biggest, newest IDE disk stuck in PIO mode; because no BIOS support? [dsb]
I've been having trouble using hdparm to speed up one of my disks.
I suspect it's because I had to disable the disk in my BIOS so the
BIOS wouldn't hang when booting.
Has anyone seen this problem or does anyone know how to fix it?
Would the "-ide" version of the 2.2.19 kernel packge help with this?
ROUGH DETAILS:
Initially, my three disks had roughly the same low performance.
Using hdparm, I got a speedup of about 4x for /dev/.hda and 2x for
/dev/hdc, but /dev/hdd didn't change.
When I installed my /dev/hdd disk, I had to disable it my BIOS
because it was too large (45GB). (The BIOS would hang when trying
to auto-detect it.)
I notice that the kernel boot message include the lines:
ide0: BM-DMA at 0xf000-0xf007, BIOS settings: hda:DMA, hdb:DMA
ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:pio
Note how hdd says "pio" while all others say "DMA".
I suspect that disabling the drive in the BIOS somehow caused Linux
to leave or put the disk in PIO-only mode, and I don't know what
to do to fix it.
MORE DETAILS:
Using hdparm, I found that I was getting around 2 MB/s:
hda: ~2.2 MB/s
hdc: ~2.7 MB/s
hdd: ~2.1 MB/s
When I enabled DMA mode ("hdparm -d1 ..."), hda quadrupled its speed
and hdc doubled its speed, but hdd changed only slightly:
hda: ~9.0 MB/s
hdc: ~5.5 MB/s
hdd: ~2.2 MB/s
I tried several other hdparm options, but nothing seemed to make
much of a difference.
When I looked at the kernel boot log messages, I notice that all the
ATA/IDE devices other than hdd were enabled for DMA mode, but hdd had
only PIO mode. Here are the seemingly relevant lines from dmesg:
ALI15X3: IDE controller on PCI bus 00 dev 78
ALI15X3: not 100% native mode: will probe irqs later
ide0: BM-DMA at 0xf000-0xf007, BIOS settings: hda:DMA, hdb:DMA
ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:pio
hda: Maxtor 90840D5, ATA DISK drive
hdb: TOSHIBA CD-ROM XM-6302B, ATAPI CDROM drive
hdc: Maxtor 91024U4, ATA DISK drive
hdd: Maxtor 54610H6, ATA DISK drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
hda: Maxtor 90840D5, 8047MB w/256kB Cache, CHS=1025/255/63
hdc: Maxtor 91024U4, 9765MB w/2048kB Cache, CHS=19841/16/63
hdd: Maxtor 54610H6, 43967MB w/2048kB Cache, CHS=5605/255/63
hdb: ATAPI 32X CD-ROM drive, 256kB Cache
...
Partition check:
hda: hda1 hda2 hda3 hda4 < hda5 hda6 hda7 hda8 hda9 hda10 >
hdc: [PTBL] [1244/255/63] hdc1 hdc2 hdc3 hdc4
hdd: hdd1 hdd2
That "pio" vs. "DMA" difference for hdd correlates with the fact that
I disabled that disk in the BIOS.
The BIOS would hang when auto-detecting the 45GB disk unless I disabled
the disk entirely or jumpered the disk to limit itself to 33GB. (The BIOS
would hang during auto-detection at boot time if the disk type was set to
"Auto". The BIOS would also hang in the BIOS setup screen for doing
auto-detection.)
With the disk disabled in the BIOS, both Linux and Windows 98 seemed able
to access it fine. (I don't have any information on its performance under
Windows 98.)
Machine Details:
Hardware:
Motherboard: GigaByte GA-5AX Rev. 3.0
BIOS: Award Modular BIOS v4.51PG
GA-5AC REV 1.5
06/17/1999-ALADDIN5-2A5KKG09C-00
Award Plug and Play BIOS Extension v1.0A
ATA/IDE controller: onboard controller (probably just ATA-33)
CPU: AMD K6-2 300MHz (yeah, I know; quit laughing)
ATA devices: primary master: Maxtor 90840D5 - 8.4 GB, ATA-?
primary slave: Toshiba XM-6302B CD-ROM
secondary master: Maxtor 91024U4 - 10.2 GB, ATA-66
secondary slave: Maxtor 54610H4 - 46 GB, ATA-100
(hdc and hdd are on what seems to be an 80-conductor
IDE ribbon cable)
OS:
Debian potato
kernel-image-2.2.19 (version 2.2.19_4potato.5)
Any ideas how to get Linux and my 45GB disk to use DMA mode (besides
upgrading my aging machine, of course)?
Thanks,
Daniel
--
Daniel Barclay
dsb@smart.net
Reply to: