© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
f33e652
(Get the code!)
I will try to debug as time permits, but the priority of MS-DOS bugs is not ... measurable with casual tools. However, there are a lot of other IDE bugs on my plate that are very important! so I am hoping to grab a bunch of IDE bugs at once, but no promises here.
Notably, our geometry detection is not very good, it's more than possible we are misreporting values and confusing DOS. Our IDE disks are also not very consistent about what standard of the spec they are trying to emulate, so there are likely other problems there, too.
If you'd like to debug on your own, I'd recommend enabling tracing and enabling some of the IDE trace points; some of them can be quite verbose -- don't enable the data dumping ones. The control flow ones can be informational sometimes to guess when the guest OS got confused and then walk your way back to a register read that would have picked up some error bits, or to detect busy-waits on registers not changing and try to guess what it was waiting for.
https:/
https:/
Ignore the AHCI and ATAPI traces, and don't use the ide_data_* traces unless you are booting a custom firmware that only performs a strict few IO accesses -- otherwise you'll get flooded off the map.