Hi all,
some month ago a change to Documentation/devices.txt was submitted by John
Cagle.
http://linux.bkbits.net:8080/linux-2.6/cset%4040586a32fpYGPUC8ysFeU7GIfmmdUA
The patch changed the major number of the s/390 dasd devices from 94 to 95.
As you can see in include/major.h and drivers/s390/block/dasd.c the change
to the documentation was bogus. The dasd device driver was using and will
be using major number 94.
Unfortunately, the "Inverse NAND Flash Translation Layer", which was added
somewhen during 2.5 now uses the same major number.
I attached a patch to restore the old state but I am not sure, how to deal
with the inftla driver.
Patch to restore the old state
Signed-of-by: Christian Borntraeger <[email protected]>
-------------
diff -ur linux-bk/Documentation/devices.txt
linux-dev/Documentation/devices.txt
--- a/Documentation/devices.txt 2004-09-18 23:20:38.000000000 +0200
+++ b/Documentation/devices.txt 2004-09-18 23:28:48.000000000 +0200
@@ -1683,11 +1683,16 @@
1 = /dev/dcxx1 Second capture card
...
- 94 block Inverse NAND Flash Translation Layer
- 0 = /dev/inftla First INFTL layer
- 16 = /dev/inftlb Second INFTL layer
+ 94 block IBM S/390 DASD block storage
+ 0 = /dev/dasda First DASD device, major
+ 1 = /dev/dasda1 First DASD device, block 1
+ 2 = /dev/dasda2 First DASD device, block 2
+ 3 = /dev/dasda3 First DASD device, block 3
+ 4 = /dev/dasdb Second DASD device, major
+ 5 = /dev/dasdb1 Second DASD device, block 1
+ 6 = /dev/dasdb2 Second DASD device, block 2
+ 7 = /dev/dasdb3 Second DASD device, block 3
...
- 240 = /dev/inftlp 16th INTFL layer
95 char IP filter
0 = /dev/ipl Filter control device/log file
@@ -1696,15 +1701,9 @@
3 = /dev/ipauth Authentication control device/log file
...
- 95 block IBM S/390 DASD block storage
- 0 = /dev/dasd0 First DASD device, major
- 1 = /dev/dasd0a First DASD device, block 1
- 2 = /dev/dasd0b First DASD device, block 2
- 3 = /dev/dasd0c First DASD device, block 3
- 4 = /dev/dasd1 Second DASD device, major
- 5 = /dev/dasd1a Second DASD device, block 1
- 6 = /dev/dasd1b Second DASD device, block 2
- 7 = /dev/dasd1c Second DASD device, block 3
+ 95 block IBM S/390 VM/ESA minidisk
+ 0 = /dev/msd0 First VM/ESA minidisk
+ 1 = /dev/msd1 Second VM/ESA minidisk
...
96 char Parallel port ATAPI tape devices
@@ -1715,11 +1714,6 @@
129 = /dev/npt1 Second p.p. ATAPI tape, no rewind
...
- 96 block IBM S/390 VM/ESA minidisk
- 0 = /dev/msd0 First VM/ESA minidisk
- 1 = /dev/msd1 Second VM/ESA minidisk
- ...
-
97 char Parallel port generic ATAPI interface
0 = /dev/pg0 First parallel port ATAPI device
1 = /dev/pg1 Second parallel port ATAPI device
> -----Original Message-----
> From: Christian Borntraeger [mailto:[email protected]]
> Sent: Saturday, September 18, 2004 5:04 PM
> To: [email protected]; Andrew Morton; Torben Mathiasen
> Cc: Cagle, John
> Subject: [Patch][RFC] conflicting device major numbers in devices.txt
>
> Hi all,
>
> some month ago a change to Documentation/devices.txt was
> submitted by John
> Cagle.
>
> http://linux.bkbits.net:8080/linux-2.6/cset%4040586a32fpYGPUC8
> ysFeU7GIfmmdUA
>
> The patch changed the major number of the s/390 dasd devices
> from 94 to 95.
> As you can see in include/major.h and
> drivers/s390/block/dasd.c the change
> to the documentation was bogus. The dasd device driver was
> using and will
> be using major number 94.
>
> Unfortunately, the "Inverse NAND Flash Translation Layer",
> which was added
> somewhen during 2.5 now uses the same major number.
>
> I attached a patch to restore the old state but I am not
> sure, how to deal
> with the inftla driver.
>
>
> Patch to restore the old state
<snip>
Christian,
Good catch. The patch that originally corrected this (which is
contained within the bitkeeper patch you can view at
http://tinyurl.com/34zoy ) never made it to LANANA, so the
correction of dasd from 95 to 94 was never in our repository.
The recent update of devices.txt into the 2.5 kernel came from
our repository so it contains the same error.
I'm copying David Woodhouse and the Linux-MTD project who (I
think) are the maintainers of the inftl driver. Hopefully, they
will be kind enough to accept a new block major for the inftl
driver. If not, there's probably little chance of device conflict
in real-world use, since intftl is for embedded systems which
probably will not be attached to S/390 storage devices.
Torben -- please correct this in devices.txt and find a new block
major for the Linux-MTD project (if they will accept it).
Thanks,
John Cagle
(former LANANA maintainer)
On Sat, 2004-09-18 at 18:14 -0500, Cagle, John wrote:
> Torben -- please correct this in devices.txt and find a new block
> major for the Linux-MTD project (if they will accept it).
Seems reasonable.
--
dwmw2
On Sun, Sep 19 2004, David Woodhouse wrote:
> On Sat, 2004-09-18 at 18:14 -0500, Cagle, John wrote:
> > Torben -- please correct this in devices.txt and find a new block
> > major for the Linux-MTD project (if they will accept it).
>
> Seems reasonable.
>
Allright. Patch attached.
s/390 dasd moved to major 94.
s/390 VM/ESA moved to major 95.
INFTL moved to major 96.
Andrew:
Patch also adds Marvell MPSC low-density device to Minor 44 & 45 (Major 204).
This one was in my queue.
Thanks,
Torben
On Dienstag, 21. September 2004 11:23, Torben Mathiasen wrote:
> s/390 dasd moved to major 94.
> s/390 VM/ESA moved to major 95.
> INFTL moved to major 96.
Actually, major 95 has never been used for VM minidisks or any other
s390 block device in any 2.4 or 2.6 based distribution, because that
driver was integrated into the dasd driver (it just uses a different
access method on the same devices). You might want to document that
this number is currently unused, even if it doesn't get assigned to
any other driver.
Arnd <><
On Tue, Sep 21 2004, Arnd Bergmann wrote:
> On Dienstag, 21. September 2004 11:23, Torben Mathiasen wrote:
> > s/390 dasd moved to major 94.
> > s/390 VM/ESA moved to major 95.
> > INFTL moved to major 96.
>
> Actually, major 95 has never been used for VM minidisks or any other
> s390 block device in any 2.4 or 2.6 based distribution, because that
> driver was integrated into the dasd driver (it just uses a different
> access method on the same devices). You might want to document that
> this number is currently unused, even if it doesn't get assigned to
> any other driver.
>
So, what you're saying is that Major 95 is not used at all in real life? Then
I'll remove it from from the list completely during the my next push. Let me
know if there's a point in keeping it assigned even if its obsolete.
Torben
On Dienstag, 21. September 2004 18:49, Torben Mathiasen wrote:
> On Tue, Sep 21 2004, Arnd Bergmann wrote:
> > On Dienstag, 21. September 2004 11:23, Torben Mathiasen wrote:
> > > s/390 dasd moved to major 94.
> > > s/390 VM/ESA moved to major 95.
> > > INFTL moved to major 96.
> >
> > Actually, major 95 has never been used for VM minidisks or any other
> > s390 block device in any 2.4 or 2.6 based distribution, because that
> > driver was integrated into the dasd driver (it just uses a different
> > access method on the same devices). You might want to document that
> > this number is currently unused, even if it doesn't get assigned to
> > any other driver.
>
> So, what you're saying is that Major 95 is not used at all in real life? Then
> I'll remove it from from the list completely during the my next push. Let me
> know if there's a point in keeping it assigned even if its obsolete.
I don't see any reason to keep it for s/390, but I added some potentially
interested parties to the CC: list. If anyone is thinking of reusing device
major number 95 for s390 minidisks, speak up now, otherwise it will finally
be gone.
Arnd <><