Re: Promise SATAII 150 TX2plus (was: 150 TX4)

Hi,

I recently bought a Promise SATAII150 TX2plus and 2 x 160 GB Hitachi disks - 
and the system recognize the ctrl. but not my disks.

My system is (now) 6.0-CURRENT from CVS this Saturday (Dec. 11th 2004) as my 
4.10-p4 sure didn't have support for this, nor did 5.3-STABLE have the chip 
ID added for this specific Promise controller (0x3d75105a).

I don't know how to save a dmesg when my boot stalls (no panic) but this is 
when I boot with the ctrl. installed and no disks attached:

atapci1: <Promise PDC20575 SATA150 controller> port 
0x7800-0x78ff,0x7400-0x747f mem 0xea800000-0xea81ffff,0xea822000-0xea822fff 
irq 10 at device 13.0 on pci0
atapci1: failed: rid 0x20 is memory, requested 4

I have tried with and w/o APIC enabled in my BIOS and I've also tried 
setting hint.apic.0.disabled="1" in /boot/loader.conf. None of this seems to 
make a significant effect. Also - I cannot see any difference when setting 
that loader.conf variable. It's like my BIOS will override this nomatter 
what I set. (I just located a floppy drive - so I will also try and upgrade 
the BIOS on this Aopen motherboard)

When I do attach my SATA drives I get a lot of this on my console:

ata1-master: ATA_IDENTIFY: timeout
ata1-slave: ATA_IDENTIFY: timeout
ata2-master: ATA_IDENTIFY: timeout
ata2-slave: ATA_IDENTIFY: timeout
ata3-master: ATA_IDENTIFY: timeout
ata3-slave: ATA_IDENTIFY: timeout

And then my boot just stalls.

One time I managed to get _one_ drive identified, but then it paniced when 
trying to set the DMA_MODE or something in that neighbourhood. Since this 
was no success I didn't think about saving output (nor do I remember how to) 
so I just rebooted and tried yet another setting.

Any hints appriciated. I have of course kept all the debugging options in my 
KERNCONF, but I'm not sure what is needed in order to help myself or others 
to solve this...

Thanks in advance.

// Nicolai

PS: I also tried importing the ata driver from 6.x into 5.x before upgrading 
to 6.x - but same timeouts occured (I added the chip ID manually). But this 
is not something I feel like doing a lot since this PII-450 takes hours to 
compile just a kernel :)


----- Original Message ----- 
From: "Kenneth Vestergaard Schmidt" <kvs_at_binarysolutions.dk>
To: <freebsd-current_at_freebsd.org>
Sent: Friday, October 15, 2004 10:36 AM
Subject: Re: Promise SATAII 150 TX4


> On Thu, Oct 14, 2004 at 07:28:27PM +0200, Søren Schmidt wrote:
>> >Adding their PCI id's to the existing driver didn't seem to be enough.
>> >
>> >Is support for this controller in the works?
>>
>> Not yet at least, but if you are in a hurry you could ship me one and
>> I'll look into it :)
>
> We might just do that. One quick question before, though:
>
> Adding the PCI ID's does get the card recognized, but it doesn't find
> any drives:
>
>  atapci0: <Promise PDC40518 SATA150 controller> port 
> 0xc400-0xc4ff,0xc000-0xc07f mem 
> 0xe9000000-0xe901ffff,0xe9020000-0xe9020fff irq 11 at device 9.0 on pci1
>  atapci0: failed: rid 0x20 is memory, requested 4
>  atapci0: Reserved 0x20000 bytes for rid 0x20 type 3 at 0xe9000000
>  atapci0: Reserved 0x1000 bytes for rid 0x1c type 3 at 0xe9020000
>  atapci0: [MPSAFE]
>  ata2: channel #0 on atapci0
>  ata2: reset tp1 mask=00 ostat0=ff ostat1=ff
>  ata2: [MPSAFE]
>
> The last three lines are repeated for ata3-5 as well.
>
> Later, the following shows up:
>
>  Interrupt storm detected on "irq11: atapci0"; throttling interrupt source
>  ata0-slave: pio=0x0c wdma=0x22 udma=0x46 cable=80pin
>  ata0-slave: setting PIO4 on Intel ICH4 chip
>  ata0-slave: setting UDMA100 on Intel ICH4 chip
>  ad1: <MAXTOR 6L040L2/A93.0500> ATA-5 disk at ata0-slave
>  ad1: 38172MB (78177792 sectors), 77557 C, 16 H, 63 S, 512 B
>  ad1: 16 secs/int, 1 depth queue, UDMA100
>
> (ata0-slave/ad1 belongs to atapci1, which is just an onboard ICH4)
>
> Does this ring a bell and inspire a quick fix, or do you need the hardware
> to fix this?
>
> Thanks in advance.
> _______________________________________________
> freebsd-current_at_freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
>