management framework")
Signed-off-by: Pin-yen Lin <treapking@chromium.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221027032149.2739912-1-treapking@chromium.org
2 years agodrm/ttm: rework on ttm_resource to use size_t type
Somalapuram Amaranath [Thu, 27 Oct 2022 09:12:37 +0000 (14:42 +0530)]
drm/ttm: rework on ttm_resource to use size_t type

Change ttm_resource structure from num_pages to size_t size in bytes.
v1 -> v2: change PFN_UP(dst_mem->size) to ttm->num_pages
v1 -> v2: change bo->resource->size to bo->base.size at some places
v1 -> v2: remove the local variable
v1 -> v2: cleanup cmp_size_smaller_first()
v2 -> v3: adding missing PFN_UP in ttm_bo_vm_fault_reserved

Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221027091237.983582-1-Amaranath.Somalapuram@amd.com
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
2 years agodrm/edid: convert to device specific logging
Jani Nikula [Mon, 24 Oct 2022 12:33:44 +0000 (15:33 +0300)]
drm/edid: convert to device specific logging

Convert to drm_kms_dbg/drm_err where possible, and reference the
connector using [CONNECTOR:%d:%s]. Pass connectors around a bit more to
enable this. Where this is not possible, unify the rest of the debugs to
DRM_DEBUG_KMS.

Rewrite tile debug logging to one line while at it.

v2:
- Use [CONNECTOR:%d:%s] throughout (Ville)
- Tile debug logging revamp
- Pass connector around a bit more

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/e48346bfe09a632d5a5faa55e3c161b196cf21e8.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: add [CONNECTOR:%d:%s] to debug logging
Jani Nikula [Mon, 24 Oct 2022 12:33:43 +0000 (15:33 +0300)]
drm/edid: add [CONNECTOR:%d:%s] to debug logging

Reference the connector using [CONNECTOR:%d:%s] in existing device based
debug logging.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Simon Ser <contact@emersion.fr>
Link: https://patchwork.freedesktop.org/patch/msgid/5884410682bcbc032de4d3af8562c0b271edaa7f.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid/firmware: convert to drm device specific logging
Jani Nikula [Mon, 24 Oct 2022 12:33:42 +0000 (15:33 +0300)]
drm/edid/firmware: convert to drm device specific logging

Conform to device specific logging.

v2: Include [CONNECTOR:%d:%s] (Ville)

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/14f3a1e55729c9157aae93fc45320d05cc4cc7bc.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: move edid load declarations to internal header
Jani Nikula [Mon, 24 Oct 2022 12:33:41 +0000 (15:33 +0300)]
drm/edid: move edid load declarations to internal header

The EDID loader is internal to drm, not for drivers.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/d58a59fede286caa8766e0813f4be492a7200287.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: use struct drm_edid for override/firmware EDID
Jani Nikula [Mon, 24 Oct 2022 12:33:40 +0000 (15:33 +0300)]
drm/edid: use struct drm_edid for override/firmware EDID

There's a lot going on here, but the main thing is switching the
firmware EDID loader to use struct drm_edid. Unfortunately, it's
difficult to reasonably split to smaller pieces.

Convert the EDID loader to struct drm_edid. There's a functional change
in validation; it no longer tries to fix errors or filter invalid
blocks. It's stricter in this sense. Hopefully this will not be an
issue.

As a by-product, this change also allows HF-EEODB extended EDIDs to be
passed via override/firmware EDID.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/e64267c28eca483e83c802bc06ddd149bdcdfc66.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid/firmware: rename drm_load_edid_firmware() to drm_edid_load_firmware()
Jani Nikula [Mon, 24 Oct 2022 12:33:39 +0000 (15:33 +0300)]
drm/edid/firmware: rename drm_load_edid_firmware() to drm_edid_load_firmware()

Follow the usual naming convention by file name.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/d6714ae737d789764bd2bdb6e7c9a5f56c99eef3.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid/firmware: drop redundant connector_name variable/parameter
Jani Nikula [Mon, 24 Oct 2022 12:33:38 +0000 (15:33 +0300)]
drm/edid/firmware: drop redundant connector_name variable/parameter

Stop passing around something that's readily available in
connector->name.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/befa53a33f2ea83094027a0e88d155779ad096e1.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: detach debugfs EDID override from EDID property update
Jani Nikula [Mon, 24 Oct 2022 12:33:37 +0000 (15:33 +0300)]
drm/edid: detach debugfs EDID override from EDID property update

Having the EDID override debugfs directly update the EDID property is
problematic. The update is partial only. The driver has no way of
knowing it's been updated. Mode list is not updated. It's an
inconsistent state.

Detach debugfs EDID override from the property update completely. Only
set and reset a separate override EDID copy from debugfs, and have it
take effect only at detect (via EDID read). The copy is at
connector->edid_override, protected by connector->edid_override_mutex.

This also brings override EDID closer to firmware EDID in behaviour.

Add validation of the override EDID which we completely lacked.

Note that IGT already forces a detect whenever tests update the override
EDID.

v2: Add locking (Ville)

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/4c875f8e06c4499f498fcf876e1233cbb155ec8a.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: add function for checking drm_edid validity
Jani Nikula [Mon, 24 Oct 2022 12:33:36 +0000 (15:33 +0300)]
drm/edid: add function for checking drm_edid validity

We've lacked a function for immutable validity check on drm_edid. Add
one.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/f96188f64e9f7f3deff348d08296609353b12316.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: split drm_edid block count helper
Jani Nikula [Mon, 24 Oct 2022 12:33:35 +0000 (15:33 +0300)]
drm/edid: split drm_edid block count helper

Split the drm_edid block count helper to a base version that reports the
block count indicated by EDID contents, and another on top that limits
the block count based on size allocated for the EDID.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/a7d63878c7fb3dd6f3b987f5257897113797b94f.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: rename drm_add_override_edid_modes() to drm_edid_override_connector_update()
Jani Nikula [Mon, 24 Oct 2022 12:33:34 +0000 (15:33 +0300)]
drm/edid: rename drm_add_override_edid_modes() to drm_edid_override_connector_update()

Follow the naming of both EDID override functions as well as
drm_edid_connector_update(). This also matches better what the function
does; a combination of EDID property update and add modes. Indeed it
should later be converted to call drm_edid_connector_update().

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ba12957e0488654e8db010a3ff1534079caec972.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: abstract debugfs override EDID show better
Jani Nikula [Mon, 24 Oct 2022 12:33:33 +0000 (15:33 +0300)]
drm/edid: abstract debugfs override EDID show better

Add a function to dump the override EDID in debugfs. This hides the
override EDID management better in drm_edid.c.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/74defa7b595f51e6c1f2eacd9c799d567d29f053.1666614699.git.jani.nikula@intel.com
2 years agodrm/edid: debug log EDID override set/reset
Jani Nikula [Mon, 24 Oct 2022 12:33:32 +0000 (15:33 +0300)]
drm/edid: debug log EDID override set/reset

It's useful debugging information to know if and when an override EDID
was set or reset.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ae352f542b4d69398c0965e33fb2e6e34156cbfb.1666614699.git.jani.nikula@intel.com
2 years agodrm/amd/display: stop using connector->override_edid
Jani Nikula [Mon, 24 Oct 2022 12:33:31 +0000 (15:33 +0300)]
drm/amd/display: stop using connector->override_edid

The connector->override_edid flag is strictly for EDID override debugfs
management, and drivers have no business using it.

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Xinhui Pan <Xinhui.Pan@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/c901869ff8a4e3aebc4abec99c7dd7b4c224f6e6.1666614699.git.jani.nikula@intel.com
2 years agodrm/i915/hdmi: stop using connector->override_edid
Jani Nikula [Mon, 24 Oct 2022 12:33:30 +0000 (15:33 +0300)]
drm/i915/hdmi: stop using connector->override_edid

The connector->override_edid flag is strictly for EDID override debugfs
management, and drivers have no business using it.

The check for override_edid was added in commit 301906290553 ("drm/i915:
Ignore TMDS clock limit for DP++ when EDID override is set") to
facilitate mode list cross-checking against modes in override EDID when
the connector in question isn't even connected. The dual mode detect
fallback would do VBT based limiting in this case.

Instead of override EDID, check for connector forcing in the fallback.

v2: Simply use !connector->force (Ville)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/c8b45867cf37134ab40be23e22825ca45adc6041.1666614699.git.jani.nikula@intel.com
2 years agodrm/i915/hdmi: do dual mode detect only if connected
Jani Nikula [Mon, 24 Oct 2022 12:33:29 +0000 (15:33 +0300)]
drm/i915/hdmi: do dual mode detect only if connected

For normal connector detect, there's really no point in trying dual mode
detect if the connector is disconnected. We can simplify the detect
sequence by skipping it. Since intel_hdmi_dp_dual_mode_detect() is only
called when EDID is present, we can drop the has_edid parameter.

The functional effect is speeding up disconnected connector detection
ever so slightly, and, combined with firmware EDID, also stop logging
about assuming dual mode adaptor.

It's a bit subtle, but this will also skip dual mode detect if the
connector is force connected and a) there's no EDID of any kind, normal
or override/firmware or b) there's EDID but it does not indicate
digital. These are corner cases no matter what, and arguably forcing
should not be limited by dual mode detect.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/f8f2a4a147e1c87ba93269a607f71fc29c4b59f6.1666614699.git.jani.nikula@intel.com
2 years agodrm/vmwgfx: Fix a sparse warning in kernel docs
Zack Rusin [Sat, 22 Oct 2022 04:02:36 +0000 (00:02 -0400)]
drm/vmwgfx: Fix a sparse warning in kernel docs

Fixes a warning about extra docs about a function argument that has been
removed a while back:
drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c:3888: warning: Excess function
parameter 'sync_file' description in 'vmw_execbuf_copy_fence_user'

Fixes: a0f90c881570 ("drm/vmwgfx: Fix stale file descriptors on failed usercopy")
Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Maaz Mombasawala <mombasawalam@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-18-zack@kde.org
2 years agodrm/vmwgfx: Optimize initial sizes of cotables
Zack Rusin [Sat, 22 Oct 2022 04:02:35 +0000 (00:02 -0400)]
drm/vmwgfx: Optimize initial sizes of cotables

It's important to get the initial size of cotables right because
otherwise every app needs to start with a synchronous cotable resize.

This has an measurable impact on system wide performance but is not
relevant for long running single full screen apps for which the cotable
resizes will happen early in the lifecycle and will continue running
just fine.

To eliminate the initial cotable resizes match the initial sizes to what
the userspace expects. The actual result of the patch is simply setting
the initial size of two of the cotables to a size that will align them
to two pages instead of one.

For a piglit run, before:
name               |  total |  per frame | per sec
vmw_cotable_resize |   1405 |       0.12 |    1.58
vmw_execbuf_ioctl  | 290805 |      25.43 |  326.05

After:
name               |  total |  per frame | per sec
vmw_cotable_resize |      4 |       0.00 |    0.00
vmw_execbuf_ioctl  | 281673 |      25.10 |  274.68

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Michael Banack <banackm@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-17-zack@kde.org
2 years agodrm/vmwgfx: Add a mksstat counter for cotable resizes
Zack Rusin [Sat, 22 Oct 2022 04:02:34 +0000 (00:02 -0400)]
drm/vmwgfx: Add a mksstat counter for cotable resizes

There's been a lot of cotable resizes on startup which we can track
by adding a mks stat to measure both the invocation count and
time spent doing cotable resizes.

This is only used if kernel is configured with CONFIG_DRM_VMWGFX_MKSSTATS
The stats are collected on the host size inside the vmware-stats.log
file.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Michael Banack <banackm@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Maaz Mombasawala <mombasawalam@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-16-zack@kde.org
2 years agodrm/vmwgfx: Remove explicit and broken vblank handling
Zack Rusin [Sat, 22 Oct 2022 04:02:33 +0000 (00:02 -0400)]
drm/vmwgfx: Remove explicit and broken vblank handling

The explicit vblank handling was never finished. The driver never had
the full implementation of vblank and what was there is emulated
by DRM when the driver doesn't pretend to be implementing it itself.

Let DRM handle the vblank emulation and stop pretending the driver is
doing anything special with vblank. In the future it would make sense
to implement helpers for full vblank handling because vkms and
amdgpu_vkms already have that code. Exporting it to common helpers and
having all three drivers share it would make sense (that would be largely
just to allow more of igt to run).

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Michael Banack <banackm@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-15-zack@kde.org
2 years agodrm/vmwgfx: Port the framebuffer code to drm fb helpers
Zack Rusin [Sat, 22 Oct 2022 04:02:32 +0000 (00:02 -0400)]
drm/vmwgfx: Port the framebuffer code to drm fb helpers

Instead of using vmwgfx specific framebuffer implementation use the drm
fb helpers. There's no change in functionality, the only difference
is a reduction in the amount of code inside the vmwgfx module.

drm fb helpers do not deal correctly with changes in crtc preferred mode
at runtime, but the old fb code wasn't dealing with it either.
Same situation applies to high-res fb consoles - the old code was
limited to 1176x885 because it was checking for legacy/deprecated
memory limites, the drm fb helpers are limited to the initial resolution
set on fb due to first problem (drm fb helpers being unable to handle
hotplug crtc preferred mode changes).

This also removes the kernel config for disabling fb support which hasn't
been used or supported in a very long time.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-14-zack@kde.org
2 years agodrm/vmwgfx: Do not allow invalid bpp's for dumb buffers
Zack Rusin [Sat, 22 Oct 2022 04:02:31 +0000 (00:02 -0400)]
drm/vmwgfx: Do not allow invalid bpp's for dumb buffers

Dumb buffers allow a very limited set of formats. Basically everything
apart from 1, 2 and 4 is expected to return an error. Make vmwgfx
follow those guidelines.

This fixes igt's dumb_buffer invalid_bpp test on vmwgfx.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Maaz Mombasawala <mombasawalam@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-13-zack@kde.org
2 years agodrm/vmwgfx: Remove vmwgfx_hashtab
Maaz Mombasawala [Sat, 22 Oct 2022 04:02:30 +0000 (00:02 -0400)]
drm/vmwgfx: Remove vmwgfx_hashtab

The vmwgfx driver has migrated from using the hashtable in vmwgfx_hashtab
to the linux/hashtable implementation. Remove the vmwgfx_hashtab from the
driver.

Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-12-zack@kde.org
2 years agodrm/vmwgfx: Refactor ttm reference object hashtable to use linux/hashtable.
Maaz Mombasawala [Sat, 22 Oct 2022 04:02:29 +0000 (00:02 -0400)]
drm/vmwgfx: Refactor ttm reference object hashtable to use linux/hashtable.

This is part of an effort to move from the vmwgfx_open_hash hashtable to
linux/hashtable implementation.
Refactor the ref_hash hashtable, used for fast lookup of reference objects
associated with a ttm file.
This also exposed a problem related to inconsistently using 32-bit and
64-bit keys with this hashtable. The hash function used changes depending
on the size of the type, and results are not consistent across numbers,
for example, hash_32(329) = 329, but hash_long(329) = 328. This would
cause the lookup to fail for objects already in the hashtable, since keys
of different sizes were being passed during adding and lookup. This was
not an issue before because vmwgfx_open_hash always used hash_long.
Fix this by always using 64-bit keys for this hashtable, which means that
hash_long is always used.

Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-11-zack@kde.org
2 years agodrm/vmwgfx: Diff cursors when using cmds
Michael Banack [Sat, 22 Oct 2022 04:02:28 +0000 (00:02 -0400)]
drm/vmwgfx: Diff cursors when using cmds

Extend the cursor diffing support to support the command-path.

Signed-off-by: Michael Banack <banackm@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-10-zack@kde.org
2 years agodrm/vmwgfx: Support cursor surfaces with mob cursor
Michael Banack [Sat, 22 Oct 2022 04:02:27 +0000 (00:02 -0400)]
drm/vmwgfx: Support cursor surfaces with mob cursor

Add support for cursor surfaces when using mob cursors.

Signed-off-by: Michael Banack <banackm@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-9-zack@kde.org
2 years agodrm/vmwgfx: Start diffing new mob cursors against old ones
Michael Banack [Sat, 22 Oct 2022 04:02:26 +0000 (00:02 -0400)]
drm/vmwgfx: Start diffing new mob cursors against old ones

Avoid making the SVGA device do extra work if the new cursor image
matches the old one.

Signed-off-by: Michael Banack <banackm@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-8-zack@kde.org
2 years agodrm/vmwgfx: Clean up cursor mobs
Michael Banack [Sat, 22 Oct 2022 04:02:25 +0000 (00:02 -0400)]
drm/vmwgfx: Clean up cursor mobs

Clean up the cursor mob path by moving ownership of the mobs into the
plane_state, and just leaving a cache of unused mobs in the plane
itself.

Signed-off-by: Michael Banack <banackm@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-7-zack@kde.org
2 years agodrm/vmwgfx: Refactor resource validation hashtable to use linux/hashtable implementation.
Maaz Mombasawala [Sat, 22 Oct 2022 04:02:24 +0000 (00:02 -0400)]
drm/vmwgfx: Refactor resource validation hashtable to use linux/hashtable implementation.

Vmwgfx's hashtab implementation needs to be replaced with linux/hashtable
to reduce maintenence burden.
As part of this effort, refactor the res_ht hashtable used for resource
validation during execbuf execution to use linux/hashtable implementation.
This also refactors vmw_validation_context to use vmw_sw_context as the
container for the hashtable, whereas before it used a vmwgfx_open_hash
directly. This makes vmw_validation_context less generic, but there is
no functional change since res_ht is the only instance where validation
context used a hashtable in vmwgfx driver.

Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-6-zack@kde.org
2 years agodrm/vmwgfx: Remove ttm object hashtable
Maaz Mombasawala [Sat, 22 Oct 2022 04:02:23 +0000 (00:02 -0400)]
drm/vmwgfx: Remove ttm object hashtable

The object_hash hashtable for ttm objects is not being used.
Remove it and perform refactoring in ttm_object init function.

Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-5-zack@kde.org
2 years agodrm/vmwgfx: Refactor resource manager's hashtable to use linux/hashtable implementation.
Maaz Mombasawala [Sat, 22 Oct 2022 04:02:22 +0000 (00:02 -0400)]
drm/vmwgfx: Refactor resource manager's hashtable to use linux/hashtable implementation.

Vmwgfx's hashtab implementation needs to be replaced with linux/hashtable
to reduce maintenance burden.
Refactor cmdbuf resource manager to use linux/hashtable.h implementation
as part of this effort.

Signed-off-by: Maaz Mombasawala <mombasawalam@vmware.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-4-zack@kde.org
2 years agodrm/vmwgfx: Fix frame-size warning in vmw_mksstat_add_ioctl
Martin Krastev [Sat, 22 Oct 2022 04:02:21 +0000 (00:02 -0400)]
drm/vmwgfx: Fix frame-size warning in vmw_mksstat_add_ioctl

Function vmw_mksstat_add_ioctl allocates three big arrays on stack.
That triggers frame-size [-Wframe-larger-than=] warning. Refactor
that function to use kmalloc_array instead.

v2: Initialize page to null to avoid possible uninitialized use of it,
    spotted by the kernel test robot <lkp@intel.com>

Signed-off-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Maaz Mombasawala <mombasawalam@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-3-zack@kde.org
2 years agodrm/vmwgfx: Write the driver id registers
Zack Rusin [Sat, 22 Oct 2022 04:02:20 +0000 (00:02 -0400)]
drm/vmwgfx: Write the driver id registers

Driver id registers are a new mechanism in the svga device to hint to the
device which driver is running. This should not change device behavior
in any way, but might be convenient to work-around specific bugs
in guest drivers.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Maaz Mombasawala <mombasawalam@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022040236.616490-2-zack@kde.org
2 years agodrm/scheduler: Set the FIFO scheduling policy as the default
Luben Tuikov [Mon, 24 Oct 2022 21:26:34 +0000 (17:26 -0400)]
drm/scheduler: Set the FIFO scheduling policy as the default

The currently default Round-Robin GPU scheduling can result in starvation
of entities which have a large number of jobs, over entities which have
a very small number of jobs (single digit).

This can be illustrated in the following diagram, where jobs are
alphabetized to show their chronological order of arrival, where job A is
the oldest, B is the second oldest, and so on, to J, the most recent job to
arrive.

    ---> entities
j | H-F-----A--E--I--
o | --G-----B-----J--
b | --------C--------
s\/ --------D--------

WLOG, assuming all jobs are "ready", then a R-R scheduling will execute them
in the following order (a slice off of the top of the entities' list),

H, F, A, E, I, G, B, J, C, D.

However, to mitigate job starvation, we'd rather execute C and D before E,
and so on, given, of course, that they're all ready to be executed.

So, if all jobs are ready at this instant, the order of execution for this
and the next 9 instances of picking the next job to execute, should really
be,

A, B, C, D, E, F, G, H, I, J,

which is their chronological order. The only reason for this order to be
broken, is if an older job is not yet ready, but a younger job is ready, at
an instant of picking a new job to execute. For instance if job C wasn't
ready at time 2, but job D was ready, then we'd pick job D, like this:

0 +1 +2  ...
A, B, D, ...

And from then on, C would be preferred before all other jobs, if it is ready
at the time when a new job for execution is picked. So, if C became ready
two steps later, the execution order would look like this:

......0 +1 +2  ...
A, B, D, E, C, F, G, H, I, J

This is what the FIFO GPU scheduling algorithm achieves. It uses a
Red-Black tree to keep jobs sorted in chronological order, where picking
the oldest job is O(1) (we use the "cached" structure), and balancing the
tree is O(log n). IOW, it picks the *oldest ready* job to execute now.

The implementation is already in the kernel, and this commit only changes
the default GPU scheduling algorithm to use.

This was tested and achieves about 1% faster performance over the Round
Robin algorithm.

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Direct Rendering Infrastructure - Development <dri-devel@lists.freedesktop.org>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221024212634.27230-1-luben.tuikov@amd.com
Signed-off-by: Christian König <christian.koenig@amd.com>
2 years agodrm/connector: send hotplug uevent on connector cleanup
Simon Ser [Mon, 17 Oct 2022 15:32:01 +0000 (15:32 +0000)]
drm/connector: send hotplug uevent on connector cleanup

A typical DP-MST unplug removes a KMS connector. However care must
be taken to properly synchronize with user-space. The expected
sequence of events is the following:

1. The kernel notices that the DP-MST port is gone.
2. The kernel marks the connector as disconnected, then sends a
   uevent to make user-space re-scan the connector list.
3. User-space notices the connector goes from connected to disconnected,
   disables it.
4. Kernel handles the IOCTL disabling the connector. On success,
   the very last reference to the struct drm_connector is dropped and
   drm_connector_cleanup() is called.
5. The connector is removed from the list, and a uevent is sent to tell
   user-space that the connector disappeared.

The very last step was missing. As a result, user-space thought the
connector still existed and could try to disable it again. Since the
kernel no longer knows about the connector, that would end up with
EINVAL and confused user-space.

Fix this by sending a hotplug uevent from drm_connector_cleanup().

Signed-off-by: Simon Ser <contact@emersion.fr>
Cc: stable@vger.kernel.org
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Lyude Paul <lyude@redhat.com>
Cc: Jonas Ådahl <jadahl@redhat.com>
Tested-by: Jonas Ådahl <jadahl@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017153150.60675-2-contact@emersion.fr
2 years agoRevert "drm: hide unregistered connectors from GETCONNECTOR IOCTL"
Simon Ser [Mon, 17 Oct 2022 15:31:57 +0000 (15:31 +0000)]
Revert "drm: hide unregistered connectors from GETCONNECTOR IOCTL"

This reverts commit 981f09295687f856d5345e19c7084aca481c1395.

It turns out this causes logically active but disconnected DP MST
connectors to disappear from the KMS resources list, and Mutter
then assumes the connector is already disabled. Later on Mutter tries
to re-use the same CRTC but fails since on the kernel side it's still
tied to the disconnected DP MST connector.

Signed-off-by: Simon Ser <contact@emersion.fr>
Tested-by: Jonas Ådahl <jadahl@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017153150.60675-1-contact@emersion.fr
2 years agodrm/client: Switch drm_client_buffer_delete() to unlocked drm_gem_vunmap
Dmitry Osipenko [Thu, 20 Oct 2022 21:33:35 +0000 (00:33 +0300)]
drm/client: Switch drm_client_buffer_delete() to unlocked drm_gem_vunmap

The drm_client_buffer_delete() wasn't switched to unlocked GEM vunmapping
by accident when rest of drm_client code transitioned to the unlocked
variants of the vmapping functions. Make drm_client_buffer_delete() use
the unlocked variant. This fixes lockdep warning splat about missing
reservation lock when framebuffer is released.

Reported-by: kernel test robot <yujie.liu@intel.com>
Link: https://lore.kernel.org/dri-devel/890f70db-68b0-8456-ca3c-c5496ef90517@collabora.com/T/
Fixes: 79e2cf2e7a19 ("drm/gem: Take reservation lock for vmap/vunmap operations")
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Acked-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020213335.309092-1-dmitry.osipenko@collabora.com
2 years agodrm/bridge: it6505: Improve synchronization between extcon subsystem
Pin-yen Lin [Thu, 13 Oct 2022 11:04:10 +0000 (19:04 +0800)]
drm/bridge: it6505: Improve synchronization between extcon subsystem

Originally, the it6505 relies on a short sleep in the IRQ handler and a
long sleep to make sure it6505->lane_swap and it6505->lane_count is
configured in it6505_extcon_work and it6505_detect, respectively.

Use completion and additional DPCD read to remove the unnecessary waits,
and use a different lock for it6505_extcon_work and the threaded IRQ
handler because they no longer need to run exclusively.

The wait time of the completion is usually less than 10ms in local
experiments, but leave it larger here just in case.

Signed-off-by: Pin-yen Lin <treapking@chromium.org>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221013110411.1674359-4-treapking@chromium.org
2 years agodrm/bridge: it6505: Setup links in it6505_irq_hpd
Pin-yen Lin [Thu, 13 Oct 2022 11:04:09 +0000 (19:04 +0800)]
drm/bridge: it6505: Setup links in it6505_irq_hpd

Move the DPCD read and link setup steps to HPD IRQ handler to remove
an unnecessary dependency between .detect callback and the HPD IRQ
handler before registering it6505 as a DRM bridge. This is safe because
there is always a .detect call after each HPD IRQ handler triggered by
the drm_helper_hpd_irq_event call.

Signed-off-by: Pin-yen Lin <treapking@chromium.org>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221013110411.1674359-3-treapking@chromium.org
2 years agodrm/bridge: it6505: Initialize AUX channel in it6505_i2c_probe
Pin-yen Lin [Thu, 13 Oct 2022 11:04:08 +0000 (19:04 +0800)]
drm/bridge: it6505: Initialize AUX channel in it6505_i2c_probe

During device boot, the HPD interrupt could be triggered before the DRM
subsystem registers it6505 as a DRM bridge. In such cases, the driver
tries to access AUX channel and causes NULL pointer dereference.
Initializing the AUX channel earlier to prevent such error.

Fixes: b5c84a9edcd4 ("drm/bridge: add it6505 driver")
Signed-off-by: Pin-yen Lin <treapking@chromium.org>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Robert Foss <robert.foss@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221013110411.1674359-2-treapking@chromium.org
2 years agogpu/drm: fix repeated words in comments
wangjianli [Sat, 22 Oct 2022 06:07:01 +0000 (14:07 +0800)]
gpu/drm: fix repeated words in comments

Delete the redundant word 'the'.

Signed-off-by: wangjianli <wangjianli@cdjrlc.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022060701.58280-1-wangjianli@cdjrlc.com
2 years agogpu/drm: fix repeated words in comments
wangjianli [Sat, 22 Oct 2022 06:05:51 +0000 (14:05 +0800)]
gpu/drm: fix repeated words in comments

Delete the redundant word 'the'.

Signed-off-by: wangjianli <wangjianli@cdjrlc.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20221022060551.56918-1-wangjianli@cdjrlc.com
2 years agodrm/vc4: Fix spelling mistake "mmaping" -> "mmapping"
Colin Ian King [Fri, 21 Oct 2022 08:40:35 +0000 (09:40 +0100)]
drm/vc4: Fix spelling mistake "mmaping" -> "mmapping"

There are a couple of spelling mistakes in DRM_DEBUG messages. Fix them.

Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20221021084035.65367-1-colin.i.king@gmail.com
2 years agonouveau/dmem: Remove duplicated include in nouveau_dmem.c
Yang Li [Mon, 17 Oct 2022 00:07:23 +0000 (08:07 +0800)]
nouveau/dmem: Remove duplicated include in nouveau_dmem.c

./drivers/gpu/drm/nouveau/nouveau_dmem.c: nvif/if000c.h is included more
than once.

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=2404
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017000723.113744-1-yang.lee@linux.alibaba.com
2 years agodrm/nouveau: Remove unused variable
Zack Rusin [Fri, 21 Oct 2022 01:07:03 +0000 (21:07 -0400)]
drm/nouveau: Remove unused variable

Trivial removal of an unused variable. Not sure how it snuck by me and
build bots in the 7c99616e3fe7.

Fixes: 7c99616e3fe7 ("drm: Remove drm_mode_config::fb_base")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Thomas Zimmermann <tzimemrmann@suse.de>
Cc: Christian König <christian.koenig@amd.com>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Karol Herbst <kherbst@redhat.com>
Cc: Lyude Paul <lyude@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221021010703.536318-1-zack@kde.org
2 years agonouveau: Fix migrate_to_ram() for faulting page
Alistair Popple [Wed, 19 Oct 2022 12:29:34 +0000 (23:29 +1100)]
nouveau: Fix migrate_to_ram() for faulting page

Commit 16ce101db85d ("mm/memory.c: fix race when faulting a device private
page") changed the migrate_to_ram() callback to take a reference on the
device page to ensure it can't be freed while handling the fault.
Unfortunately the corresponding update to Nouveau to accommodate this
change was inadvertently dropped from that patch causing GPU to CPU
migration to fail so add it here.

Signed-off-by: Alistair Popple <apopple@nvidia.com>
Fixes: 16ce101db85d ("mm/memory.c: fix race when faulting a device private page")
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221019122934.866205-1-apopple@nvidia.com
2 years agodrm/panel-edp: Add INX N116BGE-EA2 (HW: C4)
Sean Hong [Fri, 21 Oct 2022 03:10:24 +0000 (11:10 +0800)]
drm/panel-edp: Add INX N116BGE-EA2 (HW: C4)

Add support for the INX - N116BGE-EA2 (HW: C4) panel.

Signed-off-by: Sean Hong <sean.hong@quanta.corp-partner.google.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221021031024.2899082-1-sean.hong@quanta.corp-partner.google.com
2 years agodrm/panel-edp: Add INX N116BGE-EA2 (HW: C2)
Sean Hong [Fri, 21 Oct 2022 02:58:01 +0000 (10:58 +0800)]
drm/panel-edp: Add INX N116BGE-EA2 (HW: C2)

Add support for the INX - N116BGE-EA2 (HW: C2) panel.

Signed-off-by: Sean Hong <sean.hong@quanta.corp-partner.google.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221021025801.2898500-1-sean.hong@quanta.corp-partner.google.com
2 years agodrm/plane_helper: Split into parameterized test cases
Michał Winiarski [Thu, 20 Oct 2022 08:21:35 +0000 (10:21 +0200)]
drm/plane_helper: Split into parameterized test cases

The test was constructed as a single function (test case) which checks
multiple conditions, calling the function that is tested multiple times
with different arguments.
This usually means that it can be easily converted into multiple test
cases.
Split igt_check_plane_state into two parameterized test cases,
drm_check_plane_state and drm_check_invalid_plane_state.

Passing output:
============================================================
============== drm_plane_helper (2 subtests) ===============
================== drm_check_plane_state ===================
[PASSED] clipping_simple
[PASSED] clipping_rotate_reflect
[PASSED] positioning_simple
[PASSED] upscaling
[PASSED] downscaling
[PASSED] rounding1
[PASSED] rounding2
[PASSED] rounding3
[PASSED] rounding4
============== [PASSED] drm_check_plane_state ==============
============== drm_check_invalid_plane_state ===============
[PASSED] positioning_invalid
[PASSED] upscaling_invalid
[PASSED] downscaling_invalid
========== [PASSED] drm_check_invalid_plane_state ==========
================ [PASSED] drm_plane_helper =================
============================================================
Testing complete. Ran 12 tests: passed: 12

v2: Add missing EXPECT/ASSERT (Maíra)
v3: Use single EXPECT insted of condition + KUNIT_FAILURE (Maíra)
v4: Rebase after "drm_test" rename

Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020082135.779872-2-michal.winiarski@intel.com
2 years agodrm/plane_helper: Print actual/expected values on failure
Michał Winiarski [Thu, 20 Oct 2022 08:21:34 +0000 (10:21 +0200)]
drm/plane_helper: Print actual/expected values on failure

Currently the values are printed with debug log level.
Adjust the log level and link the output with the test by using kunit_err.

Example output:
foo: dst: 20x20+10+10, expected: 10x10+0+0
foo: EXPECTATION FAILED at drivers/gpu/drm/tests/drm_plane_helper_test.c:85

Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020082135.779872-1-michal.winiarski@intel.com
2 years agodrm: Remove drm_mode_config::fb_base
Zack Rusin [Wed, 19 Oct 2022 02:43:50 +0000 (22:43 -0400)]
drm: Remove drm_mode_config::fb_base

The fb_base in struct drm_mode_config has been unused for a long time.
Some drivers set it and some don't leading to a very confusing state
where the variable can't be relied upon, because there's no indication
as to which driver sets it and which doesn't.

The only usage of fb_base is internal to two drivers so instead of trying
to force it into all the drivers to get it into a coherent state
completely remove it.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Thomas Zimmermann <tzimemrmann@suse.de>
Acked-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221019024401.394617-1-zack@kde.org
2 years agodrm/mgag200: Fix PLL setup for G200_SE_A rev >=4
Jocelyn Falempe [Thu, 13 Oct 2022 13:28:10 +0000 (15:28 +0200)]
drm/mgag200: Fix PLL setup for G200_SE_A rev >=4

For G200_SE_A, PLL M setting is wrong, which leads to blank screen,
or "signal out of range" on VGA display.
previous code had "m |= 0x80" which was changed to
m |= ((pixpllcn & BIT(8)) >> 1);

Tested on G200_SE_A rev 42

This line of code was moved to another file with
commit 877507bb954e ("drm/mgag200: Provide per-device callbacks for
PIXPLLC") but can be easily backported before this commit.

v2: * put BIT(7) First to respect MSB-to-LSB (Thomas)
    * Add a comment to explain that this bit must be set (Thomas)

Fixes: 2dd040946ecf ("drm/mgag200: Store values (not bits) in struct mgag200_pll_values")
Cc: stable@vger.kernel.org
Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221013132810.521945-1-jfalempe@redhat.com
2 years agoMerge drm/drm-next into drm-misc-next
Maxime Ripard [Tue, 18 Oct 2022 13:00:03 +0000 (15:00 +0200)]
Merge drm/drm-next into drm-misc-next

Let's kick-off this release cycle.

Signed-off-by: Maxime Ripard <maxime@cerno.tech>
2 years agodma-buf: Remove obsoleted internal lock
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:29 +0000 (20:22 +0300)]
dma-buf: Remove obsoleted internal lock

The internal dma-buf lock isn't needed anymore because the updated
locking specification claims that dma-buf reservation must be locked
by importers, and thus, the internal data is already protected by the
reservation lock. Remove the obsoleted internal lock.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-22-dmitry.osipenko@collabora.com
2 years agomedia: videobuf2: Stop using internal dma-buf lock
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:28 +0000 (20:22 +0300)]
media: videobuf2: Stop using internal dma-buf lock

All drivers that use dma-bufs have been moved to the updated locking
specification and now dma-buf reservation is guaranteed to be locked
by importers during the mapping operations. There is no need to take
the internal dma-buf lock anymore. Remove locking from the videobuf2
memory allocators.

Acked-by: Tomasz Figa <tfiga@chromium.org>
Acked-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-21-dmitry.osipenko@collabora.com
2 years agodma-buf: Document dynamic locking convention
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:27 +0000 (20:22 +0300)]
dma-buf: Document dynamic locking convention

Add documentation for the dynamic locking convention. The documentation
tells dma-buf API users when they should take the reservation lock and
when not.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-20-dmitry.osipenko@collabora.com
2 years agodma-buf: Move dma_buf_mmap() to dynamic locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:26 +0000 (20:22 +0300)]
dma-buf: Move dma_buf_mmap() to dynamic locking specification

Move dma_buf_mmap() function to the dynamic locking specification by
taking the reservation lock. Neither of the today's drivers take the
reservation lock within the mmap() callback, hence it's safe to enforce
the locking.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-19-dmitry.osipenko@collabora.com
2 years agodma-buf: Move dma_buf_map_attachment() to dynamic locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:25 +0000 (20:22 +0300)]
dma-buf: Move dma_buf_map_attachment() to dynamic locking specification

Move dma-buf attachment mapping functions to the dynamic locking
specification by asserting that the reservation lock is held.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-18-dmitry.osipenko@collabora.com
2 years agodma-buf: Move dma_buf_attach() to dynamic locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:24 +0000 (20:22 +0300)]
dma-buf: Move dma_buf_attach() to dynamic locking specification

Move dma-buf attachment API functions to the dynamic locking specification
by taking the reservation lock around the mapping operations. The strict
locking convention prevents deadlock situations for dma-buf importers and
exporters.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-17-dmitry.osipenko@collabora.com
2 years agodma-buf: Move dma_buf_vmap() to dynamic locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:23 +0000 (20:22 +0300)]
dma-buf: Move dma_buf_vmap() to dynamic locking specification

Move dma_buf_vmap/vunmap() functions to the dynamic locking
specification by asserting that the reservation lock is held.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-16-dmitry.osipenko@collabora.com
2 years agomedia: tegra-vde: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:22 +0000 (20:22 +0300)]
media: tegra-vde: Prepare to dynamic dma-buf locking specification

Prepare Tegra video decoder driver to the common dynamic dma-buf
locking convention by starting to use the unlocked versions of dma-buf
API functions.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-15-dmitry.osipenko@collabora.com
2 years agomedia: videobuf2: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:21 +0000 (20:22 +0300)]
media: videobuf2: Prepare to dynamic dma-buf locking specification

Prepare V4L2 memory allocators to the common dynamic dma-buf locking
convention by starting to use the unlocked versions of dma-buf API
functions.

Acked-by: Tomasz Figa <tfiga@chromium.org>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-14-dmitry.osipenko@collabora.com
2 years agoxen/gntdev: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:20 +0000 (20:22 +0300)]
xen/gntdev: Prepare to dynamic dma-buf locking specification

Prepare gntdev driver to the common dynamic dma-buf locking convention
by starting to use the unlocked versions of dma-buf API functions.

Acked-by: Juergen Gross <jgross@suse.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-13-dmitry.osipenko@collabora.com
2 years agomisc: fastrpc: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:19 +0000 (20:22 +0300)]
misc: fastrpc: Prepare to dynamic dma-buf locking specification

Prepare fastrpc to the common dynamic dma-buf locking convention by
starting to use the unlocked versions of dma-buf API functions.

Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-12-dmitry.osipenko@collabora.com
2 years agoRDMA/umem: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:18 +0000 (20:22 +0300)]
RDMA/umem: Prepare to dynamic dma-buf locking specification

Prepare InfiniBand drivers to the common dynamic dma-buf locking
convention by starting to use the unlocked versions of dma-buf API
functions.

Acked-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-11-dmitry.osipenko@collabora.com
2 years agodrm/etnaviv: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:17 +0000 (20:22 +0300)]
drm/etnaviv: Prepare to dynamic dma-buf locking specification

Prepare Etnaviv driver to the common dynamic dma-buf locking convention
by starting to use the unlocked versions of dma-buf API functions.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-10-dmitry.osipenko@collabora.com
2 years agodrm/tegra: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:16 +0000 (20:22 +0300)]
drm/tegra: Prepare to dynamic dma-buf locking specification

Prepare Tegra DRM driver to the common dynamic dma-buf locking convention
by starting to use the unlocked versions of dma-buf API functions.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-9-dmitry.osipenko@collabora.com
2 years agodrm/omapdrm: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:15 +0000 (20:22 +0300)]
drm/omapdrm: Prepare to dynamic dma-buf locking specification

Prepare OMAP DRM driver to the common dynamic dma-buf locking convention
by starting to use the unlocked versions of dma-buf API functions.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-8-dmitry.osipenko@collabora.com
2 years agodrm/i915: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:14 +0000 (20:22 +0300)]
drm/i915: Prepare to dynamic dma-buf locking specification

Prepare i915 driver to the common dynamic dma-buf locking convention
by starting to use the unlocked versions of dma-buf API functions
and handling cases where importer now holds the reservation lock.

Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-7-dmitry.osipenko@collabora.com
2 years agodrm/armada: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:13 +0000 (20:22 +0300)]
drm/armada: Prepare to dynamic dma-buf locking specification

Prepare Armada driver to the common dynamic dma-buf locking convention
by starting to use the unlocked versions of dma-buf API functions.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-6-dmitry.osipenko@collabora.com
2 years agodrm/prime: Prepare to dynamic dma-buf locking specification
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:12 +0000 (20:22 +0300)]
drm/prime: Prepare to dynamic dma-buf locking specification

Prepare DRM prime core to the common dynamic dma-buf locking convention
by starting to use the unlocked versions of dma-buf API functions.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-5-dmitry.osipenko@collabora.com
2 years agodrm/gem: Take reservation lock for vmap/vunmap operations
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:11 +0000 (20:22 +0300)]
drm/gem: Take reservation lock for vmap/vunmap operations

The new common dma-buf locking convention will require buffer importers
to hold the reservation lock around mapping operations. Make DRM GEM core
to take the lock around the vmapping operations and update DRM drivers to
use the locked functions for the case where DRM core now holds the lock.
This patch prepares DRM core and drivers to the common dynamic dma-buf
locking convention.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-4-dmitry.osipenko@collabora.com
2 years agodma-buf: Add unlocked variant of attachment-mapping functions
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:10 +0000 (20:22 +0300)]
dma-buf: Add unlocked variant of attachment-mapping functions

Add unlocked variant of dma_buf_map/unmap_attachment() that will
be used by drivers that don't take the reservation lock explicitly.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-3-dmitry.osipenko@collabora.com
2 years agodma-buf: Add unlocked variant of vmapping functions
Dmitry Osipenko [Mon, 17 Oct 2022 17:22:09 +0000 (20:22 +0300)]
dma-buf: Add unlocked variant of vmapping functions

Add unlocked variant of dma_buf_vmap/vunmap() that will be utilized
by drivers that don't take the reservation lock explicitly.

Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017172229.42269-2-dmitry.osipenko@collabora.com
2 years agoLinux 6.1-rc1
Linus Torvalds [Sun, 16 Oct 2022 22:36:24 +0000 (15:36 -0700)]
Linux 6.1-rc1

2 years agoMerge tag 'random-6.1-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 16 Oct 2022 22:27:07 +0000 (15:27 -0700)]
Merge tag 'random-6.1-rc1-for-linus' of git://git./linux/kernel/git/crng/random

Pull more random number generator updates from Jason Donenfeld:
 "This time with some large scale treewide cleanups.

  The intent of this pull is to clean up the way callers fetch random
  integers. The current rules for doing this right are:

   - If you want a secure or an insecure random u64, use get_random_u64()

   - If you want a secure or an insecure random u32, use get_random_u32()

     The old function prandom_u32() has been deprecated for a while
     now and is just a wrapper around get_random_u32(). Same for
     get_random_int().

   - If you want a secure or an insecure random u16, use get_random_u16()

   - If you want a secure or an insecure random u8, use get_random_u8()

   - If you want secure or insecure random bytes, use get_random_bytes().

     The old function prandom_bytes() has been deprecated for a while
     now and has long been a wrapper around get_random_bytes()

   - If you want a non-uniform random u32, u16, or u8 bounded by a
     certain open interval maximum, use prandom_u32_max()

     I say "non-uniform", because it doesn't do any rejection sampling
     or divisions. Hence, it stays within the prandom_*() namespace, not
     the get_random_*() namespace.

     I'm currently investigating a "uniform" function for 6.2. We'll see
     what comes of that.

  By applying these rules uniformly, we get several benefits:

   - By using prandom_u32_max() with an upper-bound that the compiler
     can prove at compile-time is ≤65536 or ≤256, internally
     get_random_u16() or get_random_u8() is used, which wastes fewer
     batched random bytes, and hence has higher throughput.

   - By using prandom_u32_max() instead of %, when the upper-bound is
     not a constant, division is still avoided, because
     prandom_u32_max() uses a faster multiplication-based trick instead.

   - By using get_random_u16() or get_random_u8() in cases where the
     return value is intended to indeed be a u16 or a u8, we waste fewer
     batched random bytes, and hence have higher throughput.

  This series was originally done by hand while I was on an airplane
  without Internet. Later, Kees and I worked on retroactively figuring
  out what could be done with Coccinelle and what had to be done
  manually, and then we split things up based on that.

  So while this touches a lot of files, the actual amount of code that's
  hand fiddled is comfortably small"

* tag 'random-6.1-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random:
  prandom: remove unused functions
  treewide: use get_random_bytes() when possible
  treewide: use get_random_u32() when possible
  treewide: use get_random_{u8,u16}() when possible, part 2
  treewide: use get_random_{u8,u16}() when possible, part 1
  treewide: use prandom_u32_max() when possible, part 2
  treewide: use prandom_u32_max() when possible, part 1

2 years agoMerge tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sun, 16 Oct 2022 22:14:29 +0000 (15:14 -0700)]
Merge tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git./linux/kernel/git/acme/linux

Pull more perf tools updates from Arnaldo Carvalho de Melo:

 - Use BPF CO-RE (Compile Once, Run Everywhere) to support old kernels
   when using bperf (perf BPF based counters) with cgroups.

 - Support HiSilicon PCIe Performance Monitoring Unit (PMU), that
   monitors bandwidth, latency, bus utilization and buffer occupancy.

   Documented in Documentation/admin-guide/perf/hisi-pcie-pmu.rst.

 - User space tasks can migrate between CPUs, so when tracing selected
   CPUs, system-wide sideband is still needed, fix it in the setup of
   Intel PT on hybrid systems.

 - Fix metricgroups title message in 'perf list', it should state that
   the metrics groups are to be used with the '-M' option, not '-e'.

 - Sync the msr-index.h copy with the kernel sources, adding support for
   using "AMD64_TSC_RATIO" in filter expressions in 'perf trace' as well
   as decoding it when printing the MSR tracepoint arguments.

 - Fix program header size and alignment when generating a JIT ELF in
   'perf inject'.

 - Add multiple new Intel PT 'perf test' entries, including a jitdump
   one.

 - Fix the 'perf test' entries for 'perf stat' CSV and JSON output when
   running on PowerPC due to an invalid topology number in that arch.

 - Fix the 'perf test' for arm_coresight failures on the ARM Juno
   system.

 - Fix the 'perf test' attr entry for PERF_FORMAT_LOST, adding this
   option to the or expression expected in the intercepted
   perf_event_open() syscall.

 - Add missing condition flags ('hs', 'lo', 'vc', 'vs') for arm64 in the
   'perf annotate' asm parser.

 - Fix 'perf mem record -C' option processing, it was being chopped up
   when preparing the underlying 'perf record -e mem-events' and thus
   being ignored, requiring using '-- -C CPUs' as a workaround.

 - Improvements and tidy ups for 'perf test' shell infra.

 - Fix Intel PT information printing segfault in uClibc, where a NULL
   format was being passed to fprintf.

* tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux: (23 commits)
  tools arch x86: Sync the msr-index.h copy with the kernel sources
  perf auxtrace arm64: Add support for parsing HiSilicon PCIe Trace packet
  perf auxtrace arm64: Add support for HiSilicon PCIe Tune and Trace device driver
  perf auxtrace arm: Refactor event list iteration in auxtrace_record__init()
  perf tests stat+json_output: Include sanity check for topology
  perf tests stat+csv_output: Include sanity check for topology
  perf intel-pt: Fix system_wide dummy event for hybrid
  perf intel-pt: Fix segfault in intel_pt_print_info() with uClibc
  perf test: Fix attr tests for PERF_FORMAT_LOST
  perf test: test_intel_pt.sh: Add 9 tests
  perf inject: Fix GEN_ELF_TEXT_OFFSET for jit
  perf test: test_intel_pt.sh: Add jitdump test
  perf test: test_intel_pt.sh: Tidy some alignment
  perf test: test_intel_pt.sh: Print a message when skipping kernel tracing
  perf test: test_intel_pt.sh: Tidy some perf record options
  perf test: test_intel_pt.sh: Fix return checking again
  perf: Skip and warn on unknown format 'configN' attrs
  perf list: Fix metricgroups title message
  perf mem: Fix -C option behavior for perf mem record
  perf annotate: Add missing condition flags for arm64
  ...

2 years agodrm/ttm: Remove unnecessary drm_mm_clean
Zeng Heng [Wed, 12 Oct 2022 12:47:35 +0000 (20:47 +0800)]
drm/ttm: Remove unnecessary drm_mm_clean

Remove unnecessary `drm_mm_clean` calling in
`ttm_range_man_fini_nocheck`, due to effective
check is already included in the following
`drm_mm_takedown`.

Signed-off-by: Zeng Heng <zengheng4@huawei.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221012124735.1702700-1-zengheng4@huawei.com
Signed-off-by: Christian König <christian.koenig@amd.com>
2 years agoMerge tag 'kbuild-fixes-v6.1' of git://git.kernel.org/pub/scm/linux/kernel/git/masahi...
Linus Torvalds [Sun, 16 Oct 2022 18:12:22 +0000 (11:12 -0700)]
Merge tag 'kbuild-fixes-v6.1' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Fix CONFIG_DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT=y compile error for the
   combination of Clang >= 14 and GAS <= 2.35.

 - Drop vmlinux.bz2 from the rpm package as it just annoyingly increased
   the package size.

 - Fix modpost error under build environments using musl.

 - Make *.ll files keep value names for easier debugging

 - Fix single directory build

 - Prevent RISC-V from selecting the broken DWARF5 support when Clang
   and GAS are used together.

* tag 'kbuild-fixes-v6.1' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  lib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5
  kbuild: fix single directory build
  kbuild: add -fno-discard-value-names to cmd_cc_ll_c
  scripts/clang-tools: Convert clang-tidy args to list
  modpost: put modpost options before argument
  kbuild: Stop including vmlinux.bz2 in the rpm's
  Kconfig.debug: add toolchain checks for DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT
  Kconfig.debug: simplify the dependency of DEBUG_INFO_DWARF4/5

2 years agoMerge tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux
Linus Torvalds [Sun, 16 Oct 2022 18:08:19 +0000 (11:08 -0700)]
Merge tag 'clk-for-linus' of git://git./linux/kernel/git/clk/linux

Pull more clk updates from Stephen Boyd:
 "This is the final part of the clk patches for this merge window.

  The clk rate range series needed another week to fully bake. Maxime
  fixed the bug that broke clk notifiers and prevented this from being
  included in the first pull request. He also added a unit test on top
  to make sure it doesn't break so easily again. The majority of the
  series fixes up how the clk_set_rate_*() APIs work, particularly
  around when the rate constraints are dropped and how they move around
  when reparenting clks. Overall it's a much needed improvement to the
  clk rate range APIs that used to be pretty broken if you looked
  sideways.

  Beyond the core changes there are a few driver fixes for a compilation
  issue or improper data causing clks to fail to register or have the
  wrong parents. These are good to get in before the first -rc so that
  the system actually boots on the affected devices"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (31 commits)
  clk: tegra: Fix Tegra PWM parent clock
  clk: at91: fix the build with binutils 2.27
  clk: qcom: gcc-msm8660: Drop hardcoded fixed board clocks
  clk: mediatek: clk-mux: Add .determine_rate() callback
  clk: tests: Add tests for notifiers
  clk: Update req_rate on __clk_recalc_rates()
  clk: tests: Add missing test case for ranges
  clk: qcom: clk-rcg2: Take clock boundaries into consideration for gfx3d
  clk: Introduce the clk_hw_get_rate_range function
  clk: Zero the clk_rate_request structure
  clk: Stop forwarding clk_rate_requests to the parent
  clk: Constify clk_has_parent()
  clk: Introduce clk_core_has_parent()
  clk: Switch from __clk_determine_rate to clk_core_round_rate_nolock
  clk: Add our request boundaries in clk_core_init_rate_req
  clk: Introduce clk_hw_init_rate_request()
  clk: Move clk_core_init_rate_req() from clk_core_round_rate_nolock() to its caller
  clk: Change clk_core_init_rate_req prototype
  clk: Set req_rate on reparenting
  clk: Take into account uncached clocks in clk_set_rate_range()
  ...

2 years agoMerge tag '6.1-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 16 Oct 2022 18:01:40 +0000 (11:01 -0700)]
Merge tag '6.1-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6

Pull more cifs updates from Steve French:

 - fix a regression in guest mounts to old servers

 - improvements to directory leasing (caching directory entries safely
   beyond the root directory)

 - symlink improvement (reducing roundtrips needed to process symlinks)

 - an lseek fix (to problem where some dir entries could be skipped)

 - improved ioctl for returning more detailed information on directory
   change notifications

 - clarify multichannel interface query warning

 - cleanup fix (for better aligning buffers using ALIGN and round_up)

 - a compounding fix

 - fix some uninitialized variable bugs found by Coverity and the kernel
   test robot

* tag '6.1-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
  smb3: improve SMB3 change notification support
  cifs: lease key is uninitialized in two additional functions when smb1
  cifs: lease key is uninitialized in smb1 paths
  smb3: must initialize two ACL struct fields to zero
  cifs: fix double-fault crash during ntlmssp
  cifs: fix static checker warning
  cifs: use ALIGN() and round_up() macros
  cifs: find and use the dentry for cached non-root directories also
  cifs: enable caching of directories for which a lease is held
  cifs: prevent copying past input buffer boundaries
  cifs: fix uninitialised var in smb2_compound_op()
  cifs: improve symlink handling for smb2+
  smb3: clarify multichannel warning
  cifs: fix regression in very old smb1 mounts
  cifs: fix skipping to incorrect offset in emit_cached_dirents

2 years agodrm/tests: Split drm_test_dp_mst_sideband_msg_req_decode into parameterized tests
Maíra Canal [Sat, 1 Oct 2022 22:34:22 +0000 (19:34 -0300)]
drm/tests: Split drm_test_dp_mst_sideband_msg_req_decode into parameterized tests

The drm_test_dp_mst_sideband_msg_req_decode repeats the same test
structure with different parameters. This could be better represented
by parameterized tests, provided by KUnit.

In addition to the parameterization of the tests, the test case for the
client ID was changed: instead of using get_random_bytes to generate
the client ID, the client ID is now hardcoded in the test case. This
doesn't affect the assertively of the tests, as this test case only compare
the data going in with the data going out and it doesn't transform the data
itself in any way.

So, convert drm_test_dp_mst_sideband_msg_req_decode into parameterized
tests and make the tests' allocations and prints completely managed by KUnit.

Signed-off-by: Maíra Canal <mcanal@igalia.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20221001223422.857505-2-mcanal@igalia.com
2 years agodrm/tests: Split drm_test_dp_mst_calc_pbn_mode into parameterized tests
Maíra Canal [Sat, 1 Oct 2022 22:34:21 +0000 (19:34 -0300)]
drm/tests: Split drm_test_dp_mst_calc_pbn_mode into parameterized tests

The drm_test_dp_mst_calc_pbn_mode is based on a loop that executes tests
for a couple of test cases. This could be better represented by
parameterized tests, provided by KUnit.

So, convert the drm_test_dp_mst_calc_pbn_mode into parameterized tests.

Signed-off-by: Maíra Canal <mcanal@igalia.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20221001223422.857505-1-mcanal@igalia.com
2 years agoRevert "cpumask: fix checking valid cpu range".
Tetsuo Handa [Sat, 15 Oct 2022 15:53:51 +0000 (00:53 +0900)]
Revert "cpumask: fix checking valid cpu range".

This reverts commit 78e5a3399421 ("cpumask: fix checking valid cpu range").

syzbot is hitting WARN_ON_ONCE(cpu >= nr_cpumask_bits) warning at
cpu_max_bits_warn() [1], for commit 78e5a3399421 ("cpumask: fix checking
valid cpu range") is broken.  Obviously that patch hits WARN_ON_ONCE()
when e.g.  reading /proc/cpuinfo because passing "cpu + 1" instead of
"cpu" will trivially hit cpu == nr_cpumask_bits condition.

Although syzbot found this problem in linux-next.git on 2022/09/27 [2],
this problem was not fixed immediately.  As a result, that patch was
sent to linux.git before the patch author recognizes this problem, and
syzbot started failing to test changes in linux.git since 2022/10/10
[3].

Andrew Jones proposed a fix for x86 and riscv architectures [4].  But
[2] and [5] indicate that affected locations are not limited to arch
code.  More delay before we find and fix affected locations, less tested
kernel (and more difficult to bisect and fix) before release.

We should have inspected and fixed basically all cpumask users before
applying that patch.  We should not crash kernels in order to ask
existing cpumask users to update their code, even if limited to
CONFIG_DEBUG_PER_CPU_MAPS=y case.

Link: https://syzkaller.appspot.com/bug?extid=d0fd2bf0dd6da72496dd
Link: https://syzkaller.appspot.com/bug?extid=21da700f3c9f0bc40150
Link: https://syzkaller.appspot.com/bug?extid=51a652e2d24d53e75734
Link: https://lkml.kernel.org/r/20221014155845.1986223-1-ajones@ventanamicro.com
Link: https://syzkaller.appspot.com/bug?extid=4d46c43d81c3bd155060
Reported-by: Andrew Jones <ajones@ventanamicro.com>
Reported-by: syzbot+d0fd2bf0dd6da72496dd@syzkaller.appspotmail.com
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Yury Norov <yury.norov@gmail.com>
Cc: Borislav Petkov <bp@alien8.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agolib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5
Nathan Chancellor [Fri, 14 Oct 2022 20:42:11 +0000 (13:42 -0700)]
lib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5

When building with a RISC-V kernel with DWARF5 debug info using clang
and the GNU assembler, several instances of the following error appear:

  /tmp/vgettimeofday-48aa35.s:2963: Error: non-constant .uleb128 is not supported

Dumping the .s file reveals these .uleb128 directives come from
.debug_loc and .debug_ranges:

  .Ldebug_loc0:
          .byte   4                               # DW_LLE_offset_pair
          .uleb128 .Lfunc_begin0-.Lfunc_begin0    #   starting offset
          .uleb128 .Ltmp1-.Lfunc_begin0           #   ending offset
          .byte   1                               # Loc expr size
          .byte   90                              # DW_OP_reg10
          .byte   0                               # DW_LLE_end_of_list

  .Ldebug_ranges0:
          .byte   4                               # DW_RLE_offset_pair
          .uleb128 .Ltmp6-.Lfunc_begin0           #   starting offset
          .uleb128 .Ltmp27-.Lfunc_begin0          #   ending offset
          .byte   4                               # DW_RLE_offset_pair
          .uleb128 .Ltmp28-.Lfunc_begin0          #   starting offset
          .uleb128 .Ltmp30-.Lfunc_begin0          #   ending offset
          .byte   0                               # DW_RLE_end_of_list

There is an outstanding binutils issue to support a non-constant operand
to .sleb128 and .uleb128 in GAS for RISC-V but there does not appear to
be any movement on it, due to concerns over how it would work with
linker relaxation.

To avoid these build errors, prevent DWARF5 from being selected when
using clang and an assembler that does not have support for these symbol
deltas, which can be easily checked in Kconfig with as-instr plus the
small test program from the dwz test suite from the binutils issue.

Link: https://sourceware.org/bugzilla/show_bug.cgi?id=27215
Link: https://github.com/ClangBuiltLinux/linux/issues/1719
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2 years agokbuild: fix single directory build
Masahiro Yamada [Fri, 14 Oct 2022 20:18:11 +0000 (05:18 +0900)]
kbuild: fix single directory build

Commit f110e5a250e3 ("kbuild: refactor single builds of *.ko") was wrong.

KBUILD_MODULES _is_ needed for single builds.

Otherwise, "make foo/bar/baz/" does not build module objects at all.

Fixes: f110e5a250e3 ("kbuild: refactor single builds of *.ko")
Reported-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: David Sterba <dsterba@suse.com>
2 years agodrm/ast: Avoid reprogramming primary-plane scanout address
Thomas Zimmermann [Thu, 13 Oct 2022 11:29:23 +0000 (13:29 +0200)]
drm/ast: Avoid reprogramming primary-plane scanout address

Some AST-based BMCs stop display output for up to 5 seconds after
reprogramming the scanout address. As the address is fixed, avoid
re-setting the address' value.

v2:
* only update offset if it changed (Jocelyn)

Reported-by: Jocelyn Falempe <jfalempe@redhat.com>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221013112923.769-9-tzimmermann@suse.de
2 years agodrm/ast: Convert ast to SHMEM
Thomas Zimmermann [Thu, 13 Oct 2022 11:29:22 +0000 (13:29 +0200)]
drm/ast: Convert ast to SHMEM

Replace GEM VRAM helpers with GEM SHMEM helpers in ast. Avoids OOM
errors when allocating video memory. Also adds support for dma-buf
functionality.

Aspeed display hardware supports display resolutions of FullHD and
higher at 32-bit pixel depth. But the amount of video memory is in
the range of 8 MiB to 32 MiB, which adds constraints to the actually
available resolutions. As atomic modesetting with VRAM helpers
requires double buffering in video memory, ast fails to pageflip
in some configurations. For example, FullHD with an active cursor
plane does not work on devices with 16 MiB of video memory.

Resolve this problem by converting the ast driver to GEM SHMEM helpers.
Keep the buffer objects in system memory and copy to video memory
on pageflips via shadow-plane helpers. Userspace used to require shadow
planes for decent performance, but that's now provided by the driver.
To replace the memory management, the patch also implements damage
handling for the primary plane.

With GEM SHMEM helpers, dma-buf import and export is now supported
by ast. This allows easier screen mirroring across devices or with
an Aspeed-based BMC. A corresponding feature request is available
at [1].

v2:
* fix typos in commit message (Jocelyn)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://lore.kernel.org/dri-devel/20220901124451.2523077-1-oushixiong@kylinos.cn/
Link: https://patchwork.freedesktop.org/patch/msgid/20221013112923.769-8-tzimmermann@suse.de