In <Pine.LNX.4.10.10006191230440.28028-100000@mhw.ULib.IUPUI.Edu> Mark H. Wood (mwood@IUPUI.Edu) wrote:
> On Mon, 19 Jun 2000, Chris Swiedler wrote:
>> Richard Stallman <rms@gnu.org> wrote in message
>> news:<fa.h09h36v.1a1sro5@ifi.uio.no>...
>>
>> > Is there any possibility of making Linux handle file systems on
>> > floppies like MSDOS, so that there is no need to explicitly mount and
>> > unmount a floppy drive in order to access floppies through the file
>> > system?
>>
>> In discussing this problem, several people had possible solutions which were
>> all shot down with a similar counterargument: the kernel can't assume that
>> the user who is using the floppy drive is at the console. Apparently,
> Actually there are two issues here. One is: how to correctly route the
> information that the diskette has been removed. This problem exists for
> every removable medium, so we can either ignore it as has been done in
> general, or develop a general solution (which might be nice to have). I
> suspect that any such solution is going to be incompatible with the
> present request, since there's no way for the OS to know where the user
> has gone after sticking the floppy in the drive, unless the user tells it.
And in case of Windows there are easy solution: ask user on console.
>> Windows can automount floppies in part because it makes the assumption that
>> there is only one user on the machine--an assumption which Unices by design
>> can't make.
> Nope. DOS, Win/DOS, and NT don't "automount" anything.
... with exception of network devices (to some degree).
> They don't have the concept of mounting media.
They have. This concept is NOT used with removeable drives, though (they
are NOT automounted; they are just of different type with different rules
for game), but it's there: you can mount encrypted drive or network server
and it's mounting in the same sense of word as in Unix world (it's even
called "mount" internally).
> The other challenge presented by this request is that of maintaining cache
> coherency. MS' products avoid this problem by not caching writes to the
> floppy.
Not exactly right. They not caching writes to old style floppies, but they
WILL cache writes to ZIP and they WILL lock floppy in ZIP drive to do so
safely.
> Caching is the chief technical reason for operating systems to have the
> mount/unmount concept.
Not exactly true (see above).
> Actually there is a gadget called SMARTDRV which adds an I/O block cache
> to DOS or Win/DOS.
It was LONG AGO. Windows9X and NT have integrated cache (just like Linux).
> In effect it must automount any volume that it is directed to cache.
> It has no way to auto-UNmount, though; if you allow it
> to cache writes then IIRC you must give an explicit command before
> removing a removable volume or data loss can occur. So these products
> have the same difficulty that is being discussed here.
Yeah. But this difficulty was also solved. Long ago. When you'll try to eject
floppy from ZIP (push button on ZIP drive) with "files in use" you'll see nice
message in window on screen.
> So, how hard is it to make MS_SYNC default for FAT-based filesystems, and
> is this sufficient to prevent data loss due to lazy writing?
FAT-based filesystems exist not only on floppies... And floppy can use ext2
(ZIP drive can have NTFS or reiserfs as well :-)...
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:17 EST