Dexter Filmore wrote:
Am Mittwoch, 16. Mai 2007 19:14 schrieb Gus Wirth:
Dexter Filmore wrote:
Got me an Asus K8N-E Deluxe:
2 sATA ports
4 sATA ports via Sil3114 chip
Gigabit ethernet via Hypertransport on nF3, phys. layer Marvell 88E1111
Thought "nF3, tad older, probably well supported" - yeah, well, so I
thought.
Can't set hda (on IDE) to DMA mode, "Operation not permitted".
Figured that GBE only works via "forcedeth" and then only brings 100MBit.
Can't even get that one working as etho, it's up and conf'ed but I can't
ping any host outside.
Does anyone know if this is a lost cause?
Are you using 32-bit Linux or 64-bit Linux? There are messages on the
Linux Kernel Mailing List indicating this is broken in 64-bit mode but
works in 32-bit mode.
Searching around a bit for "nforce3 ide dma linux" popped up this from
NVIDIA
<http://http.download.nvidia.com/XFree86/nforce/1.0-0306/KnownProblems.html
Which kernel version are you trying to run? What does the output of
dmesg show you?
Gus
Using a 32 bit 2.6.18, custom. dmesg output:
[ 0.000000] Linux version 2.6.18-dex ([EMAIL PROTECTED]) (gcc-Version 3.4.6) #2
Wed May 16 16:59:59 CEST 2007
[snip]
[ 36.591386] Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
[ 36.591424] ide: Assuming 33MHz system bus speed for PIO modes; override
with idebus=xx
[ 36.591594] Probing IDE interface ide0...
[ 37.006607] hda: SAMSUNG SP1604N, ATA DISK drive
[ 37.678251] Probing IDE interface ide1...
[ 38.542067] hdc: IDE-CD R/RW 24x12A, ATAPI CD/DVD-ROM drive
[ 39.213721] ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
[ 39.213793] ide1 at 0x170-0x177,0x376 on irq 15
[ 39.213918] hda: max request size: 512KiB
[ 39.218079] hda: 312581808 sectors (160041 MB) w/2048KiB Cache,
CHS=19457/255/63
[ 39.218385] hda: cache flushes supported
[ 39.218436] hda: hda1 hda2 hda4 < hda5 hda6 hda7 >
[ 39.272805] hdc: ATAPI 40X CD-ROM CD-R/RW drive, 8192kB Cache
[ 39.272978] Uniform CD-ROM driver Revision: 3.20
[snip]
Looks like the generic IDE driver is finding the drives first. And much
later on we see this:
[ 45.480109] NFORCE3-250: IDE controller at PCI slot 0000:00:08.0
[ 45.480217] NFORCE3-250: chipset revision 162
[ 45.480251] NFORCE3-250: not 100% native mode: will probe irqs later
[ 45.480288] NFORCE3-250: BIOS didn't set cable bits correctly. Enabling
workaround.
[ 45.480334] NFORCE3-250: 0000:00:08.0 (rev a2) UDMA133 controller
[ 45.480374] NFORCE3-250: port 0x01f0 already claimed by ide0
[ 45.480410] NFORCE3-250: port 0x0170 already claimed by ide1
[ 45.480444] NFORCE3-250: neither IDE port enabled (BIOS)
[snip]
Looks like there might be problem with the BIOS setting up the
controller chip properly. I'm also curious about the "cable bits"
message. Are you using 80-wire ATA ribbon cables or 40-wire? You need to
use 80-wire cables to get full ATA-6 specs (DMA-100/133).
Have you tried booting something like Knoppix 5.1 (has a 2.6.19 kernel)
to see if that works properly? Maybe you need a kernel upgrade. There
may also be a way to force the detection order for the IDE controller,
either in the init scripts or in the kernel itself.
Since you are compiling your own kernel, you might try including only
the NFORCE3-250 driver and leaving out the generic IDE driver and see
what happens.
Gus
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
