The sound_osc and mclk nodes aren't actually used by anything.
Removing them firstly simplies the .dtsi and secondly avoids any
potential confusion.

The clock-names property isn't actually used by the drivers.  Removing
it firstly simplies the .dtsi and secondly avoids any potential
confusion.

While the previous clocks property worked, this patch makes it reflect
the actual clock tree parentage, reducing confusion.

While the previous assigned-clock properties worked, this patch makes
them reflect the actual clock tree parentage, reducing confusion.

Swap SAI6_RXD0 and SAI6_TXD0

See merge request Librem5/linux-emcraft!20

To enable the diversity set the module parameter antenna_diversity=1

insmod rsi_91x.ko rsi_zone_enabled=1 dev_oper_mode=13 antenna_diversity=1

librem5-evk: Enable device mapper

Closes #29

See merge request Librem5/linux-emcraft!14

Unbreak built in LCD panel

Closes #5

See merge request Librem5/linux-emcraft!18

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

Closes: #5

Thankfully they're not needed to drive the panel.

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

This makes it simple to validate panels and check basic functionality.

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

This works around packet corruption by zeros in [8:24].

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

Move it into the irq handler since the _finished function
is only for finished packets.

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

Built a deb when running gitlab-ci

See merge request Librem5/linux-emcraft!17

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

gitlab won't fetch artifacts from outside the tree so copy back the
built deb.

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

Unbreak touch for more people

See merge request Librem5/linux-emcraft!16

From what I've seen in vendor trees it's fine to treat this as gt1x¹.
Tested on the Purism Librem 5 Devkit (Rocktech JH057N00900 panel).

[1]: https://github.com/TadiT7/android_kernel_mtk-4.4/tree/master/drivers/input/touchscreen/mediatek/GT5688



This folds in the rest of the patch sent upstream making the number of
touch points probe reliably.

Signed-off-by: Guido Günther <agx@sigxcpu.org>

"Invalid config" give little idea what's wrong. Print the values that
should be != 0 so we know which one is off.

Signed-off-by: Guido Günther <agx@sigxcpu.org>

Add -lcdusbhost.dts, combining -lcdonly and -usbhost

See merge request Librem5/linux-emcraft!15

This got lost during the defconfig conversion

Closes: #29
Signed-off-by: Guido Günther <guido.gunther@puri.sm>

Drm updates

See merge request Librem5/linux-emcraft!13

Lucas Stach authored 6 years ago and Guido Gunther committed 6 years ago

This is the only place in the driver that should have to deal with
the raw hardware fences. To avoid any further confusion, consolidate
the fence handling in this file and remove any traces of this from
the header files.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>

Lucas Stach authored 6 years ago and Guido Gunther committed 6 years ago

There is no need to track the currently active fence. The GPU scheduler
keeps track of all the in-flight jobs.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>

Lucas Stach authored 6 years ago and Guido Gunther committed 6 years ago

The GPU hardware fences and the job out-fences are on different timelines
so it's wrong to compare them. Fix this by only looking at the out-fence.

Cc: <stable@vger.kernel.org>
Fixes: 2c83a726 (drm/etnaviv: bring back progress check in job
                     timeout handler)
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>

Lucas Stach authored 6 years ago and Guido Gunther committed 6 years ago

The documentation of drm_sched_job_init and drm_sched_entity_push_job has
been clarified. Both functions should be called under a shared lock, to
avoid jobs getting pushed into the scheduler queue in a different order
than their sched_fence seqnos, which will confuse checks that are looking
at the seqnos to infer information about completion order.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>

Philipp Zabel authored 6 years ago and Guido Gunther committed 6 years ago

Replace the open-coded scratch page initialization loop with memset32

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>

Use mainline goodix driver

Closes #23

See merge request Librem5/linux-emcraft!12

# Conflicts:
#   arch/arm64/configs/librem5-evk_defconfig

f/imx8-4.18/drm updates: Tie devkit panel to backlight

See merge request Librem5/linux-emcraft!11

librem5-evk-lcdonly: Bind backlight device to panel

See merge request Librem5/linux-emcraft!10

librem5-evk: Strip down defconfig

See merge request Librem5/linux-emcraft!8

We're using the mainline driver now.

This reverts commit 14aab0cf.

We need to mark the GPIO pins as active high since other wise the reset
sequence is inverted.

Using the mainline driver gives us the right scaling.

Closes: #23

Signed-off-by: Guido Günther <guido.gunther@puri.sm>

Signed-off-by: Guido Günther <guido.gunther@puri.sm>