It seems kernel version starting from 4.19.90 won't boot on my pyra. 4.19.89 is the last one that works.
They all get stuck on "Starting kernel ..." in uboot.
I've tried the latest letux image too, to confirm it's not something in my compile/packaging process, or custom uboot, but that doesn't boot either ( kernel Linux-4.19.95-letux-lpae+ , with omap5-letux-cortex15-v5.3+pyra-v5.2.dtb )
I've also tried the 5.4.11 kernel, that boots , but /sys/devices/platform/txs02612/switch doesn't exist there, so it can't find the emmc.
(and the display turns off after a second or so, but i think that was a known problem for the 5.x kernels)
Am 18.01.2020 um 22:28 schrieb aTc atc@k-n-p.org:
It seems kernel version starting from 4.19.90 won't boot on my pyra. 4.19.89 is the last one that works.
They all get stuck on "Starting kernel ..." in uboot.
Do you happen to have one of the v5.1 boards (please check which .dtb file is loaded by u-boot)? Then, there is a change starting with 4.19.90. There are now two different .dtb files. One for v5.1.0.dtb and one for v5.1.2.dtb boards. The reason is that we should have incremented the board version ID resistors for automatic detection but it was forgotten. The solution is to rename the right one to v5.1.dtb
I've tried the latest letux image too, to confirm it's not something in my compile/packaging process, or custom uboot, but that doesn't boot either ( kernel Linux-4.19.95-letux-lpae+ , with omap5-letux-cortex15-v5.3+pyra-v5.2.dtb )
Hm. Looks as if you have one of the latest ones (I am not sure if I have one to test).
What also could be is that I have cleaned up the device tree files and brought them into a better logical structure. This man have introduced a regression for your combination.
I have to check when I find time to look into that.
I've also tried the 5.4.11 kernel, that boots , but /sys/devices/platform/txs02612/switch doesn't exist there, so it can't find the emmc.
(and the display turns off after a second or so, but i think that was a known problem for the 5.x kernels)
Yes, that is a known bug. It tries to rotate the framebuffer by 90 degrees but that ends in no rotation and turning off backlight. Something in the fbdev helpers.
Am 19.01.2020 um 11:58 schrieb H. Nikolaus Schaller hns@goldelico.com:
I've also tried the 5.4.11 kernel, that boots , but /sys/devices/platform/txs02612/switch doesn't exist there, so it can't find the emmc.
It is also missing in v5.5-rc
Seems to be that the driver is not bound, although configured into the kernel.
I'll look into that.
Here is a patch to bring it back (until we have an automatic switch driver - then we can revert it):
commit 31bc826997490eb4e5016b57b4c38c32825259fa (HEAD -> work-txs02612) Author: H. Nikolaus Schaller hns@goldelico.com Date: Sun Jan 19 20:55:10 2020 +0100
ARM: DTS: LC15: bring back /sys/devices/platform/txs02612/switch
until we have an automatic switch driver
Signed-off-by: H. Nikolaus Schaller hns@goldelico.com
diff --git a/arch/arm/boot/dts/omap5-letux-cortex15-common.dtsi b/arch/arm/boot/dts/omap5-letux-cortex15-common.dtsi index 7451121d7d93..d59d9fd4c8af 100644 --- a/arch/arm/boot/dts/omap5-letux-cortex15-common.dtsi +++ b/arch/arm/boot/dts/omap5-letux-cortex15-common.dtsi @@ -263,6 +263,14 @@ }; };
+/* keep this as long as the txs02612 driver does not exist/work as mmc slave driver */ +/ { /* eMMC / uSD-card */ + txs02612 { + compatible = "ti,txs02612"; + select-gpio = <&gpio3 18 0>; /* gpio3_82 soft-select */ + }; +}; + &mmc3 { /* WiFi-SDIO */ /* has been predefined for a wl1271 in ompa5-common.dtsi */ };
-
aTc -
H. Nikolaus Schaller