I have an old Intel SC450NX server (sitka mainboard in Cabrillo
chassis) with integrated 53c810 and 53c896 controllers. The 53c810
and the primary channel of the 53c896 come up fine and work without
issues. The secondary channel initializes properly in BIOS and all
the connected drives are visible there but they are not accessible
under Linux (Ubuntu Hardy, Ubuntu Intrepid, and Debian Lenny). I can
see the secondary channel using lspci but no kernel drivers are loaded
for it as they are for the other two channels (as shown by "lspci
-nv"). Why won't Linux load a driver for the secondary channel?
>
> I have an old Intel SC450NX server (sitka mainboard in Cabrillo
> chassis) with integrated 53c810 and 53c896 controllers. The 53c810
> and the primary channel of the 53c896 come up fine and work without
> issues. The secondary channel initializes properly in BIOS and all
> the connected drives are visible there but they are not accessible
> under Linux (Ubuntu Hardy, Ubuntu Intrepid, and Debian Lenny). I can
> see the secondary channel using lspci but no kernel drivers are loaded
> for it as they are for the other two channels (as shown by "lspci
> -nv"). Why won't Linux load a driver for the secondary channel?
I would think it would use the same driver for the secondary channel as
the primary channel. What does dmesg show?
>
--
Robert Heller -- Get the Deepwoods Software FireFox Toolbar!
Deepwoods Software -- Linux Installation and Administration http://www.deepsoft.com/ -- Web Hosting, with CGI and Database heller@deepsoft.com -- Contract Programming: C/C++, Tcl/Tk
On Thu, 14 Aug 2008 20:07:57 -0500, Robert Heller
<heller@deepsoft.com> wrote:
>I would think it would use the same driver for the secondary channel as
>the primary channel. What does dmesg show?
I expected the same thing. Looking at dmesg did give another clue:
lines relating to secondary:
ACPI: PCI Interrupt 0000:01:03.1[b] -> GSI 56 (level, low) -> IRQ 56
sym53c8xx 0000:01:03.1: not initializing, driven by RAID controller.
There is no RAID controller installed. There was an old AcceleRAID
250 card installed and this motherboard does support SISL
electrically, but I removed it because SISL was disabled by recent
BIOS releases and no longer works. There is a DOS utility to
re-enable it, but Intel's web site says to contact support to get it
and they no longer support this board by phone or email (nice little
Catch-22, that). The 250 has only a VHDCI connector externally and I
cannot find my only VHDCI<>MD68M cable so I decided on that basis to
just do a software RAID setup. Besides, I have more than 15 drives
to connect in the RAID anyway, 8x36GB and 11x73GB. All 19 of these,
plus the 68pin 18GB boot drive are connected and visible in the LSI
BIOS.
BTW, the RAID card was removed before installing the OS and I
formatted the only partitioned drive during the install, so I don't
know why the kernel thinks there is a RAID controller.
Another check of dmesg shows no other messages relating to any RAID
controller.
If there is some way to enable SISL under Linux, I'm all ears. I
could just ditch the 18GB drive and install an IDE drive to boot from
then reinstall the AcceleRAID 250 and get a hardware RAID solution.
Re: sym53c896 secondary channel not active (Solved?)
I did a little poking around about the message that the secondary
channel was being controlled by a RAID controller. It seems the linux
driver does this if a specific flag is set in the controller's NVRAM.
I found no facility to change this in the LSI BIOS but decided to just
disable BIOS control of that channel altogether. This seemed to have
no effect upon rebooting so I decided to just install an old IDE
drive, copy the root partition to it and disable the SCSI BIOS
altogether. I can only assume a cold boot was necessary to completely
disable the secondary channel from BIOS control as it came up on the
next boot into Debian and all 14 drives attached to it are now visible
and populated in /dev.
Thanks for the tip about dmesg. The clue I got from there gave me
what I needed to sort this out. I'd looked at it before but never so
closely.
>
> On Thu, 14 Aug 2008 20:07:57 -0500, Robert Heller
> <heller@deepsoft.com> wrote:
> >I would think it would use the same driver for the secondary channel as
> >the primary channel. What does dmesg show?
>
> I expected the same thing. Looking at dmesg did give another clue:
>
> lines relating to primary:
> ACPI: PCI Interrupt 0000:01:03.0[A] -> GSI 57 (level, low) -> IRQ 57
> sym0: <896> rev 0x7 at pci 0000:01:03.0 irq 57
>
> lines relating to secondary:
> ACPI: PCI Interrupt 0000:01:03.1[b] -> GSI 56 (level, low) -> IRQ 56
> sym53c8xx 0000:01:03.1: not initializing, driven by RAID controller.
>
>
> There is no RAID controller installed. There was an old AcceleRAID
> 250 card installed and this motherboard does support SISL
> electrically, but I removed it because SISL was disabled by recent
> BIOS releases and no longer works. There is a DOS utility to
> re-enable it, but Intel's web site says to contact support to get it
> and they no longer support this board by phone or email (nice little
> Catch-22, that). The 250 has only a VHDCI connector externally and I
> cannot find my only VHDCI<>MD68M cable so I decided on that basis to
> just do a software RAID setup. Besides, I have more than 15 drives
> to connect in the RAID anyway, 8x36GB and 11x73GB. All 19 of these,
> plus the 68pin 18GB boot drive are connected and visible in the LSI
> BIOS.
I wonder if the controller with the secondary channel still thinks the
RAID controller is still connected somehow, or in fact has a bit set
somewhere in its NV RAM / EEPROM relating to the RAID controller. There
might be something in the SCSI BIOS (as opposed to the main motherboard
BIOS) that needs to be reset...
>
> BTW, the RAID card was removed before installing the OS and I
> formatted the only partitioned drive during the install, so I don't
> know why the kernel thinks there is a RAID controller.
The SCSI controller firmware might be (erroniously) telling it so.
>
> Another check of dmesg shows no other messages relating to any RAID
> controller.
>
> If there is some way to enable SISL under Linux, I'm all ears. I
> could just ditch the 18GB drive and install an IDE drive to boot from
> then reinstall the AcceleRAID 250 and get a hardware RAID solution.
>
--
Robert Heller -- Get the Deepwoods Software FireFox Toolbar!
Deepwoods Software -- Linux Installation and Administration http://www.deepsoft.com/ -- Web Hosting, with CGI and Database heller@deepsoft.com -- Contract Programming: C/C++, Tcl/Tk
>
> I did a little poking around about the message that the secondary
> channel was being controlled by a RAID controller. It seems the linux
> driver does this if a specific flag is set in the controller's NVRAM.
> I found no facility to change this in the LSI BIOS but decided to just
> disable BIOS control of that channel altogether. This seemed to have
> no effect upon rebooting so I decided to just install an old IDE
> drive, copy the root partition to it and disable the SCSI BIOS
> altogether. I can only assume a cold boot was necessary to completely
> disable the secondary channel from BIOS control as it came up on the
> next boot into Debian and all 14 drives attached to it are now visible
> and populated in /dev.
>
> Thanks for the tip about dmesg. The clue I got from there gave me
> what I needed to sort this out. I'd looked at it before but never so
> closely.
One of the downsides to the 'modern' graphical boot screens is the lack
of boot up messages. I *always* disable the graphical boot and remove
the -quiet flag from the kernel command line. Yes, it makes for a
verbose boot up sequence, but it something is misbehaving, I get to see
what it is.
>
--
Robert Heller -- Get the Deepwoods Software FireFox Toolbar!
Deepwoods Software -- Linux Installation and Administration http://www.deepsoft.com/ -- Web Hosting, with CGI and Database heller@deepsoft.com -- Contract Programming: C/C++, Tcl/Tk
sym53c896 secondary channel not active 
Linear Mode
