> > Charles Wilkins wrote:
> >
> > Is there a max number of ide controllers that linux-2.2.18 can
> > support?
>
Andrzej M. Krzysztofowicz says,
>"Linux supports up to 10 IDE channels, however channel numbers of PCI
controllers seem to be assigned first."
Warren Young says,
>"Kernel 2.2 is limited to 4 IDE controllers."
ok, so which is it kernel guys, 4 or 10 IDE controllers for the 2.2.x
kernel?
> I know 2.4 supports a maximum of 10 controllers, but each controller has
> to use a different I/O port. The standard ones are 0x1F0 and 0x170.
> The mobo controllers you have probably are fixed to use the common I/O
> ports. If the Promise controller can be set to use uncommon port
> values, you'll be able to use 4 controllers.
you didn't read the other posts . . .
> The Creative card is probably a loss, because it's probably fixed at I/O
> port 0x170 (second channel).
well, i know this SB32 card can operating on at least 3 different io ports .
. .
Charles Wilkins writes:
> Andrzej M. Krzysztofowicz says,
>
> >"Linux supports up to 10 IDE channels, however channel numbers of PCI
> controllers seem to be assigned first."
>
> Warren Young says,
> >"Kernel 2.2 is limited to 4 IDE controllers."
>
> ok, so which is it kernel guys, 4 or 10 IDE controllers for the 2.2.x
> kernel?
It depends if you have Andre's IDE patches applied to your kernel sources
or not.
> well, i know this SB32 card can operating on at least 3 different io ports .
> . .
It may be that there is some difficulty in the order the IDE cards are
initialized. From your previous dmesg output, it appears that ide3 and ide4
(PCI cards) are initialized before ide2 (ISA card), so they may be stealing
an ioport that the ISA card needs. Try booting with just the SB32 card
and checking /proc/ioports, and then with only the other card, and see
if anything in /proc/ioports (or /proc/interrupts) is conflicting.
Cheers, Andreas
--
Andreas Dilger \ "If a man ate a pound of pasta and a pound of antipasto,
\ would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/ -- Dogbert
> Charles Wilkins writes:
> > Andrzej M. Krzysztofowicz says,
> >
> > >"Linux supports up to 10 IDE channels, however channel numbers of PCI
> > controllers seem to be assigned first."
> >
> > Warren Young says,
> > >"Kernel 2.2 is limited to 4 IDE controllers."
> >
> > ok, so which is it kernel guys, 4 or 10 IDE controllers for the 2.2.x
> > kernel?
>
> It depends if you have Andre's IDE patches applied to your kernel sources
I have ide.2.2.18.1209.patch applied. The kernel is 2.2.18.
So what is the answer? 4 controllers max or 10 for my kernel?
> or not.
>
> > well, i know this SB32 card can operating on at least 3 different io
ports .
> > . .
>
> It may be that there is some difficulty in the order the IDE cards are
> initialized. From your previous dmesg output, it appears that ide3 and
ide4
> (PCI cards) are initialized before ide2 (ISA card), so they may be
stealing
> an ioport that the ISA card needs. Try booting with just the SB32 card
> and checking /proc/ioports, and then with only the other card, and see
> if anything in /proc/ioports (or /proc/interrupts) is conflicting.
>
I have done this. There are no conflicts.
SB32 uses 0x168-0x16f,0x36e,10 and nothing else does.
What would be the correct kernel command to manually set up the SB32 as ide
4 for the above resources?
Kernel 2.4.0-test12 picks up the SB32 controller even with the promise
controller is installed, but I would prefer to stay with the stable 2.2.x
kernel if possible.
Charles
Charles Wilkins writes:
> I have ide.2.2.18.1209.patch applied. The kernel is 2.2.18.
> So what is the answer? 4 controllers max or 10 for my kernel?
10 controllers if you have the IDE patches applied. 4 otherwise.
> I have done this. There are no conflicts.
> SB32 uses 0x168-0x16f,0x36e,10 and nothing else does.
> What would be the correct kernel command to manually set up the SB32 as ide
> 4 for the above resources?
in your lilo.conf, add a line:
append="ide4=0x168,0x36e,10"
See /usr/src/linux/Documentation/ide.txt for more info. You may need to
specify the I/O ports for ide2 and ide3 in this case, as the above I/O
ports are for ide3.
> I have the following hd devices in /dev.
> hda 3,0
> hdb 3,64
> hdc 22,0
> hdd 22,64
> hde 33,0
> hdf 33,64
> hdg 34,0
> hdh 34,64
> How do I go about making devices for hdi, hdj, hdk, hdl, hdm, hdn, etc ?
> i.e. which major and minor numbers do I use?
If you have a reasonably recent /dev/MAKEDEV installed, you should be able
to just do "cd /dev; ./MAKEDEV hdi hdj hdk hdl", but failing that,
hdi = 56,0
hdj = 56,64
hdk = 57,0
hdl = 57,64
You can always check /usr/src/linux/Documentation/devices.txt for more.
Cheers, Andreas
--
Andreas Dilger \ "If a man ate a pound of pasta and a pound of antipasto,
\ would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/ -- Dogbert
>
> Charles Wilkins writes:
> > I have ide.2.2.18.1209.patch applied. The kernel is 2.2.18.
> > So what is the answer? 4 controllers max or 10 for my kernel?
>
> 10 controllers if you have the IDE patches applied. 4 otherwise.
Look the source Luck ...
2.2.18 ide.c:
static const byte ide_hwif_to_major[] = {IDE0_MAJOR, IDE1_MAJOR,
IDE2_MAJOR, IDE3_MAJOR, IDE4_MAJOR, IDE5_MAJOR };
6 otherwise
Andrzej
Here is what worked.
append="ide6=0x168,0x36e,10"
Thanks all for your help.
Merry Christmas : )
> > Charles Wilkins writes:
> > > I have ide.2.2.18.1209.patch applied. The kernel is 2.2.18.
> > > So what is the answer? 4 controllers max or 10 for my kernel?
> >
> > 10 controllers if you have the IDE patches applied. 4 otherwise.
>
> Look the source Luck ...
>
> 2.2.18 ide.c:
>
> static const byte ide_hwif_to_major[] = {IDE0_MAJOR, IDE1_MAJOR,
> IDE2_MAJOR, IDE3_MAJOR, IDE4_MAJOR, IDE5_MAJOR };
>
> 6 otherwise
>
> Andrzej