Add a description of the bindings used by Kirin 960/970 Display
Serial Interface (DSI) controller and by its Display Engine (DPE).

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This is part of support for a panel display. Those don't come
with the Hikey 960 or 970 boards. As I don't have any of
those for tests, and we didn't port another required driver
for this to work, for now, let's drop it.

This patch can be reversed later, if one would be adding
support for it.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Add the needed bits for the DRM driver to work with the
Hikey 970 board.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Add a device tree for the HiSilicon 6421v600 SPMI PMIC, used
on HiKey970 board.

As we now have support for it, change the fixed regulators
used by the SD I/O to use the proper LDO supplies.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are several pinctrl settings that are missing at this
DT file.

Also, the entries are out of order.

Add the missing bits, as they'll be required by the DRM driver - and
probably by other drivers not upstreamed yet.

Reorder the entres, adding the missing bits.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The I2C buses are not declared at the device tree. As this will
be needed by further patches, add them, keeping all in
disabled state. Per-board settings can override it.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are lots of ifdefs checking if the SoC version is
960 or 970. Replace them all by runtime definitions.

With that, the same DRM driver should work with both versions.

The behavior will dynamically change depending on the
OF compatible strings.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are lots of things that are identical for both
Kirin 960 and Kirin 970. Place those identical ones on a
common file, removing unused structs.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are some things inside struct dss_hw_ctx that are unused.
Get rid of them.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The DPE headers define several macros for I/O. Get rid of
them by replacing by the Linux ones.

In the specific case of outp32(), I used this small
coccinelle script to change them to writel():

	@ rule1 @
	expression addr, val;
	@@
	-outp32(addr, val)
	+writel(val, addr)

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are a few typedefs inside this driver. Get rid of them.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Now that everything is in place, add the driver to the
building system.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

While this driver apparently supports both IOMMU and no-IOMMU
access, it always enable the IOMMU via some code, at the
downstream version.

Apparently, the downstream iommu is there just to get the
physical address of the logical IOMMU address. Based on
the downstream code, it sounds that the IOMMU would be an
specific one for the GPU.

Anyway, right now, the driver is set to not use the IOMMU
at all. So, let's comment out the code which allocates
IOMMU pages, and the code that would try to use it to
setup a register, as, without the IOMMU, this would cause
an OOPS.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are lots of coding style issues on this driver, as
reported by checkpatch.

Address most of them.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Use SPMI headers to identify the license of each file.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This is needed for the DRM to work. Ok, right now, this is
enabled on default fastboot logic, but, as soon as we enable
the proper SPMI driver, such code is needed.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Do some cleanups at the mode validation code. Right now, there
is a known issue with this driver which makes it to set up
some invalid modes, depending on the display.

We don't know yet what the issue is, so, instead, let's add
a table with the modes which are known to work.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Those aren't needed anymore.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

call drm_connector_register() when initializing the connector.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Currently, the method is empty. However, looking at the driver,
it sounds it shouldn't be hard to implement it.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

At least on current upstream Kernels, forcing an event parse
during init time isn't needed anymore.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Those methods belong to crtc fops

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Instead of implementing it at the code, use the default
methods from CMA helper

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

- Get rid of a global var meant to store one of its priv
  structs;
- Change the name of the driver, in order to not be confused with
  the kirin6220;
- Remove some unneeded ifdef;
- use drm_of.h helper.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The logich which sets the MIPI parameters is currently wrong:
it is using a value stored at cur_client, with actually points
to the active location, and not to the one that it is about
to be initialized.

The entire logic sounds buggy, but for now let's just keep
following it, by adding an extra var that will tell what was
the latest attached encoder.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The logic there is more complex than it needs. It also places
the device with a wrong setting if the flags are missed.

This currently happens on Kirin970 for HDMI, as there's a bug
at the part of the driver which selects between PANEL or
OUTPUT at encoder init code.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Place both adv7535 and panel logic at the same routine,
cleaning up things a little bit and fixing the includes.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Allocate the framebuffer memory via CMA, as otherwise the
drm driver may not work properly with X11.

Part of the changes here were based on a patch originally
authored by:

	alik <hou.xiaoyong@hihope.org>

The original version can be found at:
	https://github.com/Bigcountry907/linux/commit/046e29834ef1c523c73614747377d3660eec3964



Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This is not used anymore on this version, so let's get rid
of them.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This function has a too generic name to export it as a
symbol. Also, we should likely use some other macro instead.

So, for now, just copy it into the Kirin9xx dsi module, in order
for the driver to build.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There were several changes at the upstream kAPIs since
Kernel 4.4. Update the driver for it to build with the
upstream Kernel.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The includes there reflect a downstream version back on v4.4
times. change them to reflect the current upstream and to
avoid the need of using a -I flag at the Makefile.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Use the same standard as used on other Hisilicon DRM
config vars for kirin9xx.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are some #ifdefs there for non-existing CONFIG_ options
(nor even at the downstream code).

Let's get rid of those. It can be re-added later if ever needed.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

We don't need to implement legacy fbdev support. Just use
the FB emulation instead.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There's already a driver with the same namespace
for an older Kirin chipset. Rename this one, in order
to make it clearer that this is the driver for
Kirin 960/970.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The OOT had a fork of adv7535 with some changes for it to
work with a failing EDID retrival I/O.

Get rid of it, as we'll be using the upstream driver.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Xiubin Zhang authored 6 years ago and Mauro Carvalho Chehab committed 4 years ago

Adjust pixel clock for compatibility with 10.1 inch special displays.

Signed-off-by: Xiubin Zhang <zhangxiubin1@huawei.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Xiubin Zhang authored 6 years ago and Mauro Carvalho Chehab committed 4 years ago

Add some debug prints on adv7535 and kirin_drm_drv.

Signed-off-by: Xiubin Zhang <zhangxiubin1@huawei.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>