My server had been running very well with all SATA drives on it. Recently I found some IDE drives laying around and added them to the server. The BIOS was able to see them but unRaid server won't.

 

I think this is the part of syslog where it seems relevant to this problem...

 

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: controller can't do NCQ, turning off CAP_NCQ

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: controller can't do PMP, turning off CAP_PMP

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: MV_AHCI HACK: port_map 7 -> 3

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: Disabling your PATA port. Use the boot option 'ahci.marvell_enable=0' to avoid this.

Oct  5 19:34:23 Tower kernel: ahci: SSS flag set, parallel bus scan disabled

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: AHCI 0001.0000 32 slots 3 ports 3 Gbps 0x3 impl IDE mode

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: flags: 64bit stag led slum part

Oct  5 19:34:23 Tower kernel: ahci 0000:02:00.0: setting latency timer to 64

 

Just for reference, my 4.5b6 server runs on ASUS P5Q-EM board with 6 SATA drives. Any thought, guys?

First thing to check is your BIOS settings, make sure they are correct, with DMA enabled, 32bit access, etc

I can see those two IDE drives on Slax or Knoppix. So I assume it's not BIOS problem...

From http://en.wikipedia.org/wiki/List_of_Marvell_Technology_Group_chipsets

 

88SE61xx Chipsets Linux support

[edit] Linux IDE(PATA) driver

 

Actually in the Kernel 2.6.28-rc5 (15 November 2008), driver version 0.1.6 (it was the same in driver version 0.1.4 with kernel 2.6.25-rc8 the 8th April 2008), the module pata_marvell support only these chipsets :

 

    * Marvell 88SE6101 [1]

    * Marvell 88SE6121

    * Marvell 88SE6123

    * Marvell 88SE6145 [2]

 

The file pata_marvell.c in kernel sources add IDE support for the Marvell IDE Chipsets. You can browse this file here by Web Git : [linux/kernel/git/torvalds/linux-2.6.git] / drivers / ata / pata_marvell.c source : http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/ata/pata_marvell.c;h=76e399bf8c1bc2adfe9587f1ffd17cae192486f2;hb=9bf1a2445f3c569098b8de7097ca324e65abecc2

 

Code for chipset supported list (line 159 to 165):

 

static const struct pci_device_id marvell_pci_tbl[] = {

        { PCI_DEVICE(0x11AB, 0x6101), },

        { PCI_DEVICE(0x11AB, 0x6121), },

        { PCI_DEVICE(0x11AB, 0x6123), },

        { PCI_DEVICE(0x11AB, 0x6145), },

        { }    /* terminate list */

};

 

A bug ticket is open at Kernel Bug Tracker : Problem with pata_marvell kernel module : kernel 2.6.24.4 in Mandriva 2008.1 RC2 don't detect my IDE harddisk connected to a Marvell 88SE6111 IDE Controller

 

Workaround for linux and Marvell's chips that are not in the pata_marvell driver : [3] [4]

 

    * Ensure the BIOS is set to AHCI. Please go to Advanced > Drive configuration > Configure SATA as. (In later Intel® desktop boards you can enter the BIOS setup program by repeatedly pressing the key during the boot process)

    * Use the boot parameter: all-generic-ide

    * use the boot parameter: pci=nommconf

 

This workaround work only with chipset that have AHCI support. But ICH9, ICH8, ICH7 and ICH6-based chipsets, as well as ICH5 and ICH5R-based chipsets, do not use AHCI.[5]

[edit] Linux IDE(PATA) driver problem workaround for Marvell 88SE6111, 88SE6121, 88SE6145

 

Alan Cox said [6][7] that there is regressions in PATA Marvell Linux driver since March 2008 for the 6121 Marvell, and late 2007 for the 6145. The regressions are in kernel 2.6.27 and maybe 2.6.25 et 2.6.26. Kernel 2.6.24.4 works.

 

He said too : "The actual fix for the moment is very simple. If the user has included the pata_marvell driver let it drive the ports. If they've only selected for SATA support give them the AHCI driver which will run the port a fraction faster. Allow the user to control this decision via ahci.marvell_enable as a module parameter so that distributions can ship 'it works' defaults and smarter users (or config tools) can then flip it over it desired."

 

    * With PATA/IDE devices

 

    *

          o With kernel 2.6.25, 2.6.26, 2.6.27

 

Alan COX said : "You need to load pata_marvell before AHCI on current kernels (this is fixed in 2.6.28rc) for PATA support" [8]

 

    *

          o With kernel 2.6.27 and this patch , 2.6.28-rc, 2.6.28 or upper

 

By default ahci.marvell_enable=1 that is disabling your PATA port in the PATA Marvell driver. Use the boot option 'ahci.marvell_enable=0' to avoid this.

 

Use the boot parameter : ahci.marvell_enable=0 if you have PATA/IDE devices connected at you Marvell chip

or this :

added this line in the file /etc/modprobe.d/options :

 

options ahci.marvell_enable=0

 

and run update-initramfs -u

shutdown

enable Marvell

and see if it works for you

 

So I added this ahci.marvell_enable option to modprobe.d and repacked bzroot. The change did not help. Anyone had this problem before?

still cant resolve the problem. can it be linux driver incompatibility?

still cant resolve the problem. can it be linux driver incompatibility?

 

Send an email of to Tom (limetech) and get his input.  He may be able to include a newer version of the driver for you in the next beta.  Also, if you are not currently running the 4.5beta7 version try updating to that.

Some BIOSes have a setting for IDE that switches b/n "Legacy Mode" and "Native Mode".

Turning that to "Legacy Mode" has helped me see the IDE drives on some machines.