8k bcache NULL pointer deref on mount / can't boot

[Robert Collins]

After creating a bcache cache device with an 8k block size (see #1662407), a kernel oops happens every time device is probed.

I'm not sure that there is any userspace bug present in #1662407 at all - it may just be kernel. I've attached a dmesg there, but can't actually get into the system at the moment to get apport-collect or other diagnostics.

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1809748

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Robert Collins (lifeless)]

cache superblock info - sb.magic ok
sb.first_sector 8 [match]
sb.csum 5A3217E583DB5B86 [match]
sb.version 0 [cache device]

dev.label (empty)
dev.uuid 2a9e0dfd-acd0-4750-85ae-b347d7580e2b
dev.sectors_per_block 16
dev.sectors_per_bucket 1024
dev.cache.first_sector 1024
dev.cache.cache_sectors 209714176
dev.cache.total_sectors 209715200
dev.cache.ordered no
dev.cache.discard no
dev.cache.pos 0
dev.cache.replacement 0 [lru]

cset.uuid 24283111-591b-4fb8-9428-8f5db4cb5a87

[Robert Collins (lifeless)]

after doing a wipefs on the cache drive the BUG still occured; wipefs's the backing drive corrected that. I haven't yet tried wiping just the backing drive.

[Robert Collins (lifeless)]

hint if testing this - add module_blacklist=bcache to your kernel boot line to boot and reset the metadata :)

[Christian Ehrhardt  (paelzer)]

It is yet unclear if there is an issue in the userspace tools, hence I made the sibling bug a dup and added the tools as a task here marked as incomplete for now.

As I mentioned providing the extra kernel data might help to get the triaging by the kernel team started.