Putting the DRM driver to the top of the file and the PCI code to the
bottom makes ast_drv.c more readable. While at it, the patch prefixes
file-scope variables with ast_.
drm/ast: Embed CRTC and connector in struct ast_private
Only single instances of CRTC and connector are supported per
device. Embed both in ast's structure and remove the individual
memory allocations. DRM's CRTC cleanup helpers replace the rsp.
destroy function in ast.
While at it, also convert to_ast_connector() to a function.
Daniel Vetter [Sat, 1 Aug 2020 09:26:25 +0000 (11:26 +0200)]
drm/syncobj: Tune down unordered timeline DRM_ERROR
Userspace can provoke this, we generally don't allow userspace to spam
dmesg. Tune it down to debug. Unfortunately we don't have easy access
to the drm_device here (not at all without changing a few things), so
leave it as old style dmesg output for now.
References: https://patchwork.freedesktop.org/series/80146/ Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: "Christian König" <christian.koenig@amd.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200801092625.1107609-1-daniel.vetter@ffwll.ch
The previous memset operation was not correctly zeroing the alpha
channel to compute the crc, and as a result, the IGT subtest
kms_cursor_crc/pipe-A-cursor-alpha-transparent fails.
Bernard Zhao [Sat, 1 Aug 2020 12:02:13 +0000 (20:02 +0800)]
drm/panel: remove return value of function drm_panel_add
The function "int drm_panel_add(struct drm_panel *panel)"
always returns 0, this return value is meaningless.
Also, there is no need to check return value which calls
"drm_panel_add and", error branch code will never run.
Christian König [Thu, 23 Jul 2020 15:13:47 +0000 (17:13 +0200)]
drm/ttm: remove the init_mem_type callback
It is a very strange concept to call a function which just
calls back the caller for the functions parameters.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Link: https://patchwork.freedesktop.org/patch/382085/
Christian König [Thu, 23 Jul 2020 14:40:17 +0000 (16:40 +0200)]
drm/vram-helper: stop implementing init_mem_type
Instead just initialize the memory type parameters
before calling ttm_bo_init_mm.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Link: https://patchwork.freedesktop.org/patch/382086/
Christian König [Tue, 21 Jul 2020 07:58:13 +0000 (09:58 +0200)]
drm/ttm: remove TTM_MEMTYPE_FLAG_FIXED v2
Instead use a boolean field in the memory manager structure.
Also invert the meaning of the field since the use of a TT
structure is the special case here.
v2: cleanup zero init.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Link: https://patchwork.freedesktop.org/patch/382079/
Christian König [Mon, 20 Jul 2020 12:41:31 +0000 (14:41 +0200)]
drm/ttm: initialize the system domain with defaults v2
Instead of repeating that in each driver.
v2: keep the caching limitation for VMWGFX for now.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de> Link: https://patchwork.freedesktop.org/patch/382078/
Paul Cercueil [Thu, 30 Jul 2020 14:48:30 +0000 (16:48 +0200)]
drm/ingenic: ipu: Only enable clock when needed
Instead of keeping the IPU clock enabled constantly, enable and disable
it on demand, when the IPU plane is used. That way, we won't use any
extra power when the IPU is not used.
Paul Cercueil [Thu, 30 Jul 2020 14:48:29 +0000 (16:48 +0200)]
drm/ingenic: ipu: Remove YUV422 from supported formats on JZ4725B
When configuring the IPU for packed YUV 4:2:2, depending on the scaling
ratios given by the source and destination resolutions, it is possible
to crash the IPU block, to the point where a software reset of the IP
does not fix it. This can happen anytime, in the first few frames, or
after dozens of minutes. The same crash also happens when the IPU is
fully controlled by the LCD controller (in that case no HW register is
written at any moment after startup), which points towards a hardware
bug.
Thanksfully multiplanar YUV is not affected.
Until this bug is fixed or worked around, address this issue by removing
support for YUV 4:2:2 on the IPU of the JZ4725B.
v2: Update commit message (remove the "crash beyond repair" bit)
Paul Cercueil [Thu, 30 Jul 2020 14:48:28 +0000 (16:48 +0200)]
drm/ingenic: ipu: Only restart manually on older SoCs
On older SoCs, it is necessary to restart manually the IPU when a frame
is done processing. Doing so on newer SoCs (JZ4760/70) kinds of work
too, until the input or output resolutions or the framerate are too
high.
Make it work properly on newer SoCs by letting the LCD controller
trigger the IPU frame restart signal.
Pekka Paalanen [Tue, 7 Jul 2020 11:38:05 +0000 (14:38 +0300)]
drm/doc: device hot-unplug for userspace
Set up the expectations on how hot-unplugging a DRM device should look like to
userspace.
Written by Daniel Vetter's request and largely based on his comments in IRC and
from https://lists.freedesktop.org/archives/dri-devel/2020-May/265484.html .
A related Wayland protocol change proposal is at
https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/35
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Noralf Trønnes <noralf@tronnes.org> Cc: Andrey Grodzovsky <andrey.grodzovsky@amd.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Sean Paul <sean@poorly.run> Cc: Simon Ser <contact@emersion.fr> Cc: Ben Skeggs <skeggsb@gmail.com> Cc: Karol Herbst <kherbst@redhat.com> Acked-by: Simon Ser <contact@emersion.fr> Signed-off-by: Emil Velikov <emil.velikov@collabora.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200707113805.30936-1-ppaalanen@gmail.com
Tian Tao [Tue, 28 Jul 2020 12:55:07 +0000 (20:55 +0800)]
drm/hisilicon: Fixed the warning: Assignment of 0/1 to bool variable
fixed the following warning:
hibmc_drm_drv.c:296:1-18:WARNING: Assignment of 0/1 to bool variable.
hibmc_drm_drv.c:301:2-19: WARNING: Assignment of 0/1 to bool variable.
v2:
using the pci_dev.msi_enabled instead of priv->msi_enabled.
v3:
just call pci_enable_msi() and pci_disable_msi(), it's no need to
set dev->pdev->msi_enabled again.
Paul Cercueil [Tue, 28 Jul 2020 15:16:41 +0000 (17:16 +0200)]
drm/ingenic: Validate mode in a .mode_valid callback
Validate modes in the drm_crtc_helper_funcs.mode_valid() callback, which
is designed for this purpose, instead of doing it in
drm_crtc_helper_funcs.atomic_check().
Add support for Powertip PH800480T013 800x480 parallel LCD, this
one is used in the Raspberry Pi 7" touchscreen display unit.
Signed-off-by: Marek Vasut <marex@denx.de>
To: dri-devel@lists.freedesktop.org Cc: Eric Anholt <eric@anholt.net> Cc: Rob Herring <robh+dt@kernel.org> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: devicetree@vger.kernel.org Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Link: https://patchwork.freedesktop.org/patch/msgid/20200728121246.23304-3-marex@denx.de
Marek Vasut [Tue, 28 Jul 2020 12:12:45 +0000 (14:12 +0200)]
dt-bindings: Add DT bindings for Powertip PH800480T013
Add DT bindings for Powertip PH800480T013 800x480 parallel LCD,
this one is used in the Raspberry Pi 7" touchscreen display unit.
Signed-off-by: Marek Vasut <marex@denx.de>
To: dri-devel@lists.freedesktop.org Cc: Eric Anholt <eric@anholt.net> Cc: Rob Herring <robh+dt@kernel.org> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: devicetree@vger.kernel.org Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Link: https://patchwork.freedesktop.org/patch/msgid/20200728121246.23304-2-marex@denx.de
Marek Vasut [Tue, 28 Jul 2020 12:12:44 +0000 (14:12 +0200)]
dt-bindings: Add vendor prefix for Powertip
The Powertip Tech. Corp. is an LCD panel manufacturer.
Signed-off-by: Marek Vasut <marex@denx.de>
To: dri-devel@lists.freedesktop.org Cc: Eric Anholt <eric@anholt.net> Cc: Rob Herring <robh+dt@kernel.org> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: devicetree@vger.kernel.org Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Link: https://patchwork.freedesktop.org/patch/msgid/20200728121246.23304-1-marex@denx.de
drm: mxsfb: Merge mxsfb_set_pixel_fmt() and mxsfb_set_bus_fmt()
The mxsfb_set_pixel_fmt() and mxsfb_set_bus_fmt() functions both deal
with format configuration, are always called in a row from
mxsfb_crtc_mode_set_nofb(), and set fields from the LCDC_CTRL register.
This requires a read-modify-update cycle in mxsfb_set_bus_fmt(). Make
this more efficient by merging them together.
drm: mxsfb: Turn mxsfb_set_pixel_fmt() into a void function
The mxsfb_set_pixel_fmt() function returns an error when the selected
pixel format is unsupported. This can never happen, as such errors are
caught by the DRM core. Remove the error check.
drm: mxsfb: Update internal IP version number for i.MX6SX
The LCDIF present in the i.MX6SX has extra features compared to
the i.MX28. It has however lost its IP version register, so no official
version number is known. Bump the version to MXSFB_V6 following the i.MX
version, in preparation for support for the additional features.
drm: mxsfb: Add i.MX7 and i.MX8M to the list of supported SoCs in Kconfig
Extend the Kconfig option description by listing the i.MX7 and i.MX8M
SoCs, as they are supported by the same driver. Replace the list of SoCs
in the short description with just "(e)LCDIF LCD controller" to avoid
expanding it further in the future as support for more SoCs is added.
Enable vblank handling when the CRTC is turned on and disable it when it
is turned off. This requires moving vblank init after the KMS pipeline
initialisation, otherwise drm_vblank_init() gets called with 0 CRTCs.
The driver attempts agressive power management by enabling and disabling
the AXI clock around register accesses. This results in attempts to
enable and disable the clock in the IRQ handler, which is a no-go as
preparing or unpreparing the clock may sleep.
On the other hand, the driver enables the AXI clock when enabling the
CRTC and keeps it enabled until the CRTC is disabled. This is correct,
and renders the power management attempt pointless, as interrupts are
not supposed to occur when the CRTC is off.
The same reasoning can be applied to the CRTC .enable_vblank() and
.disable_vblank() that are not supposed to be called when the CRTC off
and thus don't require manual handling of the AXI clock. Furthermore,
vblank handling is never enabled, which results in the vblank enable and
disable handlers never being called.
To fix this, remove the manual clock handling in the IRQ, the CRTC
.enable_vblank() and .disable_vblank() handlers and the plane
.atomic_update() handler. We however need to handle the clock manually
in mxsfb_irq_disable() as is calls .disable_vblank() manually and is
used both at probe and remove time.
The clock disabling is also moved to the last step of the
mxsfb_crtc_atomic_disable() function, to prepare for enabling vblank
handling.
drm: mxsfb: Move vblank event arm to CRTC .atomic_flush()
The vblank event is armed in the plane .atomic_update(). This works fine
as we have a single plane, and was the only option when the driver was
using the drm_simple_kms_helper helper, but will break as soon as
multiple planes are supported. Move it to CRTC .atomic_flush().
drm: mxsfb: Stop using DRM simple display pipeline helper
The DRM simple display pipeline helper only supports a single plane. In
order to prepare for support of the alpha plane on i.MX6SX and i.MX7,
move away from the helper. No new feature is added.
drm: mxsfb: Remove register definitions from mxsfb_crtc.c
mxsfb_crtc.c defines several macros related to register addresses and
bit, which duplicates macros from mxsfb_regs.h. Use the macros from
mxsfb_regs.h instead and remove them.
drm: mxsfb: Use LCDC_CTRL register name explicitly
The LCDC_CTRL register is located at address 0x0000. Some of the
accesses to the register simply use the mxsfb->base address. Reference
the LCDC_CTRL register explicitly instead to clarify the code.
drm: mxsfb: Clarify format and bus width configuration
Replace the convoluted way to set the format and bus width through
difficult to read macros with more explicit ones. Also remove the
outdated comment related to the limitations on bus width setting as it
doesn't apply anymore (the bus width can be specified through the
display_info bus format).
drm: mxsfb: Remove unused macros from mxsfb_regs.h
mxsfb_regs.h defines macros related to register bits. Some of them are
not used and don't clearly map to any particular register, so their
purpose isn't known. Remove them.
drm: mxsfb: Use BIT() macro to define register bitfields
Using BIT() is preferred over manual shifts as it's more readable,
handles the 1 << 31 case properly, and avoids other mistakes as shown by
the DEBUG0_HSYNC and DEBUG0_VSYNC bits (that are currently unused). Use
it.
Replace the manual connector implementation based on drm_panel with the
drm_panel_bridge helper. This simplifies the mxsfb driver by removing
connector-related code, and standardizing all pipeline control
operations on bridges.
A hack is needed to get hold of the connector, as that's our only source
of bus flags and formats for now. As soon as the bridge API provides us
with that information this can be fixed.
Commit 8e93f1028d74 ("drm/mxsfb: Use drm_fbdev_generic_setup()")
replaced fbdev handling with drm_fbdev_generic_setup() but left
inclusion of the drm/drm_fb_cma_helper.h header. Remove it.
Christian König [Thu, 23 Jul 2020 08:58:12 +0000 (10:58 +0200)]
drm/ttm: fix pipelined gutting for evictions v2
We can't pipeline that during eviction because the memory needs
to be available immediately.
v2: fix how we cleanup the BOs resources
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Tested-by: Alex Sierra <alex.sierra@amd.com> Link: https://patchwork.freedesktop.org/patch/379039/
Paul Menzel [Fri, 3 Jul 2020 14:29:39 +0000 (16:29 +0200)]
drm/amdgpu: Change type of module param `ppfeaturemask` to hexint
The newly added hexint helper is more convenient for bitmasks.
Before:
$ more /sys/module/amdgpu/parameters/ppfeaturemask 4294950911
After:
$ more /sys/module/amdgpu/parameters/ppfeaturemask
0xffffbfff
Cc: amd-gfx@lists.freedesktop.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Christian König <christian.koenig@amd.com> Link: https://patchwork.freedesktop.org/patch/374724/
Paul Menzel [Fri, 3 Jul 2020 14:29:38 +0000 (16:29 +0200)]
moduleparams: Add hexint type parameter
For bitmasks printing values in hex is more convenient.
Prefix with `0x` to make it clear, that it’s a hex value, and pad it
out.
Using the helper for `amdgpu.ppfeaturemask`, it will look like below.
Before:
$ more /sys/module/amdgpu/parameters/ppfeaturemask 4294950911
After:
$ more /sys/module/amdgpu/parameters/ppfeaturemask
0xffffbfff
Cc: linux-kernel@vger.kernel.org Cc: amd-gfx@lists.freedesktop.org Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Christian König <christian.koenig@amd.com> Link: https://patchwork.freedesktop.org/patch/374726/
Paul Menzel [Fri, 3 Jul 2020 14:29:37 +0000 (16:29 +0200)]
kernel/params.c: Align last argument with a tab
The second and third arguments are aligned with tabs, so do the same for
the fourth.
Cc: linux-kernel@vger.kernel.org Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Christian König <christian.koenig@amd.com> Link: https://lore.kernel.org/patchwork/patch/1267600/
Sam Ravnborg [Mon, 27 Jul 2020 17:03:20 +0000 (19:03 +0200)]
drm/bridge: nxp-ptn3460: add drm_panel_bridge support
Prepare the bridge driver for use in a chained setup.
- Replacing direct use of drm_panel with drm_panel_bridge support.
- Make the connector creation optional
Note: the bridge panel will use the connector type from the panel.
v3:
- Fix wrong logic in connector creation (Laurent)
v2:
- Use panel_bridge for local variable name to align with other drivers
- Fix that connector was created twice (Laurent)
- Set bridge.type to DRM_MODE_CONNECTOR_LVDS.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Link: https://patchwork.freedesktop.org/patch/msgid/20200727170320.959777-6-sam@ravnborg.org
Sam Ravnborg [Mon, 27 Jul 2020 17:03:18 +0000 (19:03 +0200)]
drm/bridge: tc358767: add drm_panel_bridge support
With the bridge operations implemented the last step to prepare
this driver for a chained setup is the use of the bridge panel driver.
The bridge panel driver is only used when a port@2 is present in
the DT. So when the display driver requests a connector
support both situations:
- connector created by bridge panel driver
- connector created by this driver
And on top, support that the display driver creates the connector,
which is the preferred setup.
Note: the bridge panel will use the connector type from the panel.
v3:
- Fix wrong logic in connector creation (Laurent)
v2:
- Merge connector and drm_panel_bridge patches
and fix so we do not create two connectors (Laurent)
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Link: https://patchwork.freedesktop.org/patch/msgid/20200727170320.959777-4-sam@ravnborg.org
To prepare for a chained bridge setup add support for the
get_edid bridge operation.
There is no need for a copy of the edid - so drop
the pointer to the copy.
v2:
- Fix so we do not leak memory (Laurent)
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Peter Senna Tschudin <peter.senna@gmail.com> Cc: Martin Donnelly <martin.donnelly@ge.com> Cc: Martyn Welch <martyn.welch@collabora.co.uk> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Link: https://patchwork.freedesktop.org/patch/msgid/20200726203324.3722593-13-sam@ravnborg.org
Sam Ravnborg [Sun, 26 Jul 2020 20:33:17 +0000 (22:33 +0200)]
drm/bridge: parade-ps8622: add drm_panel_bridge support
Prepare the bridge driver for use in a chained setup by
replacing direct use of drm_panel with drm_panel_bridge support.
The connecter is now either created by the panel bridge or the display
driver. So all code for connector creation in this driver is no longer
relevant and thus dropped.
The connector code had some special polling handling:
connector.polled = DRM_CONNECTOR_POLL_HPD;
drm_helper_hpd_irq_event(ps8622->bridge.dev);
This code was most likely added to speed up detection of the connector.
If really needed then this functionality belongs somewhere else.
Note: the bridge panel will use the connector type from the panel.
v2:
- Fix to avoid creating connector twice (Laurent)
- Drop all connector code - defer to bridge panel
- Use panel_bridge for local variable to align with other drivers
- Set bridge.type to DRM_MODE_CONNECTOR_LVDS;
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Link: https://patchwork.freedesktop.org/patch/msgid/20200726203324.3722593-9-sam@ravnborg.org
Sam Ravnborg [Sun, 26 Jul 2020 20:33:13 +0000 (22:33 +0200)]
drm/bridge: tc358764: add drm_panel_bridge support
Prepare the tc358764 bridge driver for use in a chained setup by
replacing direct use of drm_panel with drm_panel_bridge support.
The bridge panel will use the connector type reported by the panel,
where the connector for this driver hardcodes DRM_MODE_CONNECTOR_LVDS.
The tc358764 did not add any additional info the the connector so the
connector creation is passed to the bridge panel driver.
v3:
- Merge with patch to make connector creation optional to avoid
creating two connectors (Laurent)
- Pass connector creation to bridge panel, as this bridge driver
did not add any extra info to the connector.
- Set bridge.type to DRM_MODE_CONNECTOR_LVDS.
v2:
- Use PTR_ERR_OR_ZERO() (kbuild test robot)
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: kbuild test robot <lkp@intel.com> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Link: https://patchwork.freedesktop.org/patch/msgid/20200726203324.3722593-5-sam@ravnborg.org
Sam Ravnborg [Sun, 26 Jul 2020 20:33:12 +0000 (22:33 +0200)]
drm/bridge: tc358764: drop drm_connector_(un)register
Drop drm_connector handling that is not needed:
- drm_dev_register() in the display controller driver takes
care of registering connectors.
So the call to drm_connector_register() call is not needed in the bridge
driver.
- Use of drm_connector_unregister() is only required for drivers that
explicit have called drm_dev_register.
- The reference counting using drm_connector_put() is likewise not needed.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> Cc: Jonas Karlman <jonas@kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec@siol.net> Link: https://patchwork.freedesktop.org/patch/msgid/20200726203324.3722593-4-sam@ravnborg.org
Warn if we detect a panel with incomplete/wrong description.
This is inspired by a similar patch by Laurent that introduced checks
for LVDS panels - this extends the checks to the remaining type of
connectors.
This is known to warn for some of the existing panels but added
despite this as we need help from people using the panels to
add the missing info.
The checks are not complete but will catch the most common mistakes.
The checks at the same time serve as documentation for the minimum
required description for a panel.
The checks uses dev_warn() as we know this will hit. WARN() was
too noisy at the moment for anything else than LVDS.
v3:
- %d => %u for bpc (Laurent)
v2:
- Use dev_warn (Laurent)
- Check for empty bus_flags
Ondrej Jirman [Fri, 3 Jul 2020 11:47:16 +0000 (13:47 +0200)]
dt-bindings: display: Fix example in nwl-dsi.yaml
The example is now validated against rocktech,jh057n00900 schema
that was ported to yaml, and didn't validate with:
- '#address-cells', '#size-cells', 'port@0' do not match any of
the regexes: 'pinctrl-[0-9]+'
- 'vcc-supply' is a required property
- 'iovcc-supply' is a required property
- 'reset-gpios' is a required property
Lyude Paul [Mon, 20 Jul 2020 19:07:36 +0000 (15:07 -0400)]
drm/vblank: Use spin_(un)lock_irq() in drm_crtc_queue_sequence_ioctl()
This is an ioctl callback, so we're guaranteed to have IRQs enabled when
calling this function. Use the plain _irq() variants of spin_(un)lock()
to make this more obvious.
Lyude Paul [Mon, 20 Jul 2020 19:07:35 +0000 (15:07 -0400)]
drm/vblank: Use spin_(un)lock_irq() in drm_queue_vblank_event()
This one's easy - we're already calling kzalloc(GFP_KERNEL) in this
function, so we must already be guaranteed to have IRQs enabled when
calling this. So, use the plain _irq() variants of spin_(un)lock() to
make this more obvious.
Lyude Paul [Mon, 20 Jul 2020 19:07:34 +0000 (15:07 -0400)]
drm/vblank: Use spin_(un)lock_irq() in drm_legacy_vblank_post_modeset()
This function is only ever called from ioctl context, so we're
guaranteed to have interrupts enabled. Stop using the irqsave/irqrestore
variants of spin_(un)lock_irq() to make this more obvious.
Lyude Paul [Mon, 20 Jul 2020 19:07:32 +0000 (15:07 -0400)]
drm/vblank: Use spin_(un)lock_irq() in drm_crtc_vblank_reset()
All of the drivers in the kernel tree only call this from one of the
following contexts:
* drm_crtc_funcs->reset
* During initial module load
Since both of these contexts are guaranteed to have interrupts enabled
beforehand, there's no need to use the irqsave/irqrestore variants of
spin_(un)lock(). So, fix this to make the irq context of this function
more obvious.
Christian König [Thu, 16 Jul 2020 08:15:17 +0000 (10:15 +0200)]
drm/vmwgfx: stop using TTM_MEMTYPE_FLAG_MAPPABLE
The driver doesn't expose any not-mapable memory resources.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/378241/
Christian König [Fri, 10 Jul 2020 13:13:41 +0000 (15:13 +0200)]
drm: remove optional dummy function from drivers using TTM
Implementing those is completely unnecessary.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Madhav Chauhan <madhav.chauhan@amd.com> Link: https://patchwork.freedesktop.org/patch/378236/
Daniel Vetter [Thu, 9 Jul 2020 12:33:38 +0000 (14:33 +0200)]
dma-buf.rst: Document why indefinite fences are a bad idea
Comes up every few years, gets somewhat tedious to discuss, let's
write this down once and for all.
What I'm not sure about is whether the text should be more explicit in
flat out mandating the amdkfd eviction fences for long running compute
workloads or workloads where userspace fencing is allowed.
v2: Now with dot graph!
v3: Typo (Dave Airlie)
Reviewed-by: Thomas Hellstrom <thomas.hellstrom@intel.com> Acked-by: Jason Ekstrand <jason@jlekstrand.net> Acked-by: Christian König <christian.koenig@amd.com> Acked-by: Daniel Stone <daniels@collabora.com> Acked-by: Dave Airlie <airlied@redhat.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Jesse Natalie <jenatali@microsoft.com> Cc: Steve Pronovost <spronovo@microsoft.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Cc: Felix Kuehling <Felix.Kuehling@amd.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Thomas Hellstrom <thomas.hellstrom@intel.com> Cc: linux-media@vger.kernel.org Cc: linaro-mm-sig@lists.linaro.org Cc: linux-rdma@vger.kernel.org Cc: amd-gfx@lists.freedesktop.org Cc: intel-gfx@lists.freedesktop.org Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200709123339.547390-1-daniel.vetter@ffwll.ch
Daniel Vetter [Tue, 7 Jul 2020 20:12:06 +0000 (22:12 +0200)]
dma-fence: prime lockdep annotations
Two in one go:
- it is allowed to call dma_fence_wait() while holding a
dma_resv_lock(). This is fundamental to how eviction works with ttm,
so required.
- it is allowed to call dma_fence_wait() from memory reclaim contexts,
specifically from shrinker callbacks (which i915 does), and from mmu
notifier callbacks (which amdgpu does, and which i915 sometimes also
does, and probably always should, but that's kinda a debate). Also
for stuff like HMM we really need to be able to do this, or things
get real dicey.
Consequence is that any critical path necessary to get to a
dma_fence_signal for a fence must never a) call dma_resv_lock nor b)
allocate memory with GFP_KERNEL. Also by implication of
dma_resv_lock(), no userspace faulting allowed. That's some supremely
obnoxious limitations, which is why we need to sprinkle the right
annotations to all relevant paths.
The one big locking context we're leaving out here is mmu notifiers,
added in
mm/mmu_notifiers: add a lockdep map for invalidate_range_start/end
that one covers a lot of other callsites, and it's also allowed to
wait on dma-fences from mmu notifiers. But there's no ready-made
functions exposed to prime this, so I've left it out for now.
v2: Also track against mmu notifier context.
v3: kerneldoc to spec the cross-driver contract. Note that currently
i915 throws in a hard-coded 10s timeout on foreign fences (not sure
why that was done, but it's there), which is why that rule is worded
with SHOULD instead of MUST.
Also some of the mmu_notifier/shrinker rules might surprise SoC
drivers, I haven't fully audited them all. Which is infeasible anyway,
we'll need to run them with lockdep and dma-fence annotations and see
what goes boom.
v4: A spelling fix from Mika
v5: #ifdef for CONFIG_MMU_NOTIFIER. Reported by 0day. Unfortunately
this means lockdep enforcement is slightly inconsistent, it won't spot
GFP_NOIO and GFP_NOFS allocations in the wrong spot if
CONFIG_MMU_NOTIFIER is disabled in the kernel config. Oh well.
v5: Note that only drivers/gpu has a reasonable (or at least
historical) excuse to use dma_fence_wait() from shrinker and mmu
notifier callbacks. Everyone else should either have a better memory
manager model, or better hardware. This reflects discussions with
Jason Gunthorpe.
Cc: Jason Gunthorpe <jgg@mellanox.com> Cc: Felix Kuehling <Felix.Kuehling@amd.com> Cc: kernel test robot <lkp@intel.com> Acked-by: Christian König <christian.koenig@amd.com> Acked-by: Dave Airlie <airlied@redhat.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reviewed-by: Thomas Hellström <thomas.hellstrom@intel.com> (v4) Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Thomas Hellstrom <thomas.hellstrom@intel.com> Cc: linux-media@vger.kernel.org Cc: linaro-mm-sig@lists.linaro.org Cc: linux-rdma@vger.kernel.org Cc: amd-gfx@lists.freedesktop.org Cc: intel-gfx@lists.freedesktop.org Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200707201229.472834-3-daniel.vetter@ffwll.ch
Daniel Vetter [Tue, 7 Jul 2020 20:12:05 +0000 (22:12 +0200)]
dma-fence: basic lockdep annotations
Design is similar to the lockdep annotations for workers, but with
some twists:
- We use a read-lock for the execution/worker/completion side, so that
this explicit annotation can be more liberally sprinkled around.
With read locks lockdep isn't going to complain if the read-side
isn't nested the same way under all circumstances, so ABBA deadlocks
are ok. Which they are, since this is an annotation only.
- We're using non-recursive lockdep read lock mode, since in recursive
read lock mode lockdep does not catch read side hazards. And we
_very_ much want read side hazards to be caught. For full details of
this limitation see
- To allow nesting of the read-side explicit annotations we explicitly
keep track of the nesting. lock_is_held() allows us to do that.
- The wait-side annotation is a write lock, and entirely done within
dma_fence_wait() for everyone by default.
- To be able to freely annotate helper functions I want to make it ok
to call dma_fence_begin/end_signalling from soft/hardirq context.
First attempt was using the hardirq locking context for the write
side in lockdep, but this forces all normal spinlocks nested within
dma_fence_begin/end_signalling to be spinlocks. That bollocks.
The approach now is to simple check in_atomic(), and for these cases
entirely rely on the might_sleep() check in dma_fence_wait(). That
will catch any wrong nesting against spinlocks from soft/hardirq
contexts.
The idea here is that every code path that's critical for eventually
signalling a dma_fence should be annotated with
dma_fence_begin/end_signalling. The annotation ideally starts right
after a dma_fence is published (added to a dma_resv, exposed as a
sync_file fd, attached to a drm_syncobj fd, or anything else that
makes the dma_fence visible to other kernel threads), up to and
including the dma_fence_wait(). Examples are irq handlers, the
scheduler rt threads, the tail of execbuf (after the corresponding
fences are visible), any workers that end up signalling dma_fences and
really anything else. Not annotated should be code paths that only
complete fences opportunistically as the gpu progresses, like e.g.
shrinker/eviction code.
The main class of deadlocks this is supposed to catch are:
Thread A:
mutex_lock(A);
mutex_unlock(A);
dma_fence_signal();
Thread B:
mutex_lock(A);
dma_fence_wait();
mutex_unlock(A);
Thread B is blocked on A signalling the fence, but A never gets around
to that because it cannot acquire the lock A.
Note that dma_fence_wait() is allowed to be nested within
dma_fence_begin/end_signalling sections. To allow this to happen the
read lock needs to be upgraded to a write lock, which means that any
other lock is acquired between the dma_fence_begin_signalling() call and
the call to dma_fence_wait(), and still held, this will result in an
immediate lockdep complaint. The only other option would be to not
annotate such calls, defeating the point. Therefore these annotations
cannot be sprinkled over the code entirely mindless to avoid false
positives.
Originally I hope that the cross-release lockdep extensions would
alleviate the need for explicit annotations:
https://lwn.net/Articles/709849/
But there's a few reasons why that's not an option:
- It's not happening in upstream, since it got reverted due to too
many false positives:
locking/lockdep: Remove the cross-release locking checks
This code (CONFIG_LOCKDEP_CROSSRELEASE=y and CONFIG_LOCKDEP_COMPLETIONS=y),
while it found a number of old bugs initially, was also causing too many
false positives that caused people to disable lockdep - which is arguably
a worse overall outcome.
- cross-release uses the complete() call to annotate the end of
critical sections, for dma_fence that would be dma_fence_signal().
But we do not want all dma_fence_signal() calls to be treated as
critical, since many are opportunistic cleanup of gpu requests. If
these get stuck there's still the main completion interrupt and
workers who can unblock everyone. Automatically annotating all
dma_fence_signal() calls would hence cause false positives.
- cross-release had some educated guesses for when a critical section
starts, like fresh syscall or fresh work callback. This would again
cause false positives without explicit annotations, since for
dma_fence the critical sections only starts when we publish a fence.
- Furthermore there can be cases where a thread never does a
dma_fence_signal, but is still critical for reaching completion of
fences. One example would be a scheduler kthread which picks up jobs
and pushes them into hardware, where the interrupt handler or
another completion thread calls dma_fence_signal(). But if the
scheduler thread hangs, then all the fences hang, hence we need to
manually annotate it. cross-release aimed to solve this by chaining
cross-release dependencies, but the dependency from scheduler thread
to the completion interrupt handler goes through hw where
cross-release code can't observe it.
In short, without manual annotations and careful review of the start
and end of critical sections, cross-relese dependency tracking doesn't
work. We need explicit annotations.
v2: handle soft/hardirq ctx better against write side and dont forget
EXPORT_SYMBOL, drivers can't use this otherwise.
v3: Kerneldoc.
v4: Some spelling fixes from Mika
v5: Amend commit message to explain in detail why cross-release isn't
the solution.
v6: Pull out misplaced .rst hunk.
Acked-by: Christian König <christian.koenig@amd.com> Acked-by: Dave Airlie <airlied@redhat.com> Cc: Felix Kuehling <Felix.Kuehling@amd.com> Reviewed-by: Thomas Hellström <thomas.hellstrom@intel.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Thomas Hellstrom <thomas.hellstrom@intel.com> Cc: linux-media@vger.kernel.org Cc: linaro-mm-sig@lists.linaro.org Cc: linux-rdma@vger.kernel.org Cc: amd-gfx@lists.freedesktop.org Cc: intel-gfx@lists.freedesktop.org Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200707201229.472834-2-daniel.vetter@ffwll.ch
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Guido Günther [Mon, 23 Mar 2020 14:52:46 +0000 (15:52 +0100)]
drm/mxsfb: Make supported modifiers explicit
In contrast to other display controllers on imx like DCSS and ipuv3
lcdif/mxsfb does not support detiling e.g. vivante tiled layouts.
Since mesa might assume otherwise make it explicit that only
DRM_FORMAT_MOD_LINEAR is supported.