Hi, Leon Merten Lohse wrote: > Although it is allowed, it might give some drives a hiccup.
It's not about the drives but about the low-level drivers in the operating systems. Those which act underneath SCSI passthrough like Linux ioclt(SG_IO). In my case i lost the connection to a USB connected DVD burner when i requested more bytes than available. The request size is also an announcement to the kernel and its drivers. Some of them took offense. I do not quite remember what was the problem with cdrecord ... Possibly this one: https://lists.debian.org/cdwrite/2007/05/msg00042.html Effects look more harmless than mine. "Input/output error. get performance: scsi sendcmd: no error" (The pointer given to a FreeBSD problem is probably a red herring. Both problems look unrelated.) I do not know from where Andy Polyakov (growisofs) got the idea or why he deemed it necessary. But i remember that i first riddled why he does it and then had the incident which made me smarter. > Also the majority of the code in mmc.c in libcdio just uses one query. That's a valid point on the one hand. But on the other hand the double call is best practice and if it ever shall become part of libcdio one has to start at some point. So why not now ? Have a nice day :) Thomas
