Posts by mglae

Quote from dan.h

do i need to do anything else before I'll test the restore?

No.

It is possible that Kodi is crashing while rendering the the Hebrew characters. I've uploaded service.libreelec.settings-10.0-20210914.zip to Testing LibreELEC-settings addon · Issue #4547 · LibreELEC/LibreELEC.tv · GitHub only showing ASCII characters. Please test.

The Kodi log file was not included because your RPI did not update the system time via NTP. Therefore the log file was directly deleted because seen as oldest one.

There is an interrupt that cannot be handled:

[    2.944255] hid-generic 0003:05AC:820B.0004: input,hidraw2: USB HID v1.11 Mouse [HID 05ac:820b] on usb-0000:00:06.0-1.3/input0
[    2.997830] irq 20: nobody cared (try booting with the "irqpoll" option)
[    2.997834] CPU: 1 PID: 234 Comm: systemd-journal Not tainted 5.1.16 #1
[    2.997835] Hardware name: Apple Inc. Macmini3,1/Mac-F22C86C8, BIOS     MM31.88Z.00AD.B00.0907171535 07/17/09
[    2.997836] Call Trace:
[    2.997839]  <IRQ>
[    2.997846]  dump_stack+0x70/0x9a
[    2.997850]  __report_bad_irq+0x3c/0xb6
[    2.997852]  note_interrupt.cold.9+0xb/0x5d
[    2.997854]  handle_irq_event_percpu+0x6f/0x80
[    2.997856]  handle_irq_event+0x3b/0x5a
[    2.997858]  handle_fasteoi_irq+0x90/0x130
[    2.997861]  handle_irq+0x20/0x30
[    2.997864]  do_IRQ+0x46/0xd0
[    2.997865]  common_interrupt+0xf/0xf
[    2.997866]  </IRQ>
[    2.997869] RIP: 0010:copy_user_generic_string+0x2c/0x40
[    2.997872] Code: 08 72 27 89 f9 83 e1 07 74 15 83 e9 08 f7 d9 29 ca 8a 06 88 07 48 ff c6 48 ff c7 ff c9 75 f2 89 d1 c1 e9 03 83 e2 07 f3 48 a5 <89> d1 f3 a4 31 c0 c3 66 66 2e 0f 1f 84 00 00 00 00 00 66 90 83 fa
[    2.997873] RSP: 0018:ffff9aebeec33ec8 EFLAGS: 00000246 ORIG_RAX: ffffffffffffffdb
[    2.997875] RAX: 00007ffd245a6728 RBX: 00007ffd245a6718 RCX: 0000000000000000
[    2.997876] RDX: 0000000000000000 RSI: ffff9aebeec33ef0 RDI: 00007ffd245a6728
[    2.997878] RBP: ffff9aebeec33ed0 R08: 00058fca8155ec8a R09: 0000000000000000
[    2.997879] R10: 0000000000000000 R11: 0000000000000000 R12: ffff9aebeec33f58
[    2.997880] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
[    2.997884]  ? ktime_get_ts64+0x46/0xe0
[    2.997887]  ? _copy_to_user+0x2b/0x40
[    2.997889]  put_timespec64+0x3f/0x60
[    2.997892]  __x64_sys_clock_gettime+0x85/0xc0
[    2.997894]  do_syscall_64+0x55/0x110
[    2.997896]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
[    2.997898] RIP: 0033:0x7ffd245bc613
[    2.997900] Code: 01 c8 53 45 8b 11 41 f6 c2 01 0f 85 01 01 00 00 8b 05 85 ca ff ff 83 f8 01 74 11 83 f8 02 74 7a 5b b8 e4 00 00 00 41 5c 0f 05 <5d> c3 0f ae e8 0f 31 48 c1 e2 20 48 09 c2 49 8b 40 08 48 8b 0d 5c
[    2.997901] RSP: 002b:00007ffd245a66e0 EFLAGS: 00000293 ORIG_RAX: 00000000000000e4
[    2.997903] RAX: ffffffffffffffda RBX: 0000000000000001 RCX: 00007ffd245bc613
[    2.997904] RDX: 0000000000006f6e RSI: 00007ffd245a6718 RDI: 0000000000000001
[    2.997906] RBP: 00007ffd245a66e0 R08: 00007ffd245b90b0 R09: 00007ffd245b9080
[    2.997907] R10: 0000000000000632 R11: 0000000000000293 R12: 00007ffd245a67a8
[    2.997908] R13: 00007ffd245a9168 R14: 0000000000000000 R15: 000000000000007f
[    2.997910] handlers:
[    2.997913] [<000000002493f4a9>] usb_hcd_irq
[    2.997915] Disabling IRQ #20

IRQ 20 is used for PCI Interrupt Link [LUS0] enabled at IRQ 20 (IMO the USB) and PCI Interrupt Link [Z00F] enabled at IRQ 20 (no idea what this is).

Likely a driver for the Z00F device is missing.

Try adding the mentioned irqpoll kernel parameter to the APPEND line in /flash/syslinux.cfg (after mount -o remount,rw /flash).

Last directory acceessed is logged with /storage/.kodi/userdata/peripheral_data. Please post the output of:

ls -la /storage/.kodi/userdata/peripheral_data

And while you are on it for the log naming issue:

ls -la /storage/.kodi/temp

Feb 02 17:30:58.586444 LibreELEC-PI4 kodi.sh[988]: Crash report available at /storage/.kodi/temp/kodi_crashlog_20210202173055.log

Surprisingly crash log is generated but not included in the paste.

Please do

paste /storage/.kodi/temp/kodi_crashlog_20210202173055.log

and post the URL.

Do you really expect that I'm able to theoretically choose the the correct combination of uid, gid, forceuid and forcegid parameters needed in your configuration?

LE is useing the quiet kernel parameter by default. If you did not have removed it before to see the messages, systemd was starting logging because of a seen error.

Unfortunately there is no useful information in the old log.

Ssh into your device and do

echo DEBUG=1 >>/storage/.config/kodi.conf

Enable debugging in Settings>System Settings>Logging of kodi GUI, reboot, reproduce the crash and upload the crash log again.

Read "FILE AND DIRECTORY OWNERSHIP AND PERMISSIONS" in man mount.cifs

1. The kernel file can not be opened. But the legacy image was only created to in case the usual GPT disk is not possible.

2. The size of the used stick is unknown so we start with a small 32MiB file system that is expanded at first boot.

The filesystem repair may have been started because you pulled the stick before any data was flushed.

A. Don't be afraid of the installer, you have to select the installation, choose the disk and then being asked twice if you really like to erase the HD.

B. We do use a GPT image. The changes will be used if accepted by the team.

Usually a log is generated when kodi crashed and is restarted. Use pastecrash or the crash log uploader of LibreELEC-Settings.

1. Yes, write 100 times: there is no /etc/fstab used in LE.

With UUIDs this should work. Only with a second copy attached the same time you don't know which is mounted.

2. A long we don't know why the LE image is not accepted this cannot be corrected. Use what is working.

I've created two additional test images, please test on a spare stick if you see the syslinux boot messages.

a) LibreELEC-Generic.x86_64-legacy.empty.img.gz.zip

b) LibreELEC-Generic.x86_64-gpt.empty.img.gz.zip

1.) The randomly generated UUIDs can be read from any linux system and then stored in the syslinux.cfg. LE is not using /etc/fstab.

boot=/dev/sdb1 disk=/dev/sdb2 is possible but not reliable because it depend on the disk initialization order. Another possibility is to use boot=LABEL=xxx disk=LABEL=yyy if the file system labels are defined.

The UUID was not incorrect, it still pointed to the original file system of the copied syslinux.cfg

2) No. ISO is read only.

Boot problems are very rare, although reported from time to time.

Great.

You can create an universal service:

[Unit]
Description=Wait for SMB server
After=time-sync.target

[Service]
Type=oneshot
ExecStart=/bin/sh -c "until nc -z %I 139;do sleep 1;done;sleep 5"
RemainAfterExit=yes

To use it in the mount service:

[email protected]
[email protected]

Quote from shippy

2. Does this mean I will have to do manual config every time there is a new install?

It is the nature of UUIDs to be unique. So yes, with any new created file system.

Quote from shippy

3. On my current config ( Syslinux 6.0.4 boot showing fat32 in my 14.9 GB USB space, and copy over Kernel and System files), how do I add ext4 partition to enable run mode?

Shrink the fat32 partition and create an additional ext4 one. (Don't forget to replace the disk=UUID= in syslinux.cfg)

I'm not aware of any Windows tool capable of doing this. Use e.g. gparted from any Linux distribution or Gparted Live.

Now it's failing with timeout. Try adding ;sleep 5 after the done.

Now the error message is different: "Could not mount UUID=1706..."

syslinux.cfg must have matching UUIDs. Replace 1706-4700 with 0E96-8A79.

But then you can only boot installer and live. For run the second ext4 partition (sdb2) is required.

It is much better: waiting 38 sec until the server is up and failing with "Connection refused"

You have to wait until the CIFS server is up, not only the device.:

ExecStart=/bin/sh -c "until nc -z 192.168.50.12 139;do sleep 1;done"