IDE controller detection 2.4 +devfs

IDE controller detection 2.4 +devfs

[Rene Rebe]

Hi all.

I like to rereport that the IDE Controllers get strange device-
node-number in Linux-2.4 (currently vanilla 2.4.16) at least using DevFS.

I reported (~a year ago) that when I disable the primary channel of the
on-board controller of the Asus-K7M (Irongate based) the second channel
will be host1 - no host0 can be found.

My today's issue is our K6-2 based server with an Gigabyte Ali-Aladin5
board. The on-board controller is host0 but the additional Promisse
TX2 Ultra100 will be host2 ??? no host1 there ... :

server1:~ # l /dev/ide/
total 0
drwxr-xr-x   1 root     root            0 Jan  1  1970 .
drwxr-xr-x   1 root     root            0 Jan  1  1970 ..
drwxr-xr-x   1 root     root            0 Jan  1  1970 host0
drwxr-xr-x   1 root     root            0 Jan  1  1970 host2

>From /var/log/boot.msg:
<4>ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
<4>PDC20268: IDE controller on PCI bus 00 dev 58
<6>PCI: Found IRQ 11 for device 00:0b.0
<4>PDC20268: chipset revision 2
<4>PDC20268: not 100%% native mode: will probe irqs later
<4>PDC20268: (U)DMA Burst Bit ENABLED Primary MASTER Mode Secondary MASTER Mode.
<4>    ide2: BM-DMA at 0xa800-0xa807, BIOS settings: hde:pio, hdf:pio
<4>    ide3: BM-DMA at 0xa808-0xa80f, BIOS settings: hdg:pio, hdh:pio
<4>ALI15X3: IDE controller on PCI bus 00 dev 78
<4>PCI: No IRQ known for interrupt pin A of device 00:0f.0. Please try using pci
=biosirq.
<4>ALI15X3: chipset revision 193
<4>ALI15X3: not 100%% native mode: will probe irqs later
<4>    ide0: BM-DMA at 0xa400-0xa407, BIOS settings: hda:DMA, hdb:pio
<4>    ide1: BM-DMA at 0xa408-0xa40f, BIOS settings: hdc:pio, hdd:pio
<4>hda: TOSHIBA MK6015MAP, ATA DISK drive
<4>ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
<4>hdc: IBM-DTLA-305040, ATA DISK drive
<4>hde: IBM-DTLA-305040, ATA DISK drive
<4>hdg: IBM-DTLA-305040, ATA DISK drive
<4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
<4>ide1 at 0x170-0x177,0x376 on irq 15

It is really odd to see the Linux kernel mangling this different on each
Workstation ...

k33p h4ck1n6
  René

-- 
René Rebe (Registered Linux user: #127875 <http://counter.li.org>)

eMail:    rene.rebe@gmx.net
          rene@rocklinux.org

Homepage: http://www.tfh-berlin.de/~s712059/index.html

Anyone sending unwanted advertising e-mail to this address will be
charged $25 for network traffic and computing time. By extracting my
address from this message or its header, you agree to these terms.

Re: IDE controller detection 2.4 +devfs

[Richard Gooch]

Rene Rebe writes:
> I like to rereport that the IDE Controllers get strange device-
> node-number in Linux-2.4 (currently vanilla 2.4.16) at least using DevFS.
> 
> I reported (~a year ago) that when I disable the primary channel of the
> on-board controller of the Asus-K7M (Irongate based) the second channel
> will be host1 - no host0 can be found.
> 
> My today's issue is our K6-2 based server with an Gigabyte Ali-Aladin5
> board. The on-board controller is host0 but the additional Promisse
> TX2 Ultra100 will be host2 ??? no host1 there ... :
> 
> server1:~ # l /dev/ide/
> total 0
> drwxr-xr-x   1 root     root            0 Jan  1  1970 .
> drwxr-xr-x   1 root     root            0 Jan  1  1970 ..
> drwxr-xr-x   1 root     root            0 Jan  1  1970 host0
> drwxr-xr-x   1 root     root            0 Jan  1  1970 host2

So what's the problem? It's a similar naming scheme as used for
SCSI. It doesn't matter if you have something plugged into a bus, the
host numbering doesn't change. This is a Feature[tm].

If you want to access your drives according to detection order, use
/dev/discs instead.

				Regards,

					Richard....
Permanent: rgooch@atnf.csiro.au
Current:   rgooch@ras.ucalgary.ca

Re: IDE controller detection 2.4 +devfs

[Rene Rebe]

On Thu, 29 Nov 2001 17:17:59 -0700
Richard Gooch <rgooch@ras.ucalgary.ca> wrote:

> Rene Rebe writes:

[...]

> So what's the problem? It's a similar naming scheme as used for
> SCSI. It doesn't matter if you have something plugged into a bus, the
> host numbering doesn't change. This is a Feature[tm].

Aeh? I can not follow. I feel completely comfortable with the names (strings)
or subdirs, you use. My problem: I have 2 ide-controllers. I would like
to get them as host0 and host1. Boths with the sub-dirs bus0 and bus1.
Reading your answer I though you mean it is fixed due to the pci-id's - but
they do not match ...

And disabling one channel in the bios shouldn't move the controller
from host0 to host1 ... - I do not see the system-behind that ...

Btw. Thanks for DevFS it really ROCKs!! (Except that USBfs exists
and I can not maintain / controll it via the devfsd :-(()

> If you want to access your drives according to detection order, use
> /dev/discs instead.

Yes I know about this.

> 				Regards,
> 
> 					Richard....
> Permanent: rgooch@atnf.csiro.au
> Current:   rgooch@ras.ucalgary.ca


k33p h4ck1n6
  René

-- 
René Rebe (Registered Linux user: #127875 <http://counter.li.org>)

eMail:    rene.rebe@gmx.net
          rene@rocklinux.org

Homepage: http://www.tfh-berlin.de/~s712059/index.html

Anyone sending unwanted advertising e-mail to this address will be
charged $25 for network traffic and computing time. By extracting my
address from this message or its header, you agree to these terms.

Re: IDE controller detection 2.4 +devfs

[Richard Gooch]

Rene Rebe writes:
> On Thu, 29 Nov 2001 17:17:59 -0700
> Richard Gooch <rgooch@ras.ucalgary.ca> wrote:
> 
> > Rene Rebe writes:
> 
> [...]
> 
> > So what's the problem? It's a similar naming scheme as used for
> > SCSI. It doesn't matter if you have something plugged into a bus, the
> > host numbering doesn't change. This is a Feature[tm].
> 
> Aeh? I can not follow. I feel completely comfortable with the names
> (strings) or subdirs, you use. My problem: I have 2
> ide-controllers. I would like to get them as host0 and host1. Boths
> with the sub-dirs bus0 and bus1.  Reading your answer I though you
> mean it is fixed due to the pci-id's - but they do not match ...

No, the "bus" I referred to was the SCSI or IDE bus. And IDE bus
supports two devices (called master and slave) while a SCSI bus
supports many more devices. It has nothing to do with PCI ID's.

> And disabling one channel in the bios shouldn't move the controller
> from host0 to host1 ... - I do not see the system-behind that ...

Um, from your previous message, it seems that host numbering doesn't
change depending on BIOS settings.

So what exactly is happening? And what is the problem? I realise you
may find the naming a little confusing, but is there an actual
problem?

				Regards,

					Richard....
Permanent: rgooch@atnf.csiro.au
Current:   rgooch@ras.ucalgary.ca

Re: IDE controller detection 2.4 +devfs

[Rene Rebe]

On Thu, 29 Nov 2001 17:34:11 -0700
Richard Gooch <rgooch@ras.ucalgary.ca> wrote:

> No, the "bus" I referred to was the SCSI or IDE bus. And IDE bus
> supports two devices (called master and slave) while a SCSI bus
> supports many more devices. It has nothing to do with PCI ID's.

Ok. Hey lets leave away such detail-teaching because I'm shure we both
know enough about all this details and technology (For me I have both
types gambling arround here - and also maintain a whole distro, too ...)

> > And disabling one channel in the bios shouldn't move the controller
> > from host0 to host1 ... - I do not see the system-behind that ...
> 
> Um, from your previous message, it seems that host numbering doesn't
> change depending on BIOS settings.

It was the first storry - my older Athlon.

> So what exactly is happening? And what is the problem? I realise you
> may find the naming a little confusing, but is there an actual
> problem?

ok - again.

I have a K6 on an ALI Aladin-5 board and an additional IDE controller
(Promisse). The onboard can be found as /dev/host0 and the additional
Promisse one appears as /dev/host2:

server1:~ # l /dev/ide/
total 0
drwxr-xr-x   1 root     root            0 Jan  1  1970 .
drwxr-xr-x   1 root     root            0 Jan  1  1970 ..
drwxr-xr-x   1 root     root            0 Jan  1  1970 host0
drwxr-xr-x   1 root     root            0 Jan  1  1970 host2

The boot messages:

<6>Uniform Multi-Platform E-IDE driver Revision: 6.31
<4>ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
<4>PDC20268: IDE controller on PCI bus 00 dev 58
<6>PCI: Found IRQ 11 for device 00:0b.0
<4>PDC20268: chipset revision 2
<4>PDC20268: not 100%% native mode: will probe irqs later
<4>PDC20268: (U)DMA Burst Bit ENABLED Primary MASTER Mode Secondary MASTER Mode.
<4>    ide2: BM-DMA at 0xa800-0xa807, BIOS settings: hde:pio, hdf:pio
<4>    ide3: BM-DMA at 0xa808-0xa80f, BIOS settings: hdg:pio, hdh:pio
<4>ALI15X3: IDE controller on PCI bus 00 dev 78
<4>PCI: No IRQ known for interrupt pin A of device 00:0f.0. Please try using pci
=biosirq.
<4>ALI15X3: chipset revision 193
<4>ALI15X3: not 100%% native mode: will probe irqs later
<4>    ide0: BM-DMA at 0xa400-0xa407, BIOS settings: hda:DMA, hdb:pio
<4>    ide1: BM-DMA at 0xa408-0xa40f, BIOS settings: hdc:pio, hdd:pio
<4>hda: TOSHIBA MK6015MAP, ATA DISK drive
<4>ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
<4>hdc: IBM-DTLA-305040, ATA DISK drive
<4>hde: IBM-DTLA-305040, ATA DISK drive
<4>hdg: IBM-DTLA-305040, ATA DISK drive
<4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
<4>ide1 at 0x170-0x177,0x376 on irq 15

I works - but sucks, because I'm not able to predict the ide-controller
entries in /dev/ide/* because they seem (for me) randomly on each workstaion
I am ...

(All info from my very first mail ...)

The other bug is: On a Athlon-600 workstation based on an Irongate
board (Asus-K7M) I have to disable the first (primarry) channel of
the onbaord IDE controller, because it has problem with the UDMA-66
mode. But when I disable this channel, Linux generates a /dev/ide/host1
entry - No host0 entry is there. Sure it works - but sucks, too!
(Generates a very unstable feeling in me ...)

Thanks for keep reading ;)

> 				Regards,
> 
> 					Richard....
> Permanent: rgooch@atnf.csiro.au
> Current:   rgooch@ras.ucalgary.ca


k33p h4ck1n6
  René

-- 
René Rebe (Registered Linux user: #127875 <http://counter.li.org>)

eMail:    rene.rebe@gmx.net
          rene@rocklinux.org

Homepage: http://www.tfh-berlin.de/~s712059/index.html

Anyone sending unwanted advertising e-mail to this address will be
charged $25 for network traffic and computing time. By extracting my
address from this message or its header, you agree to these terms.

Re: IDE controller detection 2.4 +devfs

[Richard Gooch]

Rene Rebe writes:
> On Thu, 29 Nov 2001 17:34:11 -0700
> Richard Gooch <rgooch@ras.ucalgary.ca> wrote:
> > Um, from your previous message, it seems that host numbering doesn't
> > change depending on BIOS settings.
> 
> It was the first storry - my older Athlon.

OK, so, just to be clear, BIOS settings are *not* changing host
numbering. IOW, "host1" always refers to the same piece of hardware,
whether or not "host1" appears in devfs.

> > So what exactly is happening? And what is the problem? I realise you
> > may find the naming a little confusing, but is there an actual
> > problem?
> 
> ok - again.
> 
> I have a K6 on an ALI Aladin-5 board and an additional IDE controller
> (Promisse). The onboard can be found as /dev/host0 and the additional
> Promisse one appears as /dev/host2:
> 
> server1:~ # l /dev/ide/
> total 0
> drwxr-xr-x   1 root     root            0 Jan  1  1970 .
> drwxr-xr-x   1 root     root            0 Jan  1  1970 ..
> drwxr-xr-x   1 root     root            0 Jan  1  1970 host0
> drwxr-xr-x   1 root     root            0 Jan  1  1970 host2
> 
> The boot messages:
[...]
> I works - but sucks, because I'm not able to predict the
> ide-controller entries in /dev/ide/* because they seem (for me)
> randomly on each workstaion I am ...

But it is actually predictable, isn't it? Think of it this way: the
IDE subsystem reserves "slots" (host numbers) for installed hardware.
If a piece of hardware is disabled in the BIOS, it doesn't mean that
the slot won't be reserved.

> (All info from my very first mail ...)
> 
> The other bug is: On a Athlon-600 workstation based on an Irongate
> board (Asus-K7M) I have to disable the first (primarry) channel of
> the onbaord IDE controller, because it has problem with the UDMA-66
> mode. But when I disable this channel, Linux generates a /dev/ide/host1
> entry - No host0 entry is there. Sure it works - but sucks, too!
> (Generates a very unstable feeling in me ...)

The "host0" entry isn't shown, because it is disabled. But to say
"when I disable this channel, Linux generates a /dev/ide/host1" isn't
correct, and implies a problem where there isn't. The correct way to
describe this is:
"host0" is my primary onboard IDE controller. It might not appear if I
disable it.
"host1" is my secondary onboard IDE controller. It has the same name
whether or not I disable the primary.

And this is a Feature[tm]. It means that tomorrow when a shiny new
drive arrives, you can plug it into your primary channel and enable
the channel in the BIOS. You can then boot without having to fix your
/etc/fstab, because /dev/ide/host1 is still pointing to the same
devices.

The only thing that confuses me is why the secondary onboard channel
is /dev/ide/host1 rather than /dev/ide/host0/bus1.

				Regards,

					Richard....
Permanent: rgooch@atnf.csiro.au
Current:   rgooch@ras.ucalgary.ca

Re: IDE controller detection 2.4 +devfs

[Rene Rebe]

On Thu, 29 Nov 2001 18:11:21 -0700
Richard Gooch <rgooch@ras.ucalgary.ca> wrote:

> But it is actually predictable, isn't it? Think of it this way: the
> IDE subsystem reserves "slots" (host numbers) for installed hardware.
> If a piece of hardware is disabled in the BIOS, it doesn't mean that
> the slot won't be reserved.

It would be nice if it would be that way - se below.

> > (All info from my very first mail ...)
> > 
> > The other bug is: On a Athlon-600 workstation based on an Irongate
> > board (Asus-K7M) I have to disable the first (primarry) channel of
> > the onbaord IDE controller, because it has problem with the UDMA-66
> > mode. But when I disable this channel, Linux generates a /dev/ide/host1
> > entry - No host0 entry is there. Sure it works - but sucks, too!
> > (Generates a very unstable feeling in me ...)
> 
> The "host0" entry isn't shown, because it is disabled. But to say
> "when I disable this channel, Linux generates a /dev/ide/host1" isn't
> correct, and implies a problem where there isn't. The correct way to
> describe this is:
> "host0" is my primary onboard IDE controller. It might not appear if I
> disable it.
> "host1" is my secondary onboard IDE controller. It has the same name
> whether or not I disable the primary.

No!!!! On the Althon box:
- nothing disabled in BIOS:
/dev/ide/host0/bus0/ - on-board primary channel
/deV/ide/host0/bus1/ - on-board secondary channel

- when I disable the primary channel I get this:
/dev/ide/host1/bus1/ - on-board secondary channel

So as you can see it moves!! From host0 to host1!

> And this is a Feature[tm]. It means that tomorrow when a shiny new
> drive arrives, you can plug it into your primary channel and enable
> the channel in the BIOS. You can then boot without having to fix your
> /etc/fstab, because /dev/ide/host1 is still pointing to the same
> devices.

Yes that is a really cool advantage I now for months (over a year!)!
But here is the next example again:

The K6 server (on-board ALI-Aladin-5 + PCI-Card Promisse controller):

/dev/ide/host0/bus0/ - on-board primary channel
/dev/ide/host0/bus1/ - on-board secondary channel
/dev/ide/host2/bus0/ - Promisse primary channel
/dev/ide/host2/bus1/ - Promisse secondarychannel

So where is host1 ??????

(Maybe I should switch back to Mew as MUA to not let kernel-hackers
see the Sylpheed string in the mail-header :-((()

> The only thing that confuses me is why the secondary onboard channel
> is /dev/ide/host1 rather than /dev/ide/host0/bus1.

Only when disabled in the bios (but PLEASE look at the nice lists above
...)

> 				Regards,
> 
> 					Richard....
> Permanent: rgooch@atnf.csiro.au
> Current:   rgooch@ras.ucalgary.ca


k33p h4ck1n6
  René

-- 
René Rebe (Registered Linux user: #127875 <http://counter.li.org>)

eMail:    rene.rebe@gmx.net
          rene@rocklinux.org

Homepage: http://www.tfh-berlin.de/~s712059/index.html

Anyone sending unwanted advertising e-mail to this address will be
charged $25 for network traffic and computing time. By extracting my
address from this message or its header, you agree to these terms.

Re: IDE controller detection 2.4 +devfs

[Richard Gooch]

Rene Rebe writes:
> On Thu, 29 Nov 2001 18:11:21 -0700
> Richard Gooch <rgooch@ras.ucalgary.ca> wrote:
> 
> > But it is actually predictable, isn't it? Think of it this way: the
> > IDE subsystem reserves "slots" (host numbers) for installed hardware.
> > If a piece of hardware is disabled in the BIOS, it doesn't mean that
> > the slot won't be reserved.
> 
> It would be nice if it would be that way - se below.
> 
> > > (All info from my very first mail ...)
> > > 
> > > The other bug is: On a Athlon-600 workstation based on an Irongate
> > > board (Asus-K7M) I have to disable the first (primarry) channel of
> > > the onbaord IDE controller, because it has problem with the UDMA-66
> > > mode. But when I disable this channel, Linux generates a /dev/ide/host1
> > > entry - No host0 entry is there. Sure it works - but sucks, too!
> > > (Generates a very unstable feeling in me ...)
> > 
> > The "host0" entry isn't shown, because it is disabled. But to say
> > "when I disable this channel, Linux generates a /dev/ide/host1" isn't
> > correct, and implies a problem where there isn't. The correct way to
> > describe this is:
> > "host0" is my primary onboard IDE controller. It might not appear if I
> > disable it.
> > "host1" is my secondary onboard IDE controller. It has the same name
> > whether or not I disable the primary.
> 
> No!!!! On the Althon box:
> - nothing disabled in BIOS:
> /dev/ide/host0/bus0/ - on-board primary channel
> /deV/ide/host0/bus1/ - on-board secondary channel
> 
> - when I disable the primary channel I get this:
> /dev/ide/host1/bus1/ - on-board secondary channel
> 
> So as you can see it moves!! From host0 to host1!

Oh, fuck! Now I see why you're complaining. Yeah, that is busted. I
don't know why this happens. It may be due to some deep and subtle
workings of the IDE code.

Andre: any idea why this is happening?

> > And this is a Feature[tm]. It means that tomorrow when a shiny new
> > drive arrives, you can plug it into your primary channel and enable
> > the channel in the BIOS. You can then boot without having to fix your
> > /etc/fstab, because /dev/ide/host1 is still pointing to the same
> > devices.
> 
> Yes that is a really cool advantage I now for months (over a year!)!
> But here is the next example again:
> 
> The K6 server (on-board ALI-Aladin-5 + PCI-Card Promisse controller):
> 
> /dev/ide/host0/bus0/ - on-board primary channel
> /dev/ide/host0/bus1/ - on-board secondary channel
> /dev/ide/host2/bus0/ - Promisse primary channel
> /dev/ide/host2/bus1/ - Promisse secondarychannel
> 
> So where is host1 ??????

Good question. I wonder what's taking up the "host1" slot?

				Regards,

					Richard....
Permanent: rgooch@atnf.csiro.au
Current:   rgooch@ras.ucalgary.ca

Re: IDE controller detection 2.4 +devfs

[Greg KH]

On Fri, Nov 30, 2001 at 01:27:52AM +0100, Rene Rebe wrote:
> 
> Btw. Thanks for DevFS it really ROCKs!! (Except that USBfs exists
> and I can not maintain / controll it via the devfsd :-(()

USBfs?  What's that?  I don't see that in the kernel, yet :)

thanks,

greg k-h

Re: IDE controller detection 2.4 +devfs

[Rene Rebe]

On Thu, 29 Nov 2001 21:58:55 -0800
Greg KH <greg@kroah.com> wrote:

> On Fri, Nov 30, 2001 at 01:27:52AM +0100, Rene Rebe wrote:
> > 
> > Btw. Thanks for DevFS it really ROCKs!! (Except that USBfs exists
> > and I can not maintain / controll it via the devfsd :-(()
> 
> USBfs?  What's that?  I don't see that in the kernel, yet :)

Hu? 2.4.x:

USB support  ---> Preliminary USB device filesystem

??? !!! - It is normally mounted to /proc/bus/usb ...

(But it is not that great (except it works for using a Canon IXUS
digital camera via gphoto2 ... - but controlling the permissions
really sucks (or I do not know how to do this correctly ... ?)

> thanks,
> 
> greg k-h

k33p h4ck1n6
  René

-- 
René Rebe (Registered Linux user: #127875 <http://counter.li.org>)

eMail:    rene.rebe@gmx.net
          rene@rocklinux.org

Homepage: http://www.tfh-berlin.de/~s712059/index.html

Anyone sending unwanted advertising e-mail to this address will be
charged $25 for network traffic and computing time. By extracting my
address from this message or its header, you agree to these terms.

Re: IDE controller detection 2.4 +devfs

[Greg KH]

On Fri, Nov 30, 2001 at 08:56:20AM +0100, Rene Rebe wrote:
> On Thu, 29 Nov 2001 21:58:55 -0800
> Greg KH <greg@kroah.com> wrote:
> 
> > USBfs?  What's that?  I don't see that in the kernel, yet :)
> 
> Hu? 2.4.x:
> 
> USB support  ---> Preliminary USB device filesystem
> 
> ??? !!! - It is normally mounted to /proc/bus/usb ...

Ah, that's "usbdevfs".  But people keep thinking that it has something
to do with "devfs" so I want to rename it to "usbfs".  It will not be
until 2.7 that the old name can go away.

> (But it is not that great (except it works for using a Canon IXUS
> digital camera via gphoto2 ... - but controlling the permissions
> really sucks (or I do not know how to do this correctly ... ?)

It's a load time option when you load the usbcore module.

thanks,

greg k-h

Re: IDE controller detection 2.4 +devfs

[Rene Rebe]

Are there any new results or comments regarding this issue ??

On Thu, 29 Nov 2001 19:10:52 -0700
Richard Gooch <rgooch@ras.ucalgary.ca> wrote:

> Rene Rebe writes:
> > On Thu, 29 Nov 2001 18:11:21 -0700
> > Richard Gooch <rgooch@ras.ucalgary.ca> wrote:
> > 
> > > But it is actually predictable, isn't it? Think of it this way: the
> > > IDE subsystem reserves "slots" (host numbers) for installed hardware.
> > > If a piece of hardware is disabled in the BIOS, it doesn't mean that
> > > the slot won't be reserved.
> > 
> > It would be nice if it would be that way - se below.
> > 
> > > > (All info from my very first mail ...)
> > > > 
> > > > The other bug is: On a Athlon-600 workstation based on an Irongate
> > > > board (Asus-K7M) I have to disable the first (primarry) channel of
> > > > the onbaord IDE controller, because it has problem with the UDMA-66
> > > > mode. But when I disable this channel, Linux generates a /dev/ide/host1
> > > > entry - No host0 entry is there. Sure it works - but sucks, too!
> > > > (Generates a very unstable feeling in me ...)
> > > 
> > > The "host0" entry isn't shown, because it is disabled. But to say
> > > "when I disable this channel, Linux generates a /dev/ide/host1" isn't
> > > correct, and implies a problem where there isn't. The correct way to
> > > describe this is:
> > > "host0" is my primary onboard IDE controller. It might not appear if I
> > > disable it.
> > > "host1" is my secondary onboard IDE controller. It has the same name
> > > whether or not I disable the primary.
> > 
> > No!!!! On the Althon box:
> > - nothing disabled in BIOS:
> > /dev/ide/host0/bus0/ - on-board primary channel
> > /deV/ide/host0/bus1/ - on-board secondary channel
> > 
> > - when I disable the primary channel I get this:
> > /dev/ide/host1/bus1/ - on-board secondary channel
> > 
> > So as you can see it moves!! From host0 to host1!
> 
> Oh, fuck! Now I see why you're complaining. Yeah, that is busted. I
> don't know why this happens. It may be due to some deep and subtle
> workings of the IDE code.
> 
> Andre: any idea why this is happening?
> 
> > > And this is a Feature[tm]. It means that tomorrow when a shiny new
> > > drive arrives, you can plug it into your primary channel and enable
> > > the channel in the BIOS. You can then boot without having to fix your
> > > /etc/fstab, because /dev/ide/host1 is still pointing to the same
> > > devices.
> > 
> > Yes that is a really cool advantage I now for months (over a year!)!
> > But here is the next example again:
> > 
> > The K6 server (on-board ALI-Aladin-5 + PCI-Card Promisse controller):
> > 
> > /dev/ide/host0/bus0/ - on-board primary channel
> > /dev/ide/host0/bus1/ - on-board secondary channel
> > /dev/ide/host2/bus0/ - Promisse primary channel
> > /dev/ide/host2/bus1/ - Promisse secondarychannel
> > 
> > So where is host1 ??????
> 
> Good question. I wonder what's taking up the "host1" slot?
> 
> 				Regards,
> 
> 					Richard....
> Permanent: rgooch@atnf.csiro.au
> Current:   rgooch@ras.ucalgary.ca


k33p h4ck1n6
  René

-- 
René Rebe (Registered Linux user: #248718 <http://counter.li.org>)

eMail:    rene.rebe@gmx.net
          rene@rocklinux.org

Homepage: http://www.tfh-berlin.de/~s712059/index.html

Anyone sending unwanted advertising e-mail to this address will be
charged $25 for network traffic and computing time. By extracting my
address from this message or its header, you agree to these terms.