On Thu, Oct 05, 2023 at 06:18:08AM +0200, Jan Kiszka wrote:
On 04.10.23 14:15, Nishanth Menon wrote:
On 22:26-20231003, Jan Kiszka wrote:
From: Jan Kiszka jan.kiszka@siemens.com
Since commit [1] A53 u-boot proper is broken. This is because nodes marked as 'bootph-pre-ram' are not available at u-boot proper before relocation.
To fix this we mark all nodes as 'bootph-all'.
[1] 9e644284ab812 ("dm: core: Report bootph-pre-ram/sram node as pre-reloc after relocation")
Signed-off-by: Jan Kiszka jan.kiszka@siemens.com
This may overshoot, but at least the board boots again. Could it be that [1] broke even more boards?
Jan: https://lore.kernel.org/all/b1c62a7d-a90e-4212-8972-9b622e147975@kernel.org/
I got boot without r5-beagleplay.dts modified. and it is in line with the changes in linux-next commit 944adefc7f88 ("arm64: dts: ti: k3-am625-beagleplay: Add boot phase tags marking")
Yeah, no problem, missed that.
Meanwhile, I can fix our IOT2050 because I was unfortunatenly right: more havoc in sight. Did anyone tried to look at the fallouts systematically already? Is it only affecting the TI family?
Well, I'm pretty confused right now. The visible breakage has been traced back to a commit that was in -next and is fine on my J721E EVM and is fine on my AM65x EVM. I can't figure out where my Beagleplay ended up, so I can't check that one as easily. But given how the breakage is described, mine too should be failing. But they aren't. In both cases, I have the GP versions of the chips, and am booting the unsigned files.