drm/gma500: Remove unnecessary suspend/resume wrappers

The psb_runtime_suspend/resume/thaw/freeze/restore functions are all
just 1:1 wrappers around gma_power_suspend/_resume.

Drop these wrappers and use the DEFINE_RUNTIME_DEV_PM_OPS() macro to
define the dev_pm_ops struct.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220909115646.99920-7-hdegoede@redhat.com

drm/gma500: Rewrite power management code

Rewrite the power.c code. For some reason this was doing locking +
refcounting + state (suspended or not) bookkeeping all by itself.

But there is no reason for this, this is all taken care of by
the runtime-pm core, through pm_runtime_get()/_put().

Besides this not being necessary the DIY code is also quite weird/
buggy in some places. E.g. power_begin() would manually do a resume
when not resumed already and force_on=true, followed by a
pm_runtime_get(), which will cause a call to gma_power_resume() to
get scheduled which would redo the entire resume again. Which can
all be replaced by a single pm_runtime_get_sync() call.

Note that this is just a cleanup, this does not actually fix
the (disabled through #if 0) runtime-pm support. It does now call
pm_runtime_enable(), but only after doing a pm_runtime_get() at
probe-time, so the device is never runtime suspended.

Doing this permanent get() + enable() instead of not calling
enable() at all is necessary for the pm_runtime_get_if_in_use() call
in gma_power_begin() to work properly.

Note this also removes the gma_power_is_on() call a check like this
without actually holding a reference is always racy, so it is a bad
idea (and therefor has no pm_runtime_foo() equivalent).

The 2 code paths which were using gma_power_is_on() are actually both
guaranteed to only run when the device is powered-on so the 2 checks
can simply be dropped.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220909115646.99920-6-hdegoede@redhat.com

drm/gma500: Fix (vblank) IRQs not working after suspend/resume

Fix gnome-shell (and other page-flip users) hanging after suspend/resume
because of the gma500's IRQs not working.

This fixes 2 problems with the IRQ handling:

1. gma_power_off() calls gma_irq_uninstall() which does a free_irq(), but
gma_power_on() called gma_irq_preinstall() + gma_irq_postinstall() which
do not call request_irq. Replace the pre- + post-install calls with
gma_irq_install() which does prep + request + post.

2. After fixing 1. IRQs still do not work on a Packard Bell Dot SC (Intel
Atom N2600, cedarview) netbook.

Cederview uses MSI interrupts and it seems that the BIOS re-configures
things back to normal APIC based interrupts during S3 suspend. There is
some MSI PCI-config registers save/restore code which tries to deal with
this, but on the Packard Bell Dot SC this is not sufficient to restore
MSI IRQ functionality after a suspend/resume.

Replace the PCI-config registers save/restore with pci_disable_msi() on
suspend + pci_enable_msi() on resume. Fixing e.g. gnome-shell hanging.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906203852.527663-4-hdegoede@redhat.com
(cherry picked from commit 235fdbc32d559db21e580f85035c59372704f09e)

drm/gma500: Fix (vblank) IRQs not working after suspend/resume

Fix gnome-shell (and other page-flip users) hanging after suspend/resume
because of the gma500's IRQs not working.

This fixes 2 problems with the IRQ handling:

1. gma_power_off() calls gma_irq_uninstall() which does a free_irq(), but
gma_power_on() called gma_irq_preinstall() + gma_irq_postinstall() which
do not call request_irq. Replace the pre- + post-install calls with
gma_irq_install() which does prep + request + post.

2. After fixing 1. IRQs still do not work on a Packard Bell Dot SC (Intel
Atom N2600, cedarview) netbook.

Cederview uses MSI interrupts and it seems that the BIOS re-configures
things back to normal APIC based interrupts during S3 suspend. There is
some MSI PCI-config registers save/restore code which tries to deal with
this, but on the Packard Bell Dot SC this is not sufficient to restore
MSI IRQ functionality after a suspend/resume.

Replace the PCI-config registers save/restore with pci_disable_msi() on
suspend + pci_enable_msi() on resume. Fixing e.g. gnome-shell hanging.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906203852.527663-4-hdegoede@redhat.com

drm/gma500: Cosmetic cleanup of irq code

Use the gma_ prefix instead of psb_ since the code is common for all
chips. Various coding style fixes. Removal of unused code. Removal of
duplicate function declarations.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20220317092555.17882-4-patrik.r.jakobsson@gmail.com

drm/gma500: Move GEM memory management functions to gem.c

Move GEM functions from gtt.c to gem.c. Adapt some names. No
functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220308195222.13471-11-tzimmermann@suse.de

drm/gma500: Inline psb_gtt_restore()

Inline psb_gtt_restore() into its only caller in power.c.

Perform the GTT restoration in psb_gem_mm_resume(). The restoration
step is part of GEM anyway and will be moved over at some point.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220308195222.13471-10-tzimmermann@suse.de

drm/gma500: Replace references to dev_private with helper function

Replace most references to struct drm_device.dev_private with the new
helper function to_drm_psb_private(). The only references left are in
assignments and the helper itself.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210920141051.30988-2-tzimmermann@suse.de

drm/gma500: Convert to Linux IRQ interfaces

Drop the DRM IRQ midlayer in favor of Linux IRQ interfaces. DRM's
IRQ helpers are mostly useful for UMS drivers. Modern KMS drivers
don't benefit from using it. DRM IRQ callbacks are now being called
directly or inlined.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210803090704.32152-6-tzimmermann@suse.de

gma500: Use DEFINE_SPINLOCK() for spinlock

spinlock can be initialized automatically with DEFINE_SPINLOCK()
rather than explicitly calling spin_lock_init().

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Guobin Huang <huangguobin4@huawei.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/1617710114-48071-1-git-send-email-huangguobin4@huawei.com

drm/gma500/power: Remove excess param description 'state'

Fixes the following W=1 kernel build warning(s):

drivers/gpu/drm/gma500/power.c:190: warning: Excess function parameter 'state' description in 'gma_power_suspend'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Benjamin Defnet <benjamin.r.defnet@intel.com>
Cc: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
Cc: Alan Cox <alan@linux.intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-39-lee.jones@linaro.org

drm/gma500/power: Fix a bunch of *dev documentation issues

Fixes the following W=1 kernel build warning(s):

drivers/gpu/drm/gma500/power.c:101: warning: Function parameter or member 'pdev' not described in 'gma_resume_display'
drivers/gpu/drm/gma500/power.c:155: warning: Function parameter or member 'pdev' not described in 'gma_resume_pci'
drivers/gpu/drm/gma500/power.c:155: warning: Excess function parameter 'dev' description in 'gma_resume_pci'
drivers/gpu/drm/gma500/power.c:189: warning: Function parameter or member '_dev' not described in 'gma_power_suspend'
drivers/gpu/drm/gma500/power.c:189: warning: Excess function parameter 'pdev' description in 'gma_power_suspend'
drivers/gpu/drm/gma500/power.c:189: warning: Excess function parameter 'state' description in 'gma_power_suspend'
drivers/gpu/drm/gma500/power.c:216: warning: Function parameter or member '_dev' not described in 'gma_power_resume'
drivers/gpu/drm/gma500/power.c:216: warning: Excess function parameter 'pdev' description in 'gma_power_resume'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Benjamin Defnet <benjamin.r.defnet@intel.com>
Cc: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
Cc: Alan Cox <alan@linux.intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-20-lee.jones@linaro.org

drm/gma500: Remove references to struct drm_device.pdev

Using struct drm_device.pdev is deprecated. Convert gma500 to struct
drm_device.dev. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201201103542.2182-8-tzimmermann@suse.de

drm/gma500: use to_pci_dev()

Use to_pci_dev() instead of open-coding it.

Signed-off-by: Geliang Tang <geliangtang@163.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/gma500: Add hooks for hibernation

Currently we do whatever is done during suspend/resume but we might need some
more work for hibernation so keep them in separate functions.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>

drm/gma500: Activate the gtt rebuild on suspend/resume

This patch activates the rebuilding of the gtt. Currently we reinitialize the
gtt by inserting the stolen pages again and map the rest to our scratch page.
Then we go about restoring the needed ranges. This is a bit overkill but right
now we don't have that much to restore so better safe than sorry.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>

gma500: suspend/resume support for Cedartrail

Update our tree to match the current driver head.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: Fix resume paths

We fall apart somewhat on resume because we don't invoke all the resume
methods as we should. Fix the silly error in the logic.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: make init_pm callback in struct psb_op optional

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: now move the Oaktrail save state into its own structure

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: Fix backlight crash

Initial changes to get backlight behaviour we want and to fix backlight crashes
on suspend/resume paths.

[Note: on some boxes this will now produce a warning about the backlight, this
isn't a regression it's an unfixed but non harmful case I still need to nail]

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: Add device framework

The devices have various internal differences so we have some abstractions
to hide the ugly differences and we then wrap them up in standard
interfaces. Add these bits

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>