Unable to USB boot Thinkpad T480 with nvme disk and Windows 10 installed

[austrosepp]

Hi there,

i try to install hrev-55970 64-bit on TP 480 from USB stick with no success. BIOS: secure boot: off, UEFI boot: first UEFI then legacy. the machine has windows 10 installed, with some free space on the harddisk. the stick boots fine on other thinkpad (t450).

the haiku logo with the 7 icons appears, but none of the icons get highlighted.

next try: enabled debug on screen option in boot menu: same story, haiku logo with 7 icons, no debug output.

next try: checked the last boot log via boot menu: lot of entries …, but no apparent error for me. where does he write this syslog to? is it on the stick, that i can access it on another machine, and upload it here?

I already checked the content of the stick on another haiku machine, but unfortunately could not find a syslog file on the stick.

here a linux lsblk -f output:

NAME FSTYPE FSVER LABEL UUID
nvme0n1 zfs_member 5000 lxd-pool 2735629842434347490
├─nvme0n1p1 ntfs Wiederherstellung EE50606850603A0B
├─nvme0n1p2 vfat FAT32 0460-FAE1
├─nvme0n1p3
├─nvme0n1p4 ntfs 5ADCA586DCA55D4F
└─nvme0n1p5 ntfs 5ED00298D0027715

here the blkid:

/dev/nvme0n1p1: LABEL=“Wiederherstellung” BLOCK_SIZE=“512” UUID=“EE50606850603A0B” TYPE=“ntfs” PARTLABEL=“Basic data partition” PARTUUID=“22255200-e036-4b89-81f1-e8dd6536df2e”
/dev/nvme0n1p2: UUID=“0460-FAE1” BLOCK_SIZE=“512” TYPE=“vfat” PARTLABEL=“EFI system partition” PARTUUID=“a5741332-a715-4db1-baaa-8d5596aeeb0b”
/dev/nvme0n1p4: BLOCK_SIZE=“512” UUID=“5ADCA586DCA55D4F” TYPE=“ntfs” PARTLABEL=“Basic data partition” PARTUUID=“a8123865-521c-416c-8b49-18799ecb7390”
/dev/nvme0n1p5: BLOCK_SIZE=“512” UUID=“5ED00298D0027715” TYPE=“ntfs” PARTUUID=“e4972297-8ddd-45d3-8d38-c22a33855252”
/dev/sda1: BLOCK_SIZE=“2048” UUID=“2020-08-08-02-08-38-00” LABEL=“Finnix 121” TYPE=“iso9660” PARTUUID=“7e9c6d49-01”
/dev/loop0: TYPE=“squashfs”
/dev/sda2: SEC_TYPE=“msdos” UUID=“EFA2-41DD” BLOCK_SIZE=“512” TYPE=“vfat” PARTUUID=“7e9c6d49-02”
/dev/zram0: UUID=“5a8296f7-de1f-4120-b5f8-2729b5e3c095” TYPE=“swap”
/dev/nvme0n1p3: PARTLABEL=“Microsoft reserved partition” PARTUUID=“01ad49d7-77f0-4e37-8f3a-b69a9a55b428”

[waddlesplash]

The "write syslog" option is currently only available in the bootloader on BIOS, so that indicates you are not using the EFI loader. Could you somehow try to force EFI and see if that works?

[austrosepp]

I attached the last two pages of the syslog as .jpg files

@waddlesplash: will try EFI only.

[austrosepp]

I am already booting with: UEFI only, CSM Support No

[waddlesplash]

The log is just the bootloader log, it doesn't display anything from e.g. the previous boot, unfortunately. You can try the "on-screen debug" option...

[austrosepp]

I have tried the "on-screen debug" option, but there is no debug output, it boots to the haiku logo with the 7 icons, none of them highlights, and get stuck here.

as to the bootloader log: the "Display current boot loader log" in the boot menu shows about 10 screens of logging text, from which I attached the last 2 in this ticket.

may I draw your attention to the last line of the 2nd photo, it goes like ... failed to open administrative directory ...

[waddlesplash]

That error is harmless, it always occurs on systems that have never been fully booted to the desktop at least once.

[waddlesplash]

No icons lit generally means bootloader/kernel handoff failed. No onscreen debug output would indicate the same. Any chance this machine has a serial port of some kind?

[waddlesplash]

Can you try older versions (e.g. beta3) and see if EFI works then?

[austrosepp]

@ waddlesplash

nope, no serial port, just USB. will try beta3 and maybe an older hrev tomorrow.

[austrosepp]

beta3 boot: same behaviour as previous hrev-55970. hrev-54253 boot: boots directly into boot menue (without me pressing space key)

i attach the last 2 pages from each bootmanager log

[waddlesplash]

Unfortunately if no icons are lighting up, not even the first one, the problem may be basically undiagnosable without a serial port or equivalent (there might have been experimental patches to broadcast the log over the network?)