Promise 20267 can't recognize IDE device or freezes

[Benjamin Herrenschmidt]

On Tue, 2002-11-26 at 23:29, Nicholas Riley wrote:
>
> Hi,
>
> I have an ATAPI Seagate (Conner) Travan tape drive which I recently
> removed from my Power Mac G4 running OS X.  It worked fine with
> Retrospect on the internal ATA bus and through a FireWire bridge.
>
> I want to use the drive for backups on my Power Mac 9500 running
> Linux.  I bought a Promise Ultra100 card as it appeared to be
> supported under PowerPC Linux, and installed the card and tape drive
> into the machine.  For obvious reasons (no driver support) I can't
> test whether it works under Mac OS.
>
> I was using kernel 2.4.18-ben0 compiled in March, with IDE as a module
> I compiled yesterday.  According to what I read either ide-scsi or
> ide-tape should work, but I couldn't get the device to show up.
>
> The output I saw was exactly as this message discusses:
>
> <http://lists.linuxppc.org/linuxppc-dev/200105/msg00130.html>
>
> Here is the dmesg output:
>
> Uniform Multi-Platform E-IDE driver Revision: 6.31
> ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
> PDC20267: IDE controller on PCI bus 00 dev 70
> PCI: Enabling device 00:0e.0 (0004 -> 0007)
> PDC20267: chipset revision 2
> PDC20267: not 100% native mode: will probe irqs later
> PDC20267: ROM enabled at 0x80820000
> PDC20267: (U)DMA Burst Bit DISABLED Primary PCI Mode Secondary PCI Mode.
>     ide2: BM-DMA at 0x0400-0x0407, BIOS settings: hde:pio, hdf:DMA
>     ide3: BM-DMA at 0x0408-0x040f, BIOS settings: hdg:pio, hdh:pio
>
> In the following thread:
>
> <http://lists.linuxppc.org/linuxppc-user/199907/msg01838.html>
>
> when someone had the same problem in a 2.2 kernel, they were able to
> fix it by building with ide-pmac even though internal IDE is not used.
>
> I updated to the latest 2.4.20-rc3-ben0 from yesterday and tried
> compiling with ide-pmac and the other options recommended by Andre
> Hedrick in his message in that thread (CONFIG_BLK_DEV_OFFBOARD).  Now
> I get the following on boot:
>
> Uniform Multi-Platform E-IDE driver Revision: 6.31
> ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
> PDC20267: IDE controller on PCI bus 00 dev 70
> PDC20267: chipset revision 2
> PDC20267: not 100% native mode: will probe irqs later
> PDC20267: ROM enabled at 0x80820000
> PDC20267: (U)DMA Burst Bit DISABLED Primary PCI Mode Secondary PCI Mode.
> 	ide0: BM-DMA at 0x0400-0x0407, BIOS settings: hda:pio, hdb:DMA
> 	ide1: BM-DMA at 0x0408-0x040f, BIOS settings: hdc:pio, pdd:pio
> Probing IDE interface ide0...
> disabling irq 24 defensively
> hdb: CONNER CTT8000-A, ATAPI TAPE drive
> Probing IDE interface ide1...
> [freeze]
>
> If I disable IDE completely, then the 2.4.20-pre3-ben0 kernel usually
> boots correctly (sporadically it doesn't, but much earlier than the
> IDE kernel dies, so I don't think it's related).
>
> The kernel configuration that works for me is here:
>
> <http://web.sabi.net/promise/config-no-ide-boots.txt>
>
> and the configuration which failed here:
>
> <http://web.sabi.net/promise/config-ide.txt>
>
> Any suggestions on how to troubleshoot (or fix) this problem would be
> much appreciated.  Do I need another kernel, different options,
> another IDE driver?

Promise controllers are giving me headaches everyday...

Currently, I have the 20270 of the Xserve working properly, the
driver in my tree is a mix of the old driver and the one in alan's
kernel (latest from Andre). But it still seem to not be quite right
with all controllers.

Do you have a way to prevent probing of ide1 ? (ide1=noprobe maybe)

Ben.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/