antwain.schneider added an attachment on Kernel.org Bugzilla:
Created attachment 305795
a program written in c that can be compiled and ran on a rk3328-based device to illustrate the problem which is bad
this program dumps the bits set in the grf for 2bl and 2bh
in comment 7, the otp-out thing is still vital to prove this point without extensive dtb modification
i've got it down to this
fdtput -t x /boot/efi/rockchip/rk3328-roc-cc.dtb /pinctrl/tsadc/otp-out rockchip,pins 02 0d 02 5a
what's being changed in this line is the mode (from 1 to 2) of gpio2b5
otp-out in the rk3328.dtsi is
rockchip,pins = <2 RK_PB5 1 &pcfg_pull_none>;
which when dumped by dtc is
rockchip,pins = <0x02 0x0d 0x01 0x5a>;
0x5a is the phandle for pcfg_pull_none for me currently, your value may vary
notice that it's using gpio2b5 which exists but doesn't have a listing in chapter 3 of the trm but is described elsewhere in the document
1 run the program and save output
2 fdtput the dtb and reboot
3 run the program again and compare outputs, and notice how the completely wrong register changed
File: poc.c (text/x-csrc)
Size: 1.53 KiB
Link: https://bugzilla.kernel.org/attachment.cgi?id=305795
---
a program written in c that can be compiled and ran on a rk3328-based device to illustrate the problem which is bad
You can reply to this message to join the discussion.
--
Deet-doot-dot, I am a bot.
Kernel.org Bugzilla (peebz 0.1)
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel