Bug 27999

Created attachment 12170 [details]
lspcidrake -v

Created attachment 12171 [details]
journal of failed boot

I tried to change to SDDM via tty but reboot was unsuccessful.

From attachment 12171 [details] (with localhost removed from all lines)
Jan 02 19:02:49 kernel: Kernel command line: BOOT_IMAGE=/boot/vmlinuz-5.10.4-desktop-2.mga8 root=UUID=34203a1b-dc51-47f5-9fd5-63ae6c2c5f59 ro noiswmd resume=/dev/mmcblk0p3 audit=0 vga=791
Jan 02 19:02:49 kernel:  mmcblk1: p1 p2 p3
Jan 02 19:02:49 kernel: EXT4-fs (mmcblk1p2): INFO: recovery required on readonly filesystem
Jan 02 19:02:49 kernel: EXT4-fs (mmcblk1p2): write access will be enabled during recovery
Jan 02 19:02:49 kernel: EXT4-fs (mmcblk1p2): recovery complete
Jan 02 19:02:49 kernel: EXT4-fs (mmcblk1p2): mounted filesystem with ordered data mode. Opts: (null)
Jan 02 19:02:49 kernel: EXT4-fs (mmcblk1p2): mounted filesystem with ordered data mode. Opts: (null)
Jan 02 19:02:49 dracut: Mounted root filesystem /dev/mmcblk1p2
Jan 02 19:02:49 kernel: EXT4-fs (mmcblk1p2): re-mounted. Opts: acl
Jan 02 19:02:49 systemd-fsck[461]: /dev/mmcblk1p2: clean, 225350/798112 files, 1777494/3187376 blocks
Jan 02 19:04:19 systemd[1]: dev-mmcblk0p1.device: Job dev-mmcblk0p1.device/start timed out.
Jan 02 19:04:19 systemd[1]: Timed out waiting for device /dev/mmcblk0p1.
Jan 02 19:04:19 systemd[1]: dev-mmcblk0p1.device: Job dev-mmcblk0p1.device/start failed with result 'timeout'.
Jan 02 19:04:19 systemd[1]: dev-mmcblk0p3.device: Job dev-mmcblk0p3.device/start timed out.
Jan 02 19:04:19 systemd[1]: Timed out waiting for device /dev/mmcblk0p3.
Jan 02 19:04:19 systemd[1]: Dependency failed for /dev/mmcblk0p3.
Jan 02 19:04:19 systemd[1]: dev-mmcblk0p3.swap: Job dev-mmcblk0p3.swap/start failed with result 'dependency'.
Jan 02 19:04:19 systemd[1]: dev-mmcblk0p3.device: Job dev-mmcblk0p3.device/start failed with result 'timeout'.

So the question is why the device mmcblk0 failed start.

Please try to recreate the error.

Can you provide log from a working boot with 5.9 series ?

I see in this one we fail to enumerate one gpio:

sdhci-acpi 80860F16:00: failed to setup card detect gpio

Created attachment 12174 [details]
good 5.9.12-desktop-1.mga8  journal

full journal of a 5.9.12-desktop-1.mga8 session on this hardware (Acer C730 ex-chromebook with Seabios)

Created attachment 12175 [details]
failed 5.10.4-desktop-2.mga8 boot journal

next boot that failed

Mga8 Gnome 32 installed on Asus T100TA-DK003H does NOT exhibit issue.

successful multiple bootups and reboots ok

ok, so with 5.9 series kernels it inits as primary storage:

mmc0: new HS200 MMC card at address 0001
mmcblk0: p1 p2 p3

and with 5.10 series kernels it gets enumerated as second storage:

mmc1: new HS200 MMC card at address 0001
mmcblk1: p1 p2 p3


Of course kernel does not guarantee device detetection ordering wich is why we prefer uuid or label mounts...

So if you alter the resume=/dev/mmcblk0p3 to resume=/dev/mmcblk1p3 it would probably boot up on 5.10 series 

or boot the 5.9 series, change resume and swap to use uuid reference, recreate initrds and so on...

uname -r
5.9.12-desktop-1.mga8

lsblk
NAME         MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
mmcblk0      179:0    0 14.7G  0 disk 
├─mmcblk0p1  179:1    0  299M  0 part /boot/EFI
├─mmcblk0p2  179:2    0 12.2G  0 part /
└─mmcblk0p3  179:3    0  2.2G  0 part [SWAP]
mmcblk0boot0 179:8    0    4M  1 disk 
mmcblk0boot1 179:16   0    4M  1 disk 


not sure where mccblk1..... comes from, as no external storage connected.

altering resume=/dev/mmcblk0p3 to resume=/dev/mmcblk1p3 did not seem to make a difference

Created attachment 12177 [details]
grub configuration

appears to be set to look for UUID

(In reply to ben mcmonagle from comment #10)
> uname -r
> 5.9.12-desktop-1.mga8
> 
> lsblk
> NAME         MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
> mmcblk0      179:0    0 14.7G  0 disk 
> ├─mmcblk0p1  179:1    0  299M  0 part /boot/EFI
> ├─mmcblk0p2  179:2    0 12.2G  0 part /
> └─mmcblk0p3  179:3    0  2.2G  0 part [SWAP]
> mmcblk0boot0 179:8    0    4M  1 disk 
> mmcblk0boot1 179:16   0    4M  1 disk 
> 
> 
> not sure where mccblk1..... comes from, as no external storage connected.
> 

either you have an card reader or the hardware has 2 ports and the system only uses one...

from kernel 5.9 log:
mmc0: SDHCI controller on ACPI [80860F14:00] using ADMA
sdhci-acpi 80860F16:00: failed to setup card detect gpio
mmc1: SDHCI controller on ACPI [80860F16:00] using ADMA
mmc0: new HS200 MMC card at address 0001
mmcblk0: p1 p2 p3


and in 5.10 log it gets swapped:
sdhci-acpi 80860F16:00: failed to setup card detect gpio
mmc0: SDHCI controller on ACPI [80860F16:00] using ADMA
mmc1: SDHCI controller on ACPI [80860F14:00] using ADMA
mmc1: new HS200 MMC card at address 0001
mmcblk1: p1 p2 p3


> altering resume=/dev/mmcblk0p3 to resume=/dev/mmcblk1p3 did not seem to make
> a difference

I forgot there is also the contents in:
/etc/dracut.conf.d/51-mageia-resume.conf


and since this is in initrd the 5.10 initrd need to be rebuilt after changing this file
yrs, but the resume line points to /dev stuff...

But I also

uname -r
5.10.4-desktop-3.mga8

failed today on Asus T00TA.

was able to boot usning the advanced boot option- recovery

Created attachment 12178 [details]
unsuccessful 5.10.4 boot

Created attachment 12179 [details]
successful 5.10.4 boot via advanced boot recovery option

One workaround that should work would be to add the nofail option to the swap
line in /etc/fstab, and then run dracut -f.

When the swap does time out though, the system would be running without it,
but it might make it easier to debug.

ok on MGA8 RC, so closing as fixed

Resolution: (none) => FIXED
Status: NEW => RESOLVED