AEC62XX Driver: Can ANYONE Help!?

[Corbett Battaile]

Thanks for the info, but it didn't seem to do the trick.

I passed "ide=nodma" to the kernel, and dmesg reports that DMA is indeed
disabled ("ide_setup: ide=nodmaIDE: Prevented DMA") but I get the same
behavior and same dmesg output when probing the AEC-6280M card.

I also tried adjusting the DMA setting and other parameters with hdparm, 
but that didn't help either.

Oh well, more tinkering I suppose. I plan to try the ACard in a different 
machine (G4/GigE) and also plan to try a Promise-based card in the G3/BW. 
Failing all that, I'll probably have to resort to attaching the IDE drives 
to the internal ATA/33 and booting off a SCSI card.

Thanks again.


On Wed, 17 Nov 2004, Daniel Gimpelevich wrote:
>
> Have you tried turning off DMA for /dev/hdg and /dev/hdi yet? Instructions
> on how to do so may be found at
> http://lists.terrasoftsolutions.com/pipermail/yellowdog-newbie/2003-August/003998.html
> 
> On Wed, 17 Nov 2004 16:38:04 -0800, ccb wrote:
>>
>> Sorry for the second posting, but in hindsight perhaps some additional detail
>> would be useful.
>>
>> The ACard AEC-6280M cards are installed in two Blue and White G3s, one running
>> YDL 3.0.1 and the other 4.0. The CPUs are upgraded with PowerLogix PowerForce
>> ZIF 900MHZ G3s. There's a Broadcom BCM5703X Gigabit ethernet card (pulled from
>> Apple XServes), a Texas Instruments TSB12LV23 based Firewire card, and an ATI
>> Rage 128 card in each. Each AEC-6280M has a Western Digital WD2500SB 250GB drive
>> on each IDE channel (as you can see below).
>> 
>> 
>> Here's part of the pertinent output from dmesg:
>> 
>> AEC6280R: IDE controller at PCI slot 0000:01:03.0
>> AEC6280R: chipset revision 7
>> AEC6280R: ROM enabled at 0x808a0000
>> AEC6280R: 100% native mode on irq 24
>>     ide3: BM-DMA at 0x1050-0x1057, BIOS settings: hdg:pio, hdh:pio
>>     ide4: BM-DMA at 0x1058-0x105f, BIOS settings: hdi:pio, hdj:pio
>> Probing IDE interface ide3...
>> ieee1394: Host added: ID:BUS[0-00:1023]  GUID[0000303c00090d9d]
>> hdg: WDC WD2500SB-01KBA0, ATA DISK drive
>> ide3 at 0x1090-0x1097,0x1082 on irq 24
>> hdg: max request size: 1024KiB
>> hdg: lost interrupt
>> hdg: lost interrupt
>> hdg: lost interrupt
>> hdg: 488397168 sectors (250059 MB) w/8192KiB Cache, CHS=30401/255/63, UDMA(100)
>> hdg: lost interrupt
>>  hdg:<4>hdg: dma_timer_expiry: dma status == 0x24
>>  AEC62XX time out <4>hdg: lost interrupt
>>  [mac] hdg1 hdg2 hdg3 hdg4 hdg5 hdg6 hdg7<4>hdg: dma_timer_expiry: dma status ==
>> 0x24
>>  AEC62XX time out <4>hdg: lost interrupt
>>  hdg8 hdg9 hdg10
>> hdg: lost interrupt
>> Probing IDE interface ide4...
>> hdi: WDC WD2500SB-01KBA0, ATA DISK drive
>> ide4 at 0x1070-0x1077,0x1062 on irq 24
>> hdi: max request size: 1024KiB
>> hdi: lost interrupt
>> hdi: lost interrupt
>> hdi: lost interrupt
>> hdi: 488397168 sectors (250059 MB) w/8192KiB Cache, CHS=30401/255/63, UDMA(100)
>> hdi: lost interrupt
>>  hdi:<4>hdi: dma_timer_expiry: dma status == 0x24
>>  AEC62XX time out <4>hdi: lost interrupt
>>  [mac] hdi1 hdi2 hdi3 hdi4 hdi5 hdi6 hdi7<4>hdi: dma_timer_expiry: dma status ==
>> 0x24
>>  AEC62XX time out <4>hdi: lost interrupt
>>  hdi8 hdi9 hdi10
>> hdi: lost interrupt
>> 
>> 
>> Thanks again.
>> 
>> 
>> Quoting myself:
>>>
>>> Hello all. After purchasing two ACard AEC-6280M cards for a couple of file 
>>> servers, I came to realize that the recent versions of the aec62xx driver 
>>> are flaky at best. (Or at least don't generally function properly "out of 
>>> the box.") I find it hard to believe that no one out there is using an 
>>> ACard-based IDE PCI card with YDL 3 or 4. Past discussions that I've found 
>>> on this topic haven't produced much in the way of solutions.
>>> 
>>> Does anyone have a working aec62xx driver, or know of a kernel source 
>>> distro that contains one, or know of a workaround?
>>> 
>>> Thanks.
  
--
Corbett Battaile <ccb at ydl.net>