The 2.5.39 floppy driver is still broken. Jens' fix to ll_rw_block
was included in 2.5.39, so the kernel doesn't reboot at the first
I/O operation as it did in 2.5.38, but several problems remain:
1. Accessing /dev/fd0H1440 oopses fs/block_dev.c:do_open().
Fixed by applying Al Viro's O100-get_gendisk-C38 patch.
2. The capacity of floppies is halved.
Fixed by applying Al Viro's O101-floppy_sizes-C38 patch.
3. /dev/fd0 size autodetection doesn't work properly. The very
first read or write only transmits 2K (or 4K with Al's patches)
of data. I/O works after a reopen of /dev/fd0.
Explanation: the floppy interrupt handler updates floppy_sizes[],
but doesn't set_capacity() on the corresponding gendisk, causing
the detected size for /dev/fd0 to not be applied until the next
time /dev/fd0 is opened.
Quick fix: add the missing set_capacity() calls.
With Al's two patches and this one floppies work reliably for me.
/Mikael
--- linux-2.5.39/drivers/block/floppy.c.~1~ Sat Sep 28 12:42:00 2002
+++ linux-2.5.39/drivers/block/floppy.c Sat Sep 28 18:49:27 2002
@@ -778,6 +778,7 @@
"disk change\n");
current_type[drive] = NULL;
floppy_sizes[TOMINOR(drive)] = MAX_DISK_SIZE << 1;
+ set_capacity(&disks[drive], floppy_sizes[TOMINOR(drive)]);
}
/*USETF(FD_DISK_NEWCHANGE);*/
@@ -2426,6 +2427,7 @@
}
current_type[current_drive] = _floppy;
floppy_sizes[TOMINOR(current_drive) ]= _floppy->size+1;
+ set_capacity(&disks[current_drive], floppy_sizes[TOMINOR(current_drive)]);
break;
}
@@ -2435,6 +2437,7 @@
_floppy->name,current_drive);
current_type[current_drive] = _floppy;
floppy_sizes[TOMINOR(current_drive)] = _floppy->size+1;
+ set_capacity(&disks[current_drive], floppy_sizes[TOMINOR(current_drive)]);
probing = 0;
}
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Sep 30 2002 - 22:00:38 EST