Signed-off-by: Angus Ainslie <angus@akkea.ca>

Enabling these causes USB errors on the host and device side. On the L5 we see

SDP: initialize...
SDP: handle requests...
Status: -108
Status: -108
Status: -108

And on the host side there are messages about failing to negotiate and address.

Signed-off-by: Angus Ainslie <angus@akkea.ca>

Closes #23

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

Fixes OS-issues#217 and design#24

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

The datasheet says that it can take up to 1 second for conversion to be
finished, so there's no point in waiting for longer.

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

Also, disable HVDCP and MaxCharge to not negotiate high voltages
behind TPS65982's back.

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

Although the bit is called STAT_DIS (STAT Pin Disable), we're setting it
to 0, making the comment a bit confusing.

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

The charger is disabled anyway, and once it gets re-enabled
the value is already overwritten.

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

We don't use them.

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

read_vbat does additional stuff like disabling the charger and enabling
BAT_LOADEN to get more accurate reading, so use it to read the VBAT value.

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

According to the datasheet, the conversion is not finished until CONV_START
gets cleared, which can take up to 1 second (8.2.8 Battery Monitor).

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

Not sure what was the intention behind the old code, it doesn't look right.

Sebastian Krzyszkowiak authored 3 years ago and Angus Ainslie committed 3 years ago

The next line is supposed to restore register 0x03 to the old value, which
has no chance to work if we had just stored its current value into reg03.

debian: Bump firmware to 8.12

See merge request !42

Enable WiFi SW kill switch

See merge request !40

Due to probe ordering in the kernel the SW kill switch might not get enabled
before the wifi driver is probed.

Signed-off-by: Angus Ainslie <angus@akkea.ca>

configs: librem5: set BOOTDELAY=0

Closes #18

See merge request !24

add the u-boot version to the command line

See merge request !41

Pass the u-boot version to the kernel

Signed-off-by: Angus Ainslie <angus@akkea.ca>

Update firmware 8.10

See merge request !39

arm-trusted-firmware#1

This bumps the DP firmware version to 50d385ab:

    [   65.342684] cdns-mhdp-imx 32c00000.hdmi: [drm:cdns_mhdp_read_fw_version [cdns_mhdp_drmcore]] firmware version: 50d385ab

imx8mq_pins: fix configuration for UART4 on ECSPI2 pads

See merge request !38

Move the board rev to the MAC address fuses

See merge request !37

When routing UART4 using the ECSPI2 pads, register
IOMUXC_UART4_RXD_SELECT_INPUT (offset 0x050C) should be changed only
when dealing with RX, as its name suggests.

NXP has changed the layout of the OTP fuses and the user area is no longer
writable so we'll overload the MAC address as a board revision number.

Signed-off-by: Angus Ainslie <angus@akkea.ca>

imx8: Use 8.8 firmware

See merge request !36

Use DisplayPort on phone and HDMI on Devkit

See merge request !35

Also support model names that carry  revision number at the end.

Remove redundant YYLOC global declaration

See merge request !33

Add Debian packaging

Closes #20 and image-builder#3

See merge request !34

Can be removed once we switch to bullseye