Hi, [email protected] wrote: > > Fixed now. https://en.wikipedia.org/wiki/Optical_disk_image redirects > > to https://en.wikipedia.org/wiki/Optical_disc_image.
David Wright wrote: > Bravo. +1 Apologies for not understanding the findings of David Wright on the first hand. (The abundance of human error is probably the main reason why AI looks somewhat capable despite its stupid halluzinations.) David Wright wrote: > It had occurred to me that the author(s) might be trying to emphasise > that sectors were going to be read off a magnetic disK and written > onto an optical disC. > > However, I'm not convinced that you can do that, because the disK > with the image on it could have 512-byte sectors, yet the optical > disC sectors are more likely to be twice or four times that. It's not so much about the block size of the storage device from where the image stems but about the block size which the filesystem in the image uses to convert byte addresses to block addresses for e.g. the SCSI READ command. In my experiment with ext4 and DVD+RW it looks like Linux expects the block size of the device to match the block size of the ext4 filesystem. I did not yet test whether options -b 2048 of mkfs.ext4 and losetup would make an ext4 which is readable from DVD without a loop device as block size translator. Different from ext4, the ISO 9660 filesystem driver of Linux works with at least 512, 1024, 2048, and 4192 bytes per block on the storage device. (One could dig for the code which probably translates the 2048-byte block addresses of ISO to byte offsets and then to block addresses of the device's block size.) Experiments years ago showed that Linux only accepts ISO 9660 with block size 2048 although the specs of ISO 9660 / EXMA-119 allow other sizes. Have a nice day :) Thomas
