Matt Madia wrote: > On Mon, May 30, 2011 at 12:27, Ingo Weinhold <ingo_weinhold@xxxxxx> wrote: > > Matt Madia wrote: > >> If it helps, here's the complete syslog ; http://pastebin.com/bjVXwnfn > >> > >> Is it worth moving this to Trac #4502 or another ticket? > > > > I don't see how this is related to #4502 at least. > > Should it be a new ticket for IOCache, then? Please open a ticket that an anyboot image with block size 1024 doesn't work on your hardware. According to your syslog the block size is correctly determined as 2048, which seems to point to an IOCache or DMAResource issue. I tried to emulate the problem in qemu (also introducing dummy errors), but failed to cause any unaligned I/O operations to get to the driver's do_io() callback, which suggests that IOCache/DMAResource are doing their job just fine. You could also try to disable use of the IOCache (use IOSchedulerSimple instead) [1] and see, if that changes anything. CU, Ingo [1] http://dev.haiku-os.org/browser/haiku/trunk/src/add-ons/kernel/drivers/disk/scsi/scsi_cd/scsi_cd.cpp#L1000