Dell PERC 2/SC (AMI Megaraid 466(?)) not working

Bug #32752 reported by Alex Mauer
28
Affects Status Importance Assigned to Milestone
debian-installer (Ubuntu)
Invalid
Medium
Unassigned
linux-source-2.6.12 (Ubuntu)
Invalid
Medium
Unassigned
linux-source-2.6.15 (Ubuntu)
Fix Released
Medium
Ben Collins

Bug Description

The installer fails to detect the Dell PERC 2/SC raid controller card.

This is on a Poweredge 4300, with Dapper Flight 4.

Manually loading the megraid* drivers also doesn't find the device.

It shows up in the output of lspci as "Intel Corporation 80960RP [i960RP Microprocessor] (rev 03)" when configured with "Emulation = Mass Storage", and
"PCI Bridge: Intel Corporation 80960RP [i960 RP Microprocessor/Bridge] (rev 03)
I2O: Intel Corporation 80960RP [i960 RP Microprocessor] (rev 03)" when configured with "Emulation = I2O"

The mptscsih driver mentioned in bug 23559 also does not work.

For comparison, the debian "sarge" installer also didn't detect the card but it did find/activate it when the megraid driver was loaded manually. (kernel 2.4)

Some googling indicated that this may be a problem with that driver in kernel 2.4 vs. kernel 2.6.

Revision history for this message
Chris Danis (cgd3) wrote :

This is a kernel bug, not necessarily an installer one. The legacy megaraid driver in the kernel that ships with BreezyBadger does not recognize the PERC 2/SC card. This is true even of the latest kernel (2.6.12-10.28)

A 686-smp kernel from HoaryHedgehog (2.6.10-34.11) recognizes the card correctly. Perhaps another error that occurred during the split between the new megaraid and the legacy megaraid driver?

lspci -vvv reports for this card:
0000:00:08.1 I2O: Intel Corp. 80960RP [i960RP Microprocessor] (rev 03) (prog-if 01)
        Subsystem: Dell PowerEdge Expandable RAID Controller 2/SC
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR+ FastB2B-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
        Latency: 32, Cache Line Size: 0x08 (32 bytes)
        Interrupt: pin A routed to IRQ 11
        Region 0: Memory at fc000000 (32-bit, prefetchable) [size=4M]
        Expansion ROM at fcc00000 [disabled] [size=32K]

and lspci -n
0000:00:08.1 0e00: 8086:1960 (rev 03)

-chris

Revision history for this message
Alex Mauer (hawke) wrote :

For me, lspci -vv:

0000:02:0a.0 PCI bridge: Intel Corporation 80960RP [i960RP Microprocessor/Bridge] (rev 03)
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B-
        Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <Mabort- >SERR- <PERR-
        Latency: 32, Cache Line Size: 0x08 (32 bytes)
        Bus: primary=02, secondary=03, subordinate=03, sec-latency=32
        BridgeCtl: Parity- SERR+ NoISA+ VGA- MAbort- >Reset- FastB2B-

0000:02:0a.1 I2O: Intel Corporation 80960RP [i960RP Microprocessor] (rev 03) (prog-if 01)
        Subsystem: Dell: Unknown device 1111
        Control: I/O- Mem+ BusMaster- SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR+ FastB2B-
        Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <Mabort- >SERR- <PERR-
        Interrupt: pin A routed to IRQ 185
        Region 0: Memory at f2c00000 (32-bit, prefetchable) [size=4M]
        Expansion ROM at f1000000 [disabled] [size=32K]

Note that this is copied by hand, so errors are quite possible.

Revision history for this message
Alex Mauer (hawke) wrote :

Also, as expected this is not fixed in Dapper Flight 5.

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

this is not an installer bug

Changed in debian-installer:
status: Unconfirmed → Rejected
Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Ben we want to backport the megaraid/megaraid_legacy from .16 where they did a similar but not the same kind of split and fix this problem.

Fabio

Changed in linux-source-2.6.15:
status: Unconfirmed → Confirmed
Revision history for this message
spiderwort (ironweed) wrote :

Not fixed in Dapper Flight 6, nor in Dapper Beta.

lspci -vv for my machine reports as follows:

0000:00:06.0 PCI bridge: Intel Corporation 80960RP [i960 RP Microprocessor/Bridge] (rev 03) (prog-if 00 [Normal decode])
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
        Latency: 32, Cache Line Size: 0x08 (32 bytes)
        Bus: primary=00, secondary=02, subordinate=02, sec-latency=32
        BridgeCtl: Parity+ SERR+ NoISA+ VGA- MAbort- >Reset- FastB2B-

0000:00:06.1 I2O: Intel Corporation 80960RP [i960RP Microprocessor] (rev 03) (prog-if 01)
        Subsystem: American Megatrends Inc. MegaRAID 438 Ultra2 LVD RAID Controller
        Control: I/O- Mem+ BusMaster- SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR- FastB2B-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
        Interrupt: pin A routed to IRQ 10
        Region 0: Memory at dd800000 (32-bit, prefetchable) [size=4M]
        Expansion ROM at dfff0000 [disabled] [size=32K]

During boot, this controller reports itself as:
MegaRAID Ultra2/5 PCI version 1.48
American Megatrends

Ubuntu stopped seeing my hard drives when I upgraded from Hoary to Breezy. I have been running on the old Hoary kernel for the last 6 months.

Revision history for this message
Antons Rebguns (arebgun) wrote :

Have the same problem with Dell PowerEdge 1300 and Perc2/SC RAID controller. Neither Breezy nor Dapper can see the RAID array during installation. However it it woring fine with Hoary with 2.6.10 kernel as well as custom compiled 2.6.16 kernel. I can upgrade from Horay to Breezy, but no luck with fresh install.

Revision history for this message
potsofdirt (potsofdirt) wrote :

I am having similar problems with an LSI Megaraid i4 controller (which uses the old megaraid module and not the new megaraid_mbox module). Installation does not detect the drive with Dapper Beta 1. With the live cd I tried to manually load the module (via modprobe) and I still can not see the drive in /dev. If there is any output from my system that would help, please let me know.

Revision history for this message
Mario A. Vazquez (darkcoder) wrote :

The problem with the megaraid driver is that it must be loaded prior to any other scsi driver, like aic7xxx. If is loaded after, then it will not work.

The problem is that I cannot find a way to boot the CD (install or live) without probing devices.

Revision history for this message
Mario A. Vazquez (darkcoder) wrote :

tryied with Debian 3.1-r2 net install CD, and it even do not come with any megaraid family driver at all.

Test with gentoo and yes it is a kernel problem. I'll resubmit this bug as a kernel one.

Revision history for this message
Mario A. Vazquez (darkcoder) wrote :

I was able to install another Linux distribution using Kernel 2.6.16.x on the machine, and here are part of the kernel config related to the Megaraid area, so you can give it a look.

    CONFIG_MEGARAID_NEWGEN=y
    CONFIG_MEGARAID_MM=m
    CONFIG_MEGARAID_MAILBOX=m
    CONFIG_MEGARAID_LEGACY=y
    CONFIG_MEGARAID_SAS=m

Revision history for this message
potsofdirt (potsofdirt) wrote :

I was able to get this card working on Gentoo with 2.6.16. 2.6.15 did not work, and my discussions with a Gentoo dev indicate there was some work on 2.6.16 that introduced detection (with the newgen driver) for the LSI Logic Megaraid i4. I don't know if the patch to 2.6.16 was strictly a Gentoo-sources patch, but I can certainly provide lspci output if that would help.

Revision history for this message
JulFX (julfx) wrote :

Same problem with a LSI Megaraid i4 : the card is not detected when installing.
Also i had no problem with ubuntu hoary, the support seems to be broken since breezy. None of the flight and beta of dapper work correctly with this card.

Revision history for this message
Chuck Short (zulcss) wrote :

Fabbione can you get me the patch and ill see what i can do

Revision history for this message
Maciej Biłas (maciejb) wrote :

http://lkml.org/lkml/2006/2/9/228
This is a 2.6.16 patch for the megaraid driver.
I hope it solves the problem.

Revision history for this message
Ben Collins (ben-collins) wrote : Re: [Bug 32752] Re: Dell PERC 2/SC (AMI Megaraid 466(?)) not working

On Tue, 2006-07-04 at 03:49 +0000, lvizard wrote:
> http://lkml.org/lkml/2006/2/9/228
> This is a 2.6.16 patch for the megaraid driver.
> I hope it solves the problem.

No, we already have the naming change in our code. I believe the problem
is just one of PCI ID's in the module device table. I'm inspecting the
breezy/dapper/edgy tree to see what the correct device table should look
like.

Changed in linux-source-2.6.15:
assignee: nobody → ben-collins
status: Confirmed → Fix Committed
Changed in linux-source-2.6.12:
status: Unconfirmed → Confirmed
Revision history for this message
Alex Mauer (hawke) wrote :

This works for me with the latest Dapper kernel (2.6.15-26-server) but not when booting from the 6.06.1 installer CD. Is this expected to work in both cases, or did the fix not make it into 6.06.1?

Revision history for this message
David Sedeño Fernandez (david-alderia) wrote :

Is this bug fixed in Edgy?

I'm trying to install 6.06.1 and it doesn't work.

Thanks

Revision history for this message
Alex Mauer (hawke) wrote :

It's fixed in the Dapper updates. Installing Breezy and upgrading to Dapper might work. Otherwise it's probably possible to create a custom dapper CD with the new kernel & initrd.

Hopefully there'll be a 6.06.2 including this fix...

Revision history for this message
Maciej Biłas (maciejb) wrote :

As far as I remember both the kernel in Breezy and in Dapper didn't work for me, I had to install Hoary and then upgrade to Dapper keeping the old kernel and upgrading to the fixed version in a month or two when it showed up.
However we may have a bit different hardware... try yourself.

Changed in linux-source-2.6.12:
status: Confirmed → Rejected
Revision history for this message
Marco Rodrigues (gothicx) wrote :

Isn't this one already fix released ?

Changed in linux-source-2.6.15:
status: Fix Committed → Fix Released
Revision history for this message
Tom Dickson (8-launchpad-bombcar-com) wrote :

I had a similar issue with 7.10 (installed correctly, wouldn't boot) until I followed RedHat's recommendation and turned off I2O emulation:

http://www.redhat.com/docs/manuals/enterprise/RHEL-5-manual/release-notes/RELEASE-NOTES-x86-en.html

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.