Open tomharvey opened 2 months ago
tomharvey
commented
2 months ago After a reboot I ended up in busy box.
After running fsck -y on the device I can exit out of busy box and all is well again... I assume, until the next time the queue gets borked
HeyMeco
commented
1 month ago After a reboot I ended up in busy box.
After running fsck -y on the device I can exit out of busy box and all is well again... I assume, until the next time the queue gets borked
Had the same thing happen. The above fix also worked on mine
Joshua-Riek
commented
1 month ago Interesting, dropping the eMMC down to HS200 is the proposed change here?
While this is not a fix, dropping https://github.com/Joshua-Riek/linux-rockchip/commit/d03a5046f13a8a5b4d7c4b02e442aa4508890413 may be a better option, but how will this affect other devices?
HeyMeco
commented
1 month ago @Joshua-Riek
how will this affect other devices?
Should only reduce IOPS a bit but better to have a stable device
nyanmisaka
commented
1 month ago As an alternative, add these only to known affected devices?
&sdhci {
...
/delete-property/ supports-cqe;
...
};I think I need to try a dtbo like this on Rock5B:
/dts-v1/;
/plugin/;
/ {
fragment@0 {
target = <&sdhci>;
__overlay__ {
mmc@fe2e0000 {
/* Commenting out supports-cqe */
// supports-cqe;
};
};
};
};
What happened?
I was doing some intensive disk activity on my EMMC device (no other disks are connected) when it flipped to read-only mode.
Kernel logging is attached
This seems similar to issues which resulted in a reconfiguration of the EMMC in the official kernel here https://github.com/radxa/kernel/pull/127/files and issues reported in Armbian here https://forum.armbian.com/topic/18745-emmc-errors/
Kernel version
6.1.0-1025-rockchip
SBC model
Radxa Rock 5B
What operating system are you seeing this problem on?
Ubuntu 24.04 LTS (Noble Nombat)
Relevant logs