Is it possible to support a sector size 8192

Is it possible to support a sector size 8192

[Johann Hanne]

Hi everybody,

i'm trying to access UDO/WORM mediums (which have a sector size of 8192) via a 
Plasmon UDO drive, but all I get is "unsupported sector size 8192" when I try 
to access it via /dev/sdX. drivers/scsi/sd.c suggests that the maximum sector 
size currently supported by Linux is 4096.

How difficult will it get to support 8192? Maybe we can pay somebody for doing 
it? (Kernel 2.4.21, RHEL 3.1).

Cheers, Johann

Re: Is it possible to support a sector size 8192

[Steve McIntyre]

Johann wrote:
>Hi everybody,
>
>i'm trying to access UDO/WORM mediums (which have a sector size of
>8192) via a Plasmon UDO drive, but all I get is "unsupported sector
>size 8192" when I try to access it via /dev/sdX. drivers/scsi/sd.c
>suggests that the maximum sector size currently supported by Linux is
>4096.
>
>How difficult will it get to support 8192? Maybe we can pay somebody
>for doing it? (Kernel 2.4.21, RHEL 3.1).

We (at Plasmon) investigated adding support in the kernel, but gave up
quickly as it looked like it would need a huge amount of work in
multiple areas. Instead, we've started shipping a userland filesystem
driver for UDO based on FUSE. If you'd be interested in that, let me
know...

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
"Because heaters aren't purple!" -- Catherine Pitt

Re: Is it possible to support a sector size 8192

[Christian Iversen]

On Friday 22 July 2005 20:18, Steve McIntyre wrote:
> Johann wrote:
> >Hi everybody,
> >
> >i'm trying to access UDO/WORM mediums (which have a sector size of
> >8192) via a Plasmon UDO drive, but all I get is "unsupported sector
> >size 8192" when I try to access it via /dev/sdX. drivers/scsi/sd.c
> >suggests that the maximum sector size currently supported by Linux is
> >4096.
> >
> >How difficult will it get to support 8192? Maybe we can pay somebody
> >for doing it? (Kernel 2.4.21, RHEL 3.1).
>
> We (at Plasmon) investigated adding support in the kernel, but gave up
> quickly as it looked like it would need a huge amount of work in
> multiple areas. Instead, we've started shipping a userland filesystem
> driver for UDO based on FUSE. If you'd be interested in that, let me
> know...

Wouldn't it be possible to translate the sector size 2:1? For me, it doesn't 
seem that hard to do. But then again, I usually program other things than the 
kernel :)

-- 
Regards,
Christian Iversen

Re: Is it possible to support a sector size 8192

[Bryan Henderson]

>we've started shipping a userland filesystem
>driver for UDO based on FUSE.

How does the userland filesystem driver talk to the device?

Re: Is it possible to support a sector size 8192

[Johann Hanne]

It talks to it via /dev/sg*, i.e.

mount -t udo /dev/sga /mnt/udo

On Friday 22 July 2005 23:21, Bryan Henderson wrote:
> >we've started shipping a userland filesystem
> >driver for UDO based on FUSE.
>
> How does the userland filesystem driver talk to the device?

Re: Is it possible to support a sector size 8192

[Steve McIntyre]

[-- Attachment #1: Type: text/plain, Size: 485 bytes --]

On Fri, Jul 22, 2005 at 02:21:08PM -0700, Bryan Henderson wrote:
>>we've started shipping a userland filesystem
>>driver for UDO based on FUSE.
>
>How does the userland filesystem driver talk to the device?

Using sg. It works quite well...

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
"...In the UNIX world, people tend to interpret `non-technical user'
 as meaning someone who's only ever written one device driver." -- Daniel Pead

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Is it possible to support a sector size 8192

[Johann Hanne]

Wow, an answer directly from Plasmon, that's nice. :-)

I've already got the FUSE driver (udofs-1.1-92.i386.rpm, based on FUSE 1.4, is 
this current?!). It is nice for a standalone drive, but we use a SCSI robotic 
changer in conjunction with an HSM system (Legato DiskXtender 2.9) which 
needs some kind of direct access to the drive. It probes for /dev/nst*, but I 
guess /dev/sd* would also work.

I know this exact configuration is not officially supported, but I'm a bit 
dissapointed that the limiting factor seems to be the Linux kernel. If I 
don't find a solution, it will probably become a Solaris/SPARC solution next 
week. :-(

Cheers, Johann



On Friday 22 July 2005 20:18, Steve McIntyre wrote:
> Johann wrote:
> >Hi everybody,
> >
> >i'm trying to access UDO/WORM mediums (which have a sector size of
> >8192) via a Plasmon UDO drive, but all I get is "unsupported sector
> >size 8192" when I try to access it via /dev/sdX. drivers/scsi/sd.c
> >suggests that the maximum sector size currently supported by Linux is
> >4096.
> >
> >How difficult will it get to support 8192? Maybe we can pay somebody
> >for doing it? (Kernel 2.4.21, RHEL 3.1).
>
> We (at Plasmon) investigated adding support in the kernel, but gave up
> quickly as it looked like it would need a huge amount of work in
> multiple areas. Instead, we've started shipping a userland filesystem
> driver for UDO based on FUSE. If you'd be interested in that, let me
> know...

Re: Is it possible to support a sector size 8192

[Steve McIntyre]

Johann writes:
>Wow, an answer directly from Plasmon, that's nice. :-)

On my own time at the weekend, but I like to help... :-)

>I've already got the FUSE driver (udofs-1.1-92.i386.rpm, based on
>FUSE 1.4, is this current?!).

That's the latest release, yes. We're working on a new release at the
moment.

>It is nice for a standalone drive, but we use a SCSI robotic changer
>in conjunction with an HSM system (Legato DiskXtender 2.9) which
>needs some kind of direct access to the drive. It probes for
>/dev/nst*, but I guess /dev/sd* would also work.

Ah, OK.

>I know this exact configuration is not officially supported, but I'm a bit 
>dissapointed that the limiting factor seems to be the Linux kernel. If I 
>don't find a solution, it will probably become a Solaris/SPARC solution next 
>week. :-(

:-( The problem (for us) is the hardware page size on i386. On an
architecture with a larger native page size (e.g. Alpha, sparc64), I'd
expect you'd be able to simply patch the sd driver to allow 8KB
support. A naive early attempt to do exactly that on i386 showed up
page size issues in the VM/block layer underneath, and we left it
alone at that point.

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
  Getting a SCSI chain working is perfectly simple if you remember that there
  must be exactly three terminations: one on one end of the cable, one on the
  far end, and the goat, terminated over the SCSI chain with a silver-handled
  knife whilst burning *black* candles. --- Anthony DeBoer

Re: Is it possible to support a sector size 8192

[Jay Denebeim]

In article <200507231528.20356.jhml@gmx.net>,
Johann Hanne <jhml@gmx.net> wrote:
>Wow, an answer directly from Plasmon, that's nice. :-)

>I've already got the FUSE driver (udofs-1.1-92.i386.rpm, based on
>FUSE 1.4, is this current?!). It is nice for a standalone drive, but
>we use a SCSI robotic changer in conjunction with an HSM system
>(Legato DiskXtender 2.9) which needs some kind of direct access to
>the drive. It probes for /dev/nst*, but I guess /dev/sd* would also
>work.

I'm a little confused here.  Tape devices have variable sector sizes.
I believe they're required to handle up to 64K in a unix environment.
The drives I work on go up to 256k.

Jay
-- 
* Jay Denebeim  Moderator       rec.arts.sf.tv.babylon5.moderated *
* newsgroup submission address: b5mod@deepthot.org                *
* moderator contact address:    b5mod-request@deepthot.org        *
* personal contact address:     denebeim@deepthot.org             *

Re: Is it possible to support a sector size 8192

[Steve McIntyre]

Jay Denebeim wrote:
>In article <200507231528.20356.jhml@gmx.net>,
>Johann Hanne <jhml@gmx.net> wrote:
>>Wow, an answer directly from Plasmon, that's nice. :-)
>
>>I've already got the FUSE driver (udofs-1.1-92.i386.rpm, based on
>>FUSE 1.4, is this current?!). It is nice for a standalone drive, but
>>we use a SCSI robotic changer in conjunction with an HSM system
>>(Legato DiskXtender 2.9) which needs some kind of direct access to
>>the drive. It probes for /dev/nst*, but I guess /dev/sd* would also
>>work.
>
>I'm a little confused here.  Tape devices have variable sector sizes.
>I believe they're required to handle up to 64K in a unix environment.
>The drives I work on go up to 256k.

Tapes are different. The UDO drive that Johann was asking about is an
optical _disk_ drive...

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
Support the Campaign for Audiovisual Free Expression: http://www.eff.org/cafe/

Re: Is it possible to support a sector size 8192

[Jay Denebeim]

On Mon, 25 Jul 2005, Steve McIntyre wrote:

> Jay Denebeim wrote:
>>> I've already got the FUSE driver (udofs-1.1-92.i386.rpm, based on
>>> FUSE 1.4, is this current?!). It is nice for a standalone drive, but
>>> we use a SCSI robotic changer in conjunction with an HSM system
>>> (Legato DiskXtender 2.9) which needs some kind of direct access to
>>> the drive. It probes for /dev/nst*, but I guess /dev/sd* would also
>>> work.
>>
>> I'm a little confused here.  Tape devices have variable sector sizes.
>> I believe they're required to handle up to 64K in a unix environment.
>> The drives I work on go up to 256k.
>
> Tapes are different. The UDO drive that Johann was asking about is an
> optical _disk_ drive...

Yeah, I got that.  I wasn't going to say anything until he mentioned using 
the tape driver with it.  Is the UDO drive presenting itself as a tape or 
as a disk?

Jay

-- 
* Jay Denebeim  Moderator       rec.arts.sf.tv.babylon5.moderated *
* newsgroup submission address: b5mod@deepthot.org                *
* moderator contact address:    b5mod-request@deepthot.org        *
* personal contact address:     denebeim@deepthot.org             *

Re: Is it possible to support a sector size 8192

[Steve McIntyre]

Jay Denebeim wrote:
>On Mon, 25 Jul 2005, Steve McIntyre wrote:
>
>> Jay Denebeim wrote:
>>>
>>> I'm a little confused here.  Tape devices have variable sector sizes.
>>> I believe they're required to handle up to 64K in a unix environment.
>>> The drives I work on go up to 256k.
>>
>> Tapes are different. The UDO drive that Johann was asking about is an
>> optical _disk_ drive...
>
>Yeah, I got that.  I wasn't going to say anything until he mentioned using 
>the tape driver with it.  Is the UDO drive presenting itself as a tape or 
>as a disk?

It's a disk, SCSI type 7 (optical). The tape thing is a red herring,
IMHO.

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
You lock the door
And throw away the key
There's someone in my head but it's not me 

Re: Is it possible to support a sector size 8192

[Johann Hanne]

> >I'm a little confused here.  Tape devices have variable sector sizes.
> >I believe they're required to handle up to 64K in a unix environment.
> >The drives I work on go up to 256k.
> Tapes are different. The UDO drive that Johann was asking about is an
> optical _disk_ drive...
Hmm... would it be _possible_ to support it as tape device (/dev/st* 
and /dev/nst*)?

Cheers, Johann

Re: Is it possible to support a sector size 8192

[Steve McIntyre]

In article <200507260038.22820.jhml@gmx.net> you write:
>> >I'm a little confused here.  Tape devices have variable sector sizes.
>> >I believe they're required to handle up to 64K in a unix environment.
>> >The drives I work on go up to 256k.
>> Tapes are different. The UDO drive that Johann was asking about is an
>> optical _disk_ drive...
>Hmm... would it be _possible_ to support it as tape device (/dev/st* 
>and /dev/nst*)?

Not that I'm aware, no. It's not designed to be used as a tape at
all. To be honest, if you want a device that looks like a tape,
then... :-)

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
Into the distance, a ribbon of black
Stretched to the point of no turning back