I seem to be having a problem with my new Super P6DBE motherboard and
linux 2.1.104 (and pre104, and 103).
First, the system
Super P6DBE (BX based)
2 x PII-233
64MB SDRAM
2 x WD 36400 UDMA
CPU0: Intel Pentium II (Klamath) stepping 04
CPU1: Intel Pentium II (Klamath) stepping 04
PIIX4: IDE controller on PCI bus 00 dev 39
PIIX4: not 100% native mode: will probe irqs later
ide0: BM-DMA at 0xffa0-0xffa7, BIOS settings: hda:DMA, hdb:pio
ide1: BM-DMA at 0xffa8-0xffaf, BIOS settings: hdc:DMA, hdd:pio
hda: WDC AC36400L, ATA DISK drive
hdc: WDC AC36400L, ATA DISK drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
hda: WDC AC36400L, 6149MB w/256kB Cache, CHS=13328/15/63, UDMA
hdc: WDC AC36400L, 6149MB w/256kB Cache, CHS=13328/15/63, UDMA
Okay, now the problem. If I enable DMA access (as it is by default) on
/dev/hdc, and then I do something that involves accessing both hda and
hdc, the machine with lock up SOLID in about 5 seconds. Everytime.
I tried to do some testing and haven't really found anyting. In my
BIOS, I have options for "TypeF DMA", which I played with a bit to no
avail, but since there are 9 combo's of that, +9 for another DMA, +8
other options affecting DMA, I simply couldn't try them all.
If I
hdparm -d0 /dev/hdc
everything is OK; albeit very slow (from 9MB/s to 3MB/s).
While doing a test, I was running a vmstat on another console and saw
something that looked odd to me: each second around 2000 blocks were
read in, 2000 blocks were read out and 10K+ interrupts were fired. It
seems the 10K+ ints are a bit high, but I really don't know.
I remember people talking (it seems a while ago) about lockups under
SMP kernels and IDE on 2 controllers, but looking through search
engines and back mail didn't turn up anything non-related to RAID.
Sometime today, I plan to put drive #2 on the same IDE0 as #1
(hdc->hdb) and see if that makes any difference. I expect that it
will.
And mayhaps someone could clarify this for me: at one point in Pentium
chipset history (440FX era), the second IDE interface was actually
on the ISA bus, not PCI. Is this still true for modern chipsets?
(LX,BX) ? If so, is that what this issue involves?
If anyone knows anything about this situation, I would be happy to
hear any suggestions.
Thank you in advance for your help! And as always, more info upon
request!
.............................................................................
..nicholas j. leon..........mrnick.binary9.net.........nicholas@binary9.net..
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu