DMA troubles

Q-collective · 11-05-2004, 07:12 AM

Ok, let's cut the intro and go to the problem, shall we? :P
I have 2 hdd's, both pretty new (200 GB, 8 MB cache, 7200rpm and 120 GB, 2 MB cache, 7200rpm), so you would say they have a working dma... Well, one of them has

Code:

Futility linux # hdparm -m16 -u1 -c1 -d1 -X udma6 /dev/hda

/dev/hda:
 setting 32-bit IO_support flag to 1
 setting multcount to 16
 setting unmaskirq to 1 (on)
 setting using_dma to 1 (on)
 setting xfermode to 70 (UltraDMA mode6)
 multcount    = 16 (on)
 IO_support   =  1 (32-bit)
 unmaskirq    =  1 (on)
 using_dma    =  1 (on)
Futility linux # hdparm -i /dev/hda

/dev/hda:

 Model=Maxtor 6B200P0, FwRev=BAH41B10, SerialNo=B408ERHH
 Config={ Fixed }
 RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57
 BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16
 CurCHS=4047/16/255, CurSects=16511760, LBA=yes, LBAsects=268435455
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes:  pio0 pio1 pio2 pio3 pio4
 DMA modes:  mdma0 mdma1 mdma2
 UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 udma6
 AdvancedPM=yes: disabled (255) WriteCache=enabled
 Drive conforms to: (null):

 * signifies the current active mode

Futility linux # hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   964 MB in  2.00 seconds = 481.11 MB/sec
 Timing buffered disk reads:  176 MB in  3.05 seconds =  57.69 MB/sec

But the other (120 GB one) still resides at very low speeds:

Code:

Futility linux # hdparm -m16 -u1 -c1 -d1 -X udma5 /dev/hdb

/dev/hdb:
 setting 32-bit IO_support flag to 1
 setting multcount to 16
 setting unmaskirq to 1 (on)
 setting using_dma to 1 (on)
 setting xfermode to 69 (UltraDMA mode5)
 multcount    = 16 (on)
 IO_support   =  1 (32-bit)
 unmaskirq    =  1 (on)
 using_dma    =  1 (on)
Futility linux # hdparm -i /dev/hdb

/dev/hdb:

 Model=HDS722512VLAT20, FwRev=V33OA60A, SerialNo=VNR3GEC3C41LGK
 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
 RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=52
 BuffType=DualPortCache, BuffSize=1794kB, MaxMultSect=16, MultSect=16
 CurCHS=4047/16/255, CurSects=16511760, LBA=yes, LBAsects=241254720
 IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes:  pio0 pio1 pio2 pio3 pio4
 DMA modes:  mdma0 mdma1 mdma2
 UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5
 AdvancedPM=yes: disabled (255) WriteCache=enabled
 Drive conforms to: ATA/ATAPI-6 T13 1410D revision 3a:

 * signifies the current active mode

Futility linux # hdparm -tT /dev/hdb

/dev/hdb:
 Timing cached reads:   772 MB in  2.01 seconds = 384.52 MB/sec
 Timing buffered disk reads:   20 MB in  3.30 seconds =   6.06 MB/sec

Am I the only one who thinks this is strange?

Who can help me out here?

JaseP · 11-05-2004, 09:51 AM

I'm assuming that you have a CD/DVD-ROM/RW drive on this machine, right???

If the HDs occupy seperate lines, and one of them shares the IDE line for the CD/DVD drive, then it will default to the slower DMA speed of the CD/DVD drive unit (if DMA is even enabled for that device, many distros will turn of CD/DVD DMA by default for compatibility purposes).

If both HDs are on the same line, then try and make sure you have used hdparm to set it correctly,...

Q-collective · 11-05-2004, 10:45 AM

As you may have noticed, I did my fair share of hdparm-fucking around and I'm now through my hdparm knowledge, any suggestions I haven't tried yet?
And yes, hda and hdb share one IDE cable, sorry that I didn't mentioned that before

Crashed_Again · 11-05-2004, 03:03 PM

I had a similar problem and the issue was the jumper settings on the hard drives. I had cable select on both of them. Setting one to Master and the other to Slave fixed it for me. Not sure if this helps you at all.

Q-collective · 11-05-2004, 03:14 PM

Thanks for your reply
I already have both disks defined as master/slave with hda as being master.
The slave mode could lower the speed? (I doubt it, but hey

)

JaseP · 11-05-2004, 03:48 PM

Slave mode will not lower the speed generally,...

The BIOS simply generates a message telling the HD that it wants to talk to either the Master or the Slave, and the appropriate drive responds. In Cable Select, the drives respond based upon their physical location on the cable. What kind of cabling are you using??? That MIGHT have and effect on the situation. My best guess is that this is still a hardware problem, and has nothing to do with your distro... What are the drive manufacturers??? Looks like one is a Maxtor, the other I can't tell based upon its ID.

Q-collective · 11-05-2004, 04:45 PM

Quote:

Originally posted by JaseP
What kind of cabling are you using???

Just a regular flat ata cable really

Quote:

What are the drive manufacturers??? Looks like one is a Maxtor, the other I can't tell based upon its ID.

The other is an IBM 120 GXP Deskstar (Hitachi these days)