Hi Florian. On Wed, 2025-10-29 at 16:54 +0100, Florian Fuchs wrote: > On 25 Oct 12:35, Gregor Riepl wrote: > > > We fixed that in T2 Linux (debugged that for a longer time .. ), the > > > reason was > > > that the the probe __floppy_read_block_0 tries to read 1 page from disk, > > > 8KB, > > > which was more than the derived floppy size of 8 sectors (4KB) from the > > > minor > > > number, so the "beyond end of device" message comes up, see: > > > https://github.com/rxrbln/t2sde/blob/main/package/kernel/linux/hotfix-floppy-not-4k-pages.patch > > > > Is this common for PC floppy drives? > > Perhaps /dev/fd0 should register itself with the correct minor number > > instead. > > Yes, it's common, as far as I understand the floppy driver, it > auto-detects the disk geometry by reading the first block ("bootblock") > in floppy_revalidate and __floppy_read_block_0 [1]. > > The auto-detection is also noted in admin-guide/devices.txt [2]: > > 2 block Floppy disks > 0 = /dev/fd0 Controller 0, drive 0, autodetect > > > Perhaps /dev/fd0 should register itself with the correct minor number > > instead. > > That might be difficult, as you could put in disks with a different > kind of geometry.
With your patch independently verified, could you send your patch upstream? Gregor can then add his "Tested-by:" on the patch if you remember to CC him ;-). We're tracking the state of this bug here [1] so it doesn't get lost. Adrian > [1] https://github.com/sparclinux/issues/issues/18 -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer `. `' Physicist `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
