Re: cdrecord hangs my computer

Re: cdrecord hangs my computer

[Paul Marinceu]

Just thought I'd add some positive feedback to this thread, if a bit
late (and not really related to cdrecord).

I updated my laptop to -test11 a couple of hours ago and burnt my first
cd using ide-cd. Not having made (nor planning to make) any benchmarks,
the following is to be taken with a grain of salt:

I used a recent xcdroast with a Matshita (Panasonic) ATAPI CD-RW/DVD-ROM
combo drive (model UJDA740) and was impressed. It Just Worked (TM) and
performance was excellent (at 16x), barely using any cpu at all.

Although ide-scsi has been faithful to me, I'm glad that the legacy
emulated scsi code has finally been replaced in favour of native
ide/atapi. I don't really care whether it's dev=0,0,0 or dev=/dev/cdrom,
though from a HCI point of view, the latter is better, but I definitely
want to thank those responsible. I'm sure that a lot of effort has gone
into making this possible.

My laptop _doesn't have_ any scsi hardware, so why should I then have to
select the scsi tree in my kernel config?? Happily, no more.


-- 
 Paul Marinceu
 http://elixxir.ucc.asn.au

Re: cdrecord hangs my computer

[Douglas Gilbert]

Linus Torvalds wrote:
 > On Sat, 6 Dec 2003, Wakko Warner wrote:
 > > >
 > > > One is just plain confusion - anybody who uses cdrecord has
 > > > either been
 > > > confused by the silly SCSI numbering (while "dev=/dev/hdc" is not
 > > > confusing at all, and uses the same device you use for
 > > > mounting the thing etc).
 > >
 > > Actually, it would be nice if I could use dev=/dev/scd0.  I
 > > do have a scsi burner (and an ide one too)
 >
 > It _should_ just work these days. Anything that uses "cdrom_ioctl()"
 > should automatically get the SCSI command translation code (which
 > isn't part of the scsi driver).

Yes dev=/dev/scd0 should work for "real" SCSI (and USB, IEEE1394
and sATA [via libata] attached) cd/dvd players in lk 2.6.
Copying the SG_IO ioctl and friends into the block
layer isn't exactly pretty in lk 2.6. No doubt I'll be hearing from
the author of cdrecord about some of the rough edges. Basically
cdrecord is tricked into believing it is talking to an sg device.

One rough edge is cdrecord's use of the SCSI_IOCTL_GET_IDLUN ioctl
which encodes bus/channel/target/lun into an integer. cdrecord
uses this for its dev=<n,m,q> notation. The drivers/block/scsi_ioctl.c
implementation returns 0 in all cases. So if you have 2 or more
ATAPI cd/dvd burners cdrecord's dev=<n,m,q> usage won't be able
to differentiate.

ide-scsi has always had problems (I spent about a week on it
and gave up with only a few minor fixes to report) but it
may be a useful "insurance" driver to keep around in lk 2.6 .
[It is also needed for ATAPI tapes so its deprecatation
warning might like to take into account the peripheral device
type.]

Yes the bus/channel/target/lun notation is dated but it has been
in place for around 15 years (in the Sun's bus/target/lun form).
These days the "bus/channel" component is defined by the host OS,
the "target" by the transport (see "object identifiers" in SAM-3
Annex A). Only the "lun" is defined within the SCSI^H^H^Htorage
Architecture Model (SAM-3 section 4.9 in the latest draft).

 > But hey, the scsi layer confuses me. Less than it used to, but still.

Well making the scsi layer handle some of the the most
sophisticated storage devices and some of the most brain
damaged at the same time is proving quite a challenge.
With libata (and later SAS) sATA disks will be getting to
the application space via the sd driver. And how will
object storage devices fit into Linux's block-centric I/O
architecture?

Doug Gilbert

Re: cdrecord hangs my computer

[Jens Axboe]

On Mon, Dec 08 2003, Douglas Gilbert wrote:
> Yes dev=/dev/scd0 should work for "real" SCSI (and USB, IEEE1394
> and sATA [via libata] attached) cd/dvd players in lk 2.6.
> Copying the SG_IO ioctl and friends into the block
> layer isn't exactly pretty in lk 2.6. No doubt I'll be hearing from
> the author of cdrecord about some of the rough edges. Basically
> cdrecord is tricked into believing it is talking to an sg device.

It's not tricked, cdrecord should not care about device type. All it
needs to care about is SG_IO working or not.

> One rough edge is cdrecord's use of the SCSI_IOCTL_GET_IDLUN ioctl
> which encodes bus/channel/target/lun into an integer. cdrecord
> uses this for its dev=<n,m,q> notation. The drivers/block/scsi_ioctl.c
> implementation returns 0 in all cases. So if you have 2 or more
> ATAPI cd/dvd burners cdrecord's dev=<n,m,q> usage won't be able
> to differentiate.

Not sure I see any merrit in supporting that at all, but if so it would
be pretty easy to fake these.

> ide-scsi has always had problems (I spent about a week on it
> and gave up with only a few minor fixes to report) but it
> may be a useful "insurance" driver to keep around in lk 2.6 .
> [It is also needed for ATAPI tapes so its deprecatation
> warning might like to take into account the peripheral device
> type.]

It does, the 2.6 warning printed is just for CDROM devices.

> Well making the scsi layer handle some of the the most
> sophisticated storage devices and some of the most brain
> damaged at the same time is proving quite a challenge.
> With libata (and later SAS) sATA disks will be getting to
> the application space via the sd driver. And how will
> object storage devices fit into Linux's block-centric I/O
> architecture?

What makes you say that Linux has a block-centric IO architecture? 2.6
block io layer is quite happy to do byte-granularity SCSI commands for
you.

-- 
Jens Axboe

Re: cdrecord hangs my computer

[Linus Torvalds]

On Thu, 11 Dec 2003, Jens Axboe wrote:
>
> What makes you say that Linux has a block-centric IO architecture? 2.6
> block io layer is quite happy to do byte-granularity SCSI commands for
> you.

Indeed.

I don't think some people really _realize_ how much cleaner and generic
the generic block layer is compared to SCSI.

Yes, we call it "block layer" for historical reasons, but the fact is,
it's a "packet command" layer with knowledge of blocking (ie the merging
and sorting code has the ability to merge packets that are marked as
mergeable and fit certain criteria).

And the reason it is so much superior to SCSI is that it's designed to be
generic enough that it doesn't _care_ what the device is. The generic
block layer can work with MD, with floppy disks, with traditional SCSI
devices, and it just _works_.

The block layer doesn't have any silly assumptions about what it is
talking to, although it has some helper functions that are directly aimed
at a block device that implements a SCSI-like packet command set. But they
literally are helper functions - the block layer does not force your
floppy device to pretend that it is some kind of strange SCSI disk when it
isn't.

			Linus

Re: cdrecord hangs my computer

[Tero Knuutila]

Hi all!

I put my kernel to pre-emptive (or something like that in processor opts) in 
menuconfig.
After recompilation kernel test11 now performs fine. No more hangs while 
burning cdroms!

So now we know that it was Linus's patch which helped me on ide-scsi 
problems.

However, I can't remember if I tried multiple disks when testing with ide-cd 
mode.
It is possible that this new way also works. Too bad that  cdrecord does not 
tell when there's faulty cd on drive.

But my system works now so thank you everyone!!! Keep up the good work.

Rgds,
    Tero

_________________________________________________________________
Add photos to your messages with MSN 8. Get 2 months FREE*. 
http://join.msn.com/?page=features/featuredemail

Re: cdrecord hangs my computer

[Tero Knuutila]

Hi all!

>Good. However, I'd still like to hear if ide-scsi.c works with the patch:
>it's deprecated and I don't actually encourage people to use it, but it
>would be interesting to hear whether it works for people..

I had to throw away one faulty cd-disk, and now my 2.6-test11 works!!
I did two things while having troubles: Applied Linus's patch and
tried John Bradford's tip to turn off pre-emptive kernel (in processor 
options).

I am currently in a process to put pre-emptive back to enabled so soon I can 
tell if it was the patch or the option change.

I'll be back in one hour. Thanks already for Your help, this test11 kernel 
is fast and goog, I LIKE it!!!

Rgds,

    Tero

_________________________________________________________________
Add photos to your messages with MSN 8. Get 2 months FREE*. 
http://join.msn.com/?page=features/featuredemail

Re: cdrecord hangs my computer

[Tero Knuutila]

Hi Linus, John, Ethan and others!

I tried everything suggested. First I compiled my kernel to use ide-cd 
instead of ide-scsi. No success, cdrecord did hang. Fortunately the whole 
machine did not hang.

Then I tried apply a patch Linus send me. So I changed settings so that HP 
7200 ide burner is seen as ide-scsi device. After that cdrecord still 
freezes totally my machine.
I forgot to mention earlier: I tried cdrecord with older 2.4 kernel and it 
works fine (as ide-scsi device).

So I try to give more accurate info so You can debug this ide-scsi thing 
which is broken. I repeat already mentioned things so that "everything" is 
on one mail:

- Kernel 2.6.0-test11 with one patch on ide-scsi.c
- AMD Duron(tm) processor AuthenticAMD GNU/Linux
- Soltek motherboard: SL-75DRV
- Another CDROM drive: LG GCR-8520B idereader
- HP7200 Ide burner:
cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
  Vendor: HP       Model: CD-Writer+ 7200  Rev: 3.01
  Type:   CD-ROM                           ANSI SCSI revision: 02

And then I guess this is useful too:
-------
lspci
00:00.0 Host bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo KT266/A/333]
00:01.0 PCI bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo KT266/A/333 
AGP]
00:08.0 VGA compatible controller: nVidia Corporation NV5M64 [RIVA TNT2 
Model 64/Model 64 Pro] (rev 15)
00:0a.0 Ethernet controller: Realtek Semiconductor Co., Ltd. 
RTL-8139/8139C/8139C+ (rev 10)
00:0b.0 Multimedia audio controller: Creative Labs SB Live! EMU10k1 (rev 0a)
00:0b.1 Input device controller: Creative Labs SB Live! MIDI/Game Port (rev 
0a)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8233A ISA Bridge
00:11.1 IDE interface: VIA Technologies, Inc. 
VT82C586A/B/VT82C686/A/B/VT8233/A/C/VT8235 PIPC Bus Master IDE (rev 06)
00:11.2 USB Controller: VIA Technologies, Inc. USB (rev 23)
00:11.3 USB Controller: VIA Technologies, Inc. USB (rev 23)
-------

So here's everything I could find. Please let me know what other useful info 
I may submit. I am not guru like You are, so I need advices what to do.

With best regards,
     Tero Knuutila

_________________________________________________________________
Help STOP SPAM with the new MSN 8 and get 2 months FREE*  
http://join.msn.com/?page=features/junkmail

cdrecord hangs my computer

[Tero Knuutila]

Hi!

I am running 2.6.0-test11 kernel on Gentoo (i686 AMD Duron). I have HP 7200 
cd burner.
My system hangs everytime I try to  burn my .wav files to cd with cdrecord. 
Usually few tracks go succesfully but then everything stops. Even the mouse 
won't move and powerbutton does not affect.

And if I have understood correctly, Linux should never hang, so hopefully 
some of You will help me.

I am glad if any responses are CC:t to me (tero1001@hotmail.com). I can't 
subscribe to this list because traffic is so heavy.

With best regards,
       Tero

_________________________________________________________________
STOP MORE SPAM with the new MSN 8 and get 2 months FREE* 
http://join.msn.com/?page=features/junkmail

Re: cdrecord hangs my computer

[Linus Torvalds]

On Sat, 6 Dec 2003, Tero Knuutila wrote:
>
> My system hangs everytime I try to  burn my .wav files to cd with cdrecord.
> Usually few tracks go succesfully but then everything stops. Even the mouse
> won't move and powerbutton does not affect.

Is this with ide-scsi? If so, does the appended patch help?

If not, we really need a whole lot more information (hw config, messages
etc).

		Linus

---
# This is a BitKeeper generated patch for the following project:
# Project Name: Linux kernel tree
# This patch format is intended for GNU patch command version 2.5 or higher.
# This patch includes the following deltas:
#	           ChangeSet	1.1499  -> 1.1500
#	drivers/scsi/ide-scsi.c	1.33    -> 1.34
#
# The following is the BitKeeper ChangeSet Log
# --------------------------------------------
# 03/12/02	torvalds@home.osdl.org	1.1500
# Fix ide-scsi.c uninitialized variable
# --------------------------------------------
#
diff -Nru a/drivers/scsi/ide-scsi.c b/drivers/scsi/ide-scsi.c
--- a/drivers/scsi/ide-scsi.c	Sat Dec  6 00:12:13 2003
+++ b/drivers/scsi/ide-scsi.c	Sat Dec  6 00:12:13 2003
@@ -517,6 +517,7 @@
 	pc->current_position=pc->buffer;
 	bcount.all = IDE_MIN(pc->request_transfer, 63 * 1024);		/* Request to transfer the entire buffer at once */

+	feature.all = 0;
 	if (drive->using_dma && rq->bio) {
 		if (test_bit(PC_WRITING, &pc->flags))
 			feature.b.dma = !HWIF(drive)->ide_dma_write(drive);

Re: cdrecord hangs my computer

[Ethan Weinstein]

Linus Torvalds wrote:
> 
> On Sat, 6 Dec 2003, Tero Knuutila wrote:
> 
>>My system hangs everytime I try to  burn my .wav files to cd with cdrecord.
>>Usually few tracks go succesfully but then everything stops. Even the mouse
>>won't move and powerbutton does not affect.
> 
> 
> Is this with ide-scsi? If so, does the appended patch help?
> 
> If not, we really need a whole lot more information (hw config, messages
> etc).
> 
> 		Linus

I've noted this at boot several times with 2.6.0-test11

Dec  4 23:59:21 e-d0uble kernel: ide-scsi is deprecated for cd burning! 
Use ide-cd and give dev=/dev/hdX as device

Removing the ide-cd bootparams: (I didn't try the patch)

Cdrecord 2.00.3 seems to like the sony-dru500a, denoted as 
--dev=/dev/hdc, I burned several disks this way.
crdao and growisofs don't like to parse /dev/hdc as an argumenet, but 
I'm sure these are easily fixed.

Ethan

Re: cdrecord hangs my computer

[Måns Rullgård]

Ethan Weinstein <lists@stinkfoot.org> writes:

>>>My system hangs everytime I try to  burn my .wav files to cd with cdrecord.
>>>Usually few tracks go succesfully but then everything stops. Even the mouse
>>>won't move and powerbutton does not affect.
>> Is this with ide-scsi? If so, does the appended patch help?
>> If not, we really need a whole lot more information (hw config,
>> messages
>> etc).
>
> I've noted this at boot several times with 2.6.0-test11
>
> Dec  4 23:59:21 e-d0uble kernel: ide-scsi is deprecated for cd
> burning! Use ide-cd and give dev=/dev/hdX as device

And cdrecord keeps saying "Warning: Open by 'devname' is unintentional
and not supported.".  Maybe it's about time someone got rid of that
message.

-- 
Måns Rullgård
mru@kth.se

Re: cdrecord hangs my computer

[Wakko Warner]

> I've noted this at boot several times with 2.6.0-test11
> 
> Dec  4 23:59:21 e-d0uble kernel: ide-scsi is deprecated for cd burning! 
> Use ide-cd and give dev=/dev/hdX as device

At the moment, I don't have a burner on a 2.6.0 machine, however, why is
ide-scsi depreciated?  On every PC I have that has an ide cd drive, I use
ide-scsi.  I like the fact that scd0 is the cdrom drive.  Instead of
guessing if it's hdb hdc or hdd (in the case of this laptop, the dvd was hdb
and the modular cdrw was hdc).

On a side note, has anyone had any luck with Acard's SCSIDE AEC-7722 adapter
on dvd burners?  I was planning on this adapter on an optorite dd0401 dvd
burner.

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals

Re: cdrecord hangs my computer

[John Bradford]

> At the moment, I don't have a burner on a 2.6.0 machine, however, why is
> ide-scsi depreciated?

Basically IDE-SCSI is a work-around to allow userspace programs that
were designed to talk to SCSI devices to use SCSI-like devices
connected to an IDE bus.

This works, but obviously it is better to support things natively.
IDE CD recorders are probably the most popular SCSI-like IDE device
and were therefore quickly supported.  Less common hardware, such as
some IDE MO drives, continues to require IDE-SCSI for the time being.

> On every PC I have that has an ide cd drive, I use
> ide-scsi.  I like the fact that scd0 is the cdrom drive.  Instead of
> guessing if it's hdb hdc or hdd (in the case of this laptop, the dvd was hdb
> and the modular cdrw was hdc).

It's easy enough to write something in userspace to identify which
devices are which and create devices such as /dev/cdrom
automatically - no need to use IDE-SCSI for that.

John.

Re: cdrecord hangs my computer

[Linus Torvalds]

On Sat, 6 Dec 2003, Wakko Warner wrote:
>
> At the moment, I don't have a burner on a 2.6.0 machine, however, why is
> ide-scsi depreciated?

Several reasons.

One is just plain confusion - anybody who uses cdrecord has either been
confused by the silly SCSI numbering (while "dev=/dev/hdc" is not
confusing at all, and uses the same device you use for mounting the thing
etc).

Another is that several things did _not_ work well with ide-scsi. Some
people ended up having to boot with ide-scsi enabled to burn CD's, but
then if they wanted to watch DVD's (on the same drive), they needed to
boot without it.

>		 On every PC I have that has an ide cd drive, I use
> ide-scsi.  I like the fact that scd0 is the cdrom drive.

And you liked the fact that you were supposed to write "dev=0,0,0" or
something strange like that? What a piece of crap it was.

		Linus

Re: cdrecord hangs my computer

[Larry McVoy]

On Sat, Dec 06, 2003 at 01:57:03PM -0800, Linus Torvalds wrote:
> >		 On every PC I have that has an ide cd drive, I use
> > ide-scsi.  I like the fact that scd0 is the cdrom drive.
> 
> And you liked the fact that you were supposed to write "dev=0,0,0" or
> something strange like that? What a piece of crap it was.

Hey, that "piece of crap" has burned one heck of a lot of ISO images of
Linux over the years.  How about a nod of thanks to the author before you
tell him you don't like his interface?  And how about acknowledgement that
he made that "piece of crap" work on a lot of different Unix platforms?
-- 
---
Larry McVoy              lm at bitmover.com          http://www.bitmover.com/lm

Re: cdrecord hangs my computer

[Linus Torvalds]

On Sat, 6 Dec 2003, Larry McVoy wrote:
>
> Hey, that "piece of crap" has burned one heck of a lot of ISO images of
> Linux over the years.

And so does windows. That doesn't make it good.

> How about a nod of thanks to the author before you tell him you don't
> like his interface?

I tried to tell him why numbers are bad. Very politely, explaining that a
lot of devices cannot be enumerated by a traditional "bus/dev/lun" scheme.
He basically cursed at me, and told me that that is how SCSI works. Never
mind that IDE isn't SCSI, and even SCSI doesn't work that way any more
(iSCSI comes to mind).

I can be polite. But when there is no reason to be polite, I can be blunt
too.

				Linus

Re: cdrecord hangs my computer

[Eduard Bloch]

#include <hallo.h>
* Linus Torvalds [Sat, Dec 06 2003, 02:32:08PM]:

> > Hey, that "piece of crap" has burned one heck of a lot of ISO images of
> > Linux over the years.
> 
> And so does windows. That doesn't make it good.

But somehow most Windows programers have easy way to deal with devices,
they have clear paths to get hardware access where on Linux there is
often something not thought out well which ruins your day. Examples?

If you have worked on a SCSI generic device and wish to get all the
block devices associated with this one in upper level, how do you get
that information? I tried to find a way a while ago and it simply
sucked.

For details: http://bugs.debian.org/93633/ (CDROM TOC cache is not
flushed on mount, kernel does not see new sessions on the disk). And
don't tell me that I am stupid and I have simply to reject and reinsert
the CD as everybody does. It is not mandatory on non-Linux, and why? ;)
There must be a more simple way to flush the TOC cache in the block
device driver (where it should not be located whatsoever, IMO).

Another example: how to get a list of module names for each
module corresponding to drivers that have detected hardware? AFAICS
there is no policy, no subsystem to manage this and there are no
guidelines for the driver writers. The only thing you can do is looking
around in procfs, guessing with some hocus pocus.

How to assign a network interface name when loading the NIC driver?

> > How about a nod of thanks to the author before you tell him you don't
> > like his interface?
> 
> I tried to tell him why numbers are bad. Very politely, explaining that a
> lot of devices cannot be enumerated by a traditional "bus/dev/lun" scheme.

Imagine, there would be an internal mapping between devfs names and
"bus/dev/lun" so cdrecord&Co. could query it to get the scsi-generic
device name they need. There is almost always a way to a compromise.

> He basically cursed at me, and told me that that is how SCSI works. Never
> mind that IDE isn't SCSI, and even SCSI doesn't work that way any more
> (iSCSI comes to mind).

Sorry, but if you drive SCSI protocols over IDE hardware, using the same
terminology and the same control methods, why should it not be used as
such? Or was it because Windows works that way so we should make it
different (worse)?

Telling that someone is a dick is easy if you have the absolute power in
your area of control, but listening to him and extracting good arguments
may make sence sometimes.

MfG,
Eduard.

Re: cdrecord hangs my computer

[Linus Torvalds]

On Sun, 7 Dec 2003, Eduard Bloch wrote:
>
> But somehow most Windows programers have easy way to deal with devices,
> they have clear paths to get hardware access where on Linux there is
> often something not thought out well which ruins your day. Examples?

What? There _is_ a very well thought out way of accessing devices in UNIX.
It is way superior to the mess that is windows. It is called a "device
node", and a hypothetical program might use a syntax like this:

	record dev=/dev/hdc

to access the device "/dev/hdc".

This is nothing new. This is how UNIX has worked for the last thirty years
or so. It's not only quite readable, but because everybody uses device
nodes the same way, it's consistent which makes it even more pleasant to
use. So if you were to want to mount the thing you recorded, you'd use

	mount /dev/hdc mntpoint

and notice how we used the same name again?

In contrast, the old cdrecord interfaces are an UNBELIEVABLE PILE OF CRAP!
It's an interface that is based on some random hardware layout mechanism
that isn't even TRUE any more, and hasn't been true for a long time. It's
not helpful to the user, and it doesn't match how devices are accessed by
everything else on the system.

It's bad from a technical standpoint (anybody who names a generic device
with a flat namespace is just basically clueless), and it's bad from a
usability standpoint. It has _zero_ redeeming qualities.

		Linus

Re: cdrecord hangs my computer

[bill davidsen]

In article <Pine.LNX.4.58.0312070812080.2057@home.osdl.org>,
Linus Torvalds  <torvalds@osdl.org> wrote:

| In contrast, the old cdrecord interfaces are an UNBELIEVABLE PILE OF CRAP!
| It's an interface that is based on some random hardware layout mechanism
| that isn't even TRUE any more, and hasn't been true for a long time. It's
| not helpful to the user, and it doesn't match how devices are accessed by
| everything else on the system.
| 
| It's bad from a technical standpoint (anybody who names a generic device
| with a flat namespace is just basically clueless), and it's bad from a
| usability standpoint. It has _zero_ redeeming qualities.

And the redeeming features of naming disks, CDs, and ide-floppy devices
hda..hdx in an order depending on the loading order of the device
drivers?
-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: cdrecord hangs my computer

[Linus Torvalds]

On Mon, 8 Dec 2003, bill davidsen wrote:
> |
> | It's bad from a technical standpoint (anybody who names a generic device
> | with a flat namespace is just basically clueless), and it's bad from a
> | usability standpoint. It has _zero_ redeeming qualities.
>
> And the redeeming features of naming disks, CDs, and ide-floppy devices
> hda..hdx in an order depending on the loading order of the device
> drivers?

.. but you can fix that. Several ways. Make up your own names. Make it
have "/dev/the-cd-with-the-blue-faceplate" if you want, and it will all
still work quite intuitively.

And when you switch the hardware around, and the CD-ROM breaks and you
replace it with another one (still with a blue face-plate, just to not
confuse the user unnecessarily, but this time it ends up being on another
bus entirely), the "/dev/the-cd...-faceplate" thing still works with
minimal effort on the admin part.

And it works in _all_ situations.

Sure, you can have all programs use their own random naming scheme and use
.cdrecordrc and edit that instead, but then you have to remember to edit
the .k3drc thing too, and the /etc/fstab, and so on and so on.

Isn't it saner to use a naming scheme that everybody can agree on, and
that is generic enough that it really _does_ work for everybody, and that
allows localised names?

		Linus

Re: cdrecord hangs my computer

[Bill Davidsen]

On Mon, 8 Dec 2003, Linus Torvalds wrote:

> On Mon, 8 Dec 2003, bill davidsen wrote:
> > |
> > | It's bad from a technical standpoint (anybody who names a generic device
> > | with a flat namespace is just basically clueless), and it's bad from a
> > | usability standpoint. It has _zero_ redeeming qualities.
> >
> > And the redeeming features of naming disks, CDs, and ide-floppy devices
> > hda..hdx in an order depending on the loading order of the device
> > drivers?
> 
> .. but you can fix that. Several ways. Make up your own names. Make it
> have "/dev/the-cd-with-the-blue-faceplate" if you want, and it will all
> still work quite intuitively.

Actually my point was that you were pretty harsh about Joerg's naming
scheme, and the default in Linux is also less than perfect. As you note
these can be fixed, and in current versions of cdrecord after they are
fixed you can use them.

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: cdrecord hangs my computer

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 325 bytes --]

On Mon, 08 Dec 2003 16:21:54 GMT, davidsen@tmr.com (bill davidsen)  said:

> And the redeeming features of naming disks, CDs, and ide-floppy devices
> hda..hdx in an order depending on the loading order of the device
> drivers?

Amen.

At least when network interfaces do it, I can use 'nameif' to beat them
into submission.

[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: cdrecord hangs my computer

[Linus Torvalds]

On Mon, 8 Dec 2003 Valdis.Kletnieks@vt.edu wrote:
>
> Amen.
>
> At least when network interfaces do it, I can use 'nameif' to beat them
> into submission.

What's _wrong_ with you people?

The reason you need 'nameif' for network devices is that the kernel
actually cares about them. But for normal device nodes, you have thousands
of tools to rename them, and you can have a million different names for
the same thing if you want to.

Valdis: for /dev/hdxx, you can rename it with such esoteric programs as
'mv', 'ln', 'perl', 'cp', 'mknod', 'emacs', and a few hundred others. What
is your beef with it?

In fact, every distribution I know of comes with it already aliased to
/dev/cdrom, without you having to lift a pinky to do _anything_ about it.

And quite frankly, anybody who finds

	cdrecord dev=/dev/cdrom

less intuitive than

	cdrecord dev=1,0,0

is so drugged and mindwashed by the cdrecord authors that it's not even
funny any more.

So stop spreading this incredible crap, guys. How about you just admit
that I was right. If that's hard to do, add a comment like

  "Just this once Linus happened to pick a winner. Incredible, but it
   was probably just a fluke. He's still a drugged-out idiot most of the
   time."

to make it feel a bit better.

			Linus

Re: cdrecord hangs my computer

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 1799 bytes --]

On Mon, 08 Dec 2003 10:53:24 PST, Linus Torvalds said:

> Valdis: for /dev/hdxx, you can rename it with such esoteric programs as
> 'mv', 'ln', 'perl', 'cp', 'mknod', 'emacs', and a few hundred others. What
> is your beef with it?

The difference is that with nameif, I can feed it the MAC address and use
that as a "find this interface" key.  And given that at the moment, my lap
top has *4* ethernet devices (an onboard one, one in the docking station,
a wireless card, and one that happens to be be on a Xircom modem card),
it's really handy to be able to be able to nail down the names.

Yes, there's 3 zillion ways to rename the device, once I figure out what it's
name *is*.   Currently, my machine has a nice symlink set up:

% ls -l /dev/cdroms/
total 0
   0 lr-xr-xr-x    1 root     root           33 Dec 31  1969 cdrom0 -> ../ide/host0/bus0/target1/lun0/cd

which devfs was nice enough to do.  My beef is that if I had 2 cdroms, then
there's no guarantee of stability for cdrom0/cdrom1, and unlike the nameif
example, there's no really good way to deal with it (especially when you start
dealing with hotplug devices).

Or as another poster commented, it's easy to use /dev/cdrom-blue-faceplate once
you make it a symlink to the right place.  It's getting that symlink into place
that's the fun part. I admit I haven't looked at the udev stuff - is it able to
look closely enough at devices to do things like "I want the Mitsubishi CDrom
to be cdrom0 and the FireWire/USB/whatever to be cdrom1 if it's my Fujitsu,
but call it cdrom2 otherwise"? If so, then I don't have a beef with it... ;)

The stuff that supports LABEL= on a partition is a *partial* solution to
decouple the name of the device as the system found it from a logical name, but
as many have noted, it has its own issues.


[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: cdrecord hangs my computer

[Ian Soboroff]

Valdis.Kletnieks@vt.edu writes:

> The stuff that supports LABEL= on a partition is a *partial*
> solution to decouple the name of the device as the system found it
> from a logical name, but as many have noted, it has its own issues.

Yes, labels and UUIDs are great for those of us with lots of SCSI
devices, so that adding a controller or changing a cable doesn't
require _two_ boots (one to figure out where everything went and edit
/etc/fstab, one for real).

I wish I could LABEL swap partitions.

Ian

Re: cdrecord hangs my computer

[Mikael Pettersson]

Ian Soboroff writes:
 > Valdis.Kletnieks@vt.edu writes:
 > 
 > > The stuff that supports LABEL= on a partition is a *partial*
 > > solution to decouple the name of the device as the system found it
 > > from a logical name, but as many have noted, it has its own issues.
 > 
 > Yes, labels and UUIDs are great for those of us with lots of SCSI
 > devices, so that adding a controller or changing a cable doesn't
 > require _two_ boots (one to figure out where everything went and edit
 > /etc/fstab, one for real).
 > 
 > I wish I could LABEL swap partitions.

In March 2001 I posted a patch for util-linux to support that:
http://www.csd.uu.se/~mikpe/linux/swap-label/swap-label-patch-2001-03-15

There was only minor interest in the thing, so I've not
pursued it any further.

Re: cdrecord hangs my computer

[Martin J. Bligh]

> In contrast, the old cdrecord interfaces are an UNBELIEVABLE PILE OF CRAP!
> It's an interface that is based on some random hardware layout mechanism
> that isn't even TRUE any more, and hasn't been true for a long time. It's
> not helpful to the user, and it doesn't match how devices are accessed by
> everything else on the system.
> 
> It's bad from a technical standpoint (anybody who names a generic device
> with a flat namespace is just basically clueless), and it's bad from a
> usability standpoint. It has _zero_ redeeming qualities.

I think the appropriate phrase is "user malevolent" software. Making
the user interface fit some arcane technica rather than the user is
rather tragic. Reality is quite complicated enough as it is, without
deliberately setting out to make it more so.

M.

Re: cdrecord hangs my computer

[Bob]

Larry McVoy wrote:

>On Sat, Dec 06, 2003 at 01:57:03PM -0800, Linus Torvalds wrote:
>  
>
>>>		 On every PC I have that has an ide cd drive, I use
>>>ide-scsi.  I like the fact that scd0 is the cdrom drive.
>>>      
>>>
>>And you liked the fact that you were supposed to write "dev=0,0,0" or
>>something strange like that? What a piece of crap it was.
>>    
>>
>
>Hey, that "piece of crap" has burned one heck of a lot of ISO images of
>Linux over the years.  How about a nod of thanks to the author before you
>tell him you don't like his interface?  And how about acknowledgement that
>he made that "piece of crap" work on a lot of different Unix platforms?
>  
>
Naming "1,0,0" won't work everywhere for me.

I started using a 3ware ide card because promise and siig sis
were crashing nforce2, so it's a unified concept ;-) to use
ide-scsi scsi-sg for hd's on 3ware and for cdr cdrw. I can rip from
cdr through pipes to yamaha cdrw, pre-emptive kernel, anticipatory
sched, use dma, devfs. The only quirk is naming "1,0,0" won't work
everywhere--

#/etc/default/cdrecord
CDR_DEVICE=ATAPI:/dev/scsi/host1/bus0/target0/lun0/generic
# ATAPI:1,0,0 won't work in CDR_DEVICE but below it does--
yamaha=   ATAPI:1,0,0   -1      -1      ""

#/usr/src/linux/.config--
# ATA/ATAPI/MFM/RLL support
#
CONFIG_IDE=y
CONFIG_BLK_DEV_IDE=y

# CONFIG_BLK_DEV_HD_IDE is not set
# CONFIG_BLK_DEV_IDEDISK is not set
# CONFIG_BLK_DEV_IDECS is not set
# CONFIG_BLK_DEV_IDECD is not set
# CONFIG_BLK_DEV_IDETAPE is not set
# CONFIG_BLK_DEV_IDEFLOPPY is not set
CONFIG_BLK_DEV_IDESCSI=y

# SCSI device support
#
CONFIG_SCSI=y
CONFIG_SCSI_PROC_FS=y

#
# SCSI support type (disk, tape, CD-ROM)
#
CONFIG_BLK_DEV_SD=y
# CONFIG_CHR_DEV_ST is not set
# CONFIG_CHR_DEV_OSST is not set
CONFIG_BLK_DEV_SR=y
# CONFIG_BLK_DEV_SR_VENDOR is not set
CONFIG_CHR_DEV_SG=y

#
# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
#
CONFIG_SCSI_MULTI_LUN=y
# CONFIG_SCSI_REPORT_LUNS is not set
# CONFIG_SCSI_CONSTANTS is not set
# CONFIG_SCSI_LOGGING is not set

#
# SCSI low-level drivers
#
CONFIG_BLK_DEV_3W_XXXX_RAID=y

-Bob D

Re: cdrecord hangs my computer

[Larry McVoy]

On Sun, Dec 07, 2003 at 11:36:09PM -0500, Bob wrote:
> Larry McVoy wrote:
> >Hey, that "piece of crap" has burned one heck of a lot of ISO images of
> >Linux over the years.  How about a nod of thanks to the author before you
> >tell him you don't like his interface?  And how about acknowledgement that
> >he made that "piece of crap" work on a lot of different Unix platforms?
> > 
> Naming "1,0,0" won't work everywhere for me.

You're missing my point.  I was not claiming that cdrecord's naming 
scheme was great and I'm still not claiming that.  I was objecting to
what I saw as a needless slam on someone who has produced a useful tool,
maintained it for years, and we've all used it for years.

Personally, I think a

    The cdrecord naming scheme is broken, we aren't going to support it.

would have been enough.  Maybe Linus has some issues with the author but
that's no excuse to be that rude.  I'm a little tired of this mailing
list hammering on people more than is needed.
-- 
---
Larry McVoy              lm at bitmover.com          http://www.bitmover.com/lm

Re: cdrecord hangs my computer

[Wakko Warner]

> > At the moment, I don't have a burner on a 2.6.0 machine, however, why is
> > ide-scsi depreciated?
> 
> Several reasons.
> 
> One is just plain confusion - anybody who uses cdrecord has either been
> confused by the silly SCSI numbering (while "dev=/dev/hdc" is not
> confusing at all, and uses the same device you use for mounting the thing
> etc).

Actually, itwould be nice if I could use dev=/dev/scd0.  I do have a scsi
burner (and an ide one too)

> Another is that several things did _not_ work well with ide-scsi. Some
> people ended up having to boot with ide-scsi enabled to burn CD's, but
> then if they wanted to watch DVD's (on the same drive), they needed to
> boot without it.

The joys of modules =)

> >		 On every PC I have that has an ide cd drive, I use
> > ide-scsi.  I like the fact that scd0 is the cdrom drive.
> 
> And you liked the fact that you were supposed to write "dev=0,0,0" or
> something strange like that? What a piece of crap it was.

I have it named using cdrecord's defaults.  People who have real scsi
burners still have to use that format.

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals

Re: cdrecord hangs my computer

[William Lee Irwin III]

At some point in the past, Linus wrote:
>> And you liked the fact that you were supposed to write "dev=0,0,0" or
>> something strange like that? What a piece of crap it was.

On Sat, Dec 06, 2003 at 07:16:50PM -0500, Wakko Warner wrote:
> I have it named using cdrecord's defaults.  People who have real scsi
> burners still have to use that format.

Since I appear to be one of the very few who have a SCSI burner around,
I might as well chime in. ISTR being able to use dev=/dev/scd0 at some
point, though axboe might know more (I'm pretty much at an end-user
level on these kinds issues for the moment).


-- wli

Re: cdrecord hangs my computer

[Linus Torvalds]

On Sat, 6 Dec 2003, Wakko Warner wrote:
> >
> > One is just plain confusion - anybody who uses cdrecord has either been
> > confused by the silly SCSI numbering (while "dev=/dev/hdc" is not
> > confusing at all, and uses the same device you use for mounting the thing
> > etc).
>
> Actually, it would be nice if I could use dev=/dev/scd0.  I do have a scsi
> burner (and an ide one too)

It _should_ just work these days. Anything that uses "cdrom_ioctl()"
should automatically get the SCSI command translation code (which isn't
part of the scsi driver).

But hey, the scsi layer confuses me. Less than it used to, but still.

		Linus

Re: cdrecord hangs my computer

[bill davidsen]

In article <Pine.LNX.4.58.0312061044450.2092@home.osdl.org>,
Linus Torvalds  <torvalds@osdl.org> wrote:

| >		 On every PC I have that has an ide cd drive, I use
| > ide-scsi.  I like the fact that scd0 is the cdrom drive.
| 
| And you liked the fact that you were supposed to write "dev=0,0,0" or
| something strange like that? What a piece of crap it was.

  Actually, dev=0,0,0 or dev=/dev/hdc are neither particularly portable;
each can be something else on another machine. At least /dev/sr0 (or
scd0 if you go to that church) are a bit less likely to change.

  Joerg made the point at one time that the 0,0,0 notation will allow
use of devices with no inode. That's not been useful to me, but it's
probably true ;-)

  If I were going to do that at all, I would have used controller, bus,
device, LUN notation, (0,0,0,0) and been done with it. Joerg marches to
the beat of another drummer, however, maybe even a whole other brass
band. He wrote it, he invites people to not use it if they don't like
it, which I've heard in other contexts. ;-)

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: cdrecord hangs my computer

[Linus Torvalds]

On Mon, 8 Dec 2003, bill davidsen wrote:

> In article <Pine.LNX.4.58.0312061044450.2092@home.osdl.org>,
> Linus Torvalds  <torvalds@osdl.org> wrote:
> |
> | And you liked the fact that you were supposed to write "dev=0,0,0" or
> | something strange like that? What a piece of crap it was.
>
>   Actually, dev=0,0,0 or dev=/dev/hdc are neither particularly portable;
> each can be something else on another machine. At least /dev/sr0 (or
> scd0 if you go to that church) are a bit less likely to change.

Actually, the sane thing to do is to use "/dev/cdrom", which is likely to
be right, and if it isn't, you can always fix it (and you can fix it
_dynamically_ with something like "udev" - so it will do the right thing
even if your cdrom happens to be hot-pluggable).

The only time that ends up being confusing is if you have multiple CD-roms
(which used to be common - DVD reader and CD writer - but is going away
again on "normal" machines due to combo drives). And then you really
actually care about position, so then you are likely happy to go back to
"/dev/hdc" or "/dev/usb-cdrom" or something like that to specify _which_
CD-RW you're talking about.

>   If I were going to do that at all, I would have used controller, bus,
> device, LUN notation, (0,0,0,0) and been done with it.

Well, even that isn't enough.

What's the format for iSCSI? What's the format for buses that are
hierarchical? The thing is, naming is _hard_, and the only way to name
things in a generic manner is:
 - allow multiple levels (ie not a fixed "always 3/4 levels" format)
 - don't use numbers (part of the name might be a hostname or whatever).

Which means that you have to have names that are (a) hierarchical and (b)
strings as the individual path components.

In addition, you clearly _do_ want to have the "simplified" form (aka
the "just point me to the 'cdrom', dammit!" format). Because quite
frankly, the regular user that doesn't care, doesn't really want to know
how his (single) CD-ROM is connected, and if he has to look it up he'll
have all the right in the world to say "This computer is _stupid_".

And guess what you end up with if you have these requirements? A
filesystem. No, "/dev/cdrom" and "/dev/iscsi/cdrom.work.com/cd5" aren't
the _only_ ways of naming things, but they are clear, and they are
sufficient.

And "0,0,0,0" fails _both_ of these obvious requirements. It's neither
clear _nor_ sufficient.

			Linus

Re: cdrecord hangs my computer

[Linus Torvalds]

On Sat, 6 Dec 2003, Ethan Weinstein wrote:
>
> I've noted this at boot several times with 2.6.0-test11
>
> Dec  4 23:59:21 e-d0uble kernel: ide-scsi is deprecated for cd burning!
> Use ide-cd and give dev=/dev/hdX as device

Yup.

> Removing the ide-cd bootparams: (I didn't try the patch)
>
> Cdrecord 2.00.3 seems to like the sony-dru500a, denoted as
> --dev=/dev/hdc, I burned several disks this way.

Good. However, I'd still like to hear if ide-scsi.c works with the patch:
it's deprecated and I don't actually encourage people to use it, but it
would be interesting to hear whether it works for people..

		Linus