https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=287928
--- Comment #3 from Warner Losh <i...@freebsd.org> --- So the CACHE PAGE TOO SHORT shouldn't be printed except under boot verbose. what it means is that we asked for the caching mode page, and it was shorter than we expected. We concluded from that we should not send SYNCHRONIZE CACHE commands to the device. That's it. That's all that message means. It's a common problem that I saw in my testing: the device would return the mode page, but it would be 0 length or would just be a copy of the CDB we sent down! For the 80-odd different thumb drives, disks, etc I cycled through, it worked on all the others, so I'm puzzled here. I don't know why it's disconnecting. You can create a Quirk for this device which disables the SYNCHRONIZED CACHE feature and see if that helps. That will short-circuit the test entirely, and also disable SYNCHRONIZED CACHE. If it works manually adding a SYNC CACHE quirk, then we can add that to the kernel. As a sanity check, does 14.3 work or fail? I agree that Mark's problem is completely different. -- You are receiving this mail because: You are the assignee for the bug.
