atapi write support? No

atapi write support? No

[Ed Sweetman]

Is anyone able to actually use the atapi write support present in the 
later cdrecord releases?  2.6 can't seem to work with it at all.  Is 
this due to the kernel being broken or cdrecord not being up to date 
with 2.6 semantics?

Re: atapi write support? No

[Jens Axboe]

On Tue, Sep 09 2003, Ed Sweetman wrote:
> Is anyone able to actually use the atapi write support present in the 
> later cdrecord releases?  2.6 can't seem to work with it at all.  Is 

Based on the clues you pass above, noone can help you. What are you
trying to do and how? What kernel version? What cdrecord version?

> this due to the kernel being broken or cdrecord not being up to date 
> with 2.6 semantics?

I'll polish my crystal ball and let you know!

-- 
Jens Axboe

Re: atapi write support? No

[Ed Sweetman]

Jens Axboe wrote:
> On Tue, Sep 09 2003, Ed Sweetman wrote:
> 
>>Is anyone able to actually use the atapi write support present in the 
>>later cdrecord releases?  2.6 can't seem to work with it at all.  Is 
> 
> 
> Based on the clues you pass above, noone can help you. What are you
> trying to do and how? What kernel version? What cdrecord version?

There is no other information needed. By use atapi write support i mean 
Get it to do anything besides error out reporting that it cant access 
the drive. If you can query the drive much less actually write anything 
to it using the ATAPI interface than that's more than i've been able to do.

for example   cdrecord dev=ATAPI:1,0,0 checkdisk

1,0,0 should conform to secondary channel master as this is how devfs 
sets the cdr up too.

kernel version, if you look above it's 2.6, but for the sake of 
comparison, test4 or test5 with or without mm patches.

cdrecord version: atapi support was introduced with 2.0 but again, for 
the sake of comparison, anything after a13.



> 
>>this due to the kernel being broken or cdrecord not being up to date 
>>with 2.6 semantics?
> 
> 
> I'll polish my crystal ball and let you know!
> 

Re: atapi write support? No

[Jens Axboe]

On Tue, Sep 09 2003, Ed Sweetman wrote:
> Jens Axboe wrote:
> >On Tue, Sep 09 2003, Ed Sweetman wrote:
> >
> >>Is anyone able to actually use the atapi write support present in the 
> >>later cdrecord releases?  2.6 can't seem to work with it at all.  Is 
> >
> >
> >Based on the clues you pass above, noone can help you. What are you
> >trying to do and how? What kernel version? What cdrecord version?
> 
> There is no other information needed. By use atapi write support i mean 
> Get it to do anything besides error out reporting that it cant access 
> the drive. If you can query the drive much less actually write anything 

That's no info in itself. Clearly you have never had to deal with any
sort of support, or you would know that you should not post and ask for
help in such a way.

> to it using the ATAPI interface than that's more than i've been able to do.
> 
> for example   cdrecord dev=ATAPI:1,0,0 checkdisk
> 
> 1,0,0 should conform to secondary channel master as this is how devfs 
> sets the cdr up too.

And a quick seach of lkml would have shown you that you should use
dev=/dev/hdX.

> cdrecord version: atapi support was introduced with 2.0 but again, for 
> the sake of comparison, anything after a13.

You must mistake that support for something else. Support for block
SG_IO was introduced with 1.11a37 iirc, so anything later than that
should work.

-- 
Jens Axboe

Re: atapi write support? No

[Markus Plail]

On Tue, 09 Sep 2003, Ed Sweetman wrote:

>Jens Axboe wrote: 
>> On Tue, Sep 09 2003, Ed Sweetman wrote:
> There is no other information needed.

There is...

> By use atapi write support i mean Get it to do anything besides error
> out reporting that it cant access the drive. If you can query the
> drive much less actually write anything to it using the ATAPI
> interface than that's more than i've been able to do.
> 
> for example   cdrecord dev=ATAPI:1,0,0 checkdisk

ATAPI: is most likely wrong for what you want to do. It's meant for
notebooks (PCATA or something).
If you just want to get rid of ide-scsi, you have to use dev=/dev/hdX in
cdrecord.

regards
Markus

PS: A little change in attitude towards people who are willing to help
you wouldn't be the worst idea. IMHO of course.

Re: atapi write support? No

[Ed Sweetman]

Markus Plail wrote:
> On Tue, 09 Sep 2003, Ed Sweetman wrote:
> 
> 
>>Jens Axboe wrote: 
>>
>>>On Tue, Sep 09 2003, Ed Sweetman wrote:
>>
>>There is no other information needed.
> 
> 
> There is...

You seemed to get it without any more.

> 
>>By use atapi write support i mean Get it to do anything besides error
>>out reporting that it cant access the drive. If you can query the
>>drive much less actually write anything to it using the ATAPI
>>interface than that's more than i've been able to do.
>>
>>for example   cdrecord dev=ATAPI:1,0,0 checkdisk
> 
> 
> ATAPI: is most likely wrong for what you want to do. It's meant for
> notebooks (PCATA or something).
> If you just want to get rid of ide-scsi, you have to use dev=/dev/hdX in
> cdrecord.

this method states that the method of access is unsupported and 
unintentional.  Which is why i didn't think that it was the right way to 
use cdrecord on atapi devices without ide-scsi.


> regards
> Markus
> 
> PS: A little change in attitude towards people who are willing to help
> you wouldn't be the worst idea. IMHO of course.
> 

If you make what is a general question too specific with details you 
limit your responses if anyone thinks their response is correct for you 
anyway.  I limited my question as much as i wanted to, with the desired 
effect no less.


apparently cdrecord's documention is a little behind it's code. Now 
tracking down why it seems to be botching audio cds for me would require 
a full bugreport style mail now that i know cdrecord is being used in 
the correct manner.

Re: atapi write support? No

[Markus Plail]

On Tue, 09 Sep 2003, Ed Sweetman wrote:

>Markus Plail wrote: 
>> On Tue, 09 Sep 2003, Ed Sweetman wrote:
>>>There is no other information needed.
>> There is...
> 
> You seemed to get it without any more.

Nope. Without the dev=ATAPI command line I wouldn't have had a clue
what could be wrong.

>> ATAPI: is most likely wrong for what you want to do. It's meant for
>> notebooks (PCATA or something).  If you just want to get rid of
>> ide-scsi, you have to use dev=/dev/hdX in cdrecord.
> 
> this method states that the method of access is unsupported and
> unintentional.  Which is why i didn't think that it was the right way
> to use cdrecord on atapi devices without ide-scsi.

The message is in because Jörg Schilling doesn't like the current
implementation at all. So he says one shouldn't use it, in order to
'force' kernel developers in another direction, e.g. Jeff Garzik's
libata.

>> PS: A little change in attitude towards people who are willing to
>> help you wouldn't be the worst idea. IMHO of course.
> 
> If you make what is a general question too specific with details you
> limit your responses if anyone thinks their response is correct for
> you anyway.  I limited my question as much as i wanted to, with the
> desired effect no less.

It had the effect that one had to ask you for more information before
one could give you an answer.

> apparently cdrecord's documention is a little behind it's code.

I don't think it is. If you don't find documentation for dev=/dev/hdX
then that's probably because Jörg doesn't want it to be used. I have
not checked myself though.

regars
Markus

Re: atapi write support? No

[Jens Axboe]

On Tue, Sep 09 2003, Markus Plail wrote:
> On Tue, 09 Sep 2003, Ed Sweetman wrote:
> 
> >Jens Axboe wrote: 
> >> On Tue, Sep 09 2003, Ed Sweetman wrote:
> > There is no other information needed.
> 
> There is...
> 
> > By use atapi write support i mean Get it to do anything besides error
> > out reporting that it cant access the drive. If you can query the
> > drive much less actually write anything to it using the ATAPI
> > interface than that's more than i've been able to do.
> > 
> > for example   cdrecord dev=ATAPI:1,0,0 checkdisk
> 
> ATAPI: is most likely wrong for what you want to do. It's meant for
> notebooks (PCATA or something).
> If you just want to get rid of ide-scsi, you have to use dev=/dev/hdX in
> cdrecord.

That ATAPI support is slow and unreliable, Joerg was a fool to merge it.
It shold _not_ be used! Using dev=/dev/hdX is required for SG_IO in 2.6
right now, other methods should be usable in the future. So Markus is
dead right.

> PS: A little change in attitude towards people who are willing to help
> you wouldn't be the worst idea. IMHO of course.

Indeed, and doing just a little work before showing up with an attitude
would be even better. It's amazing how people asking for help think they
can define the rules as well.

-- 
Jens Axboe