dma problems with 2.6 kernel and vt8233A southbridge

From: Duane Evenson (duane4_at_shaw.ca)
Date: 08/23/05

  • Next message: John Thompson: "Re: ultraportable?"
    Date: Tue, 23 Aug 2005 02:24:51 GMT
    
    

    I have a problem activating dma on my vt8233A southbridge.
    I have 2 identical disks, the second controller works fine -- dma
    automatically. I cannot get any dma through my southbridge though.
    I have identical disks on the vt8233A controller (hdb) and a HPT372
    controller (hdg). The HighPoint controller works fine with the drive,
    Maxtor 60G, ATA100.
    In the kernel, I've compiled support for the HighPoint chipset in the
    kernel, BLK_DEV_HPT366. I've also set automatic use of DMA for IDE
    drives, IDEDMA_PCI_AUTO. I don't see any special compile options in
    regards to the VT8233A chipset.
    On the net, there was some discussion about early 2.4 kernels not
    supporting this chipset, but as of 2.4.19, it had been fixed. I've
    recently upgraded from my old 2.4.xx to 2.6.11 and now to 2.6.12-6 and
    have noticed this performance problem since then.

    Here is a sample of some commands I've used to gather information and
    correct the problem:
    # hdparm -i /dev/hd[abeg]
    /dev/hda:
      Model=WDC WD1600JB-00FUA0, FwRev=15.05R15, SerialNo=WD-WMAEP1214820
      Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
      RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=74
      BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16
      CurCHS=65535/1/63, CurSects=4128705, 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
      AdvancedPM=no WriteCache=enabled
      Drive conforms to: device does not report version:
      * signifies the current active mode

    /dev/hdb:
      Model=Maxtor 96147H6, FwRev=ZAH814Y0, SerialNo=V603P98C
      Config={ Fixed }
      RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57
      BuffType=DualPortCache, BuffSize=2048kB, MaxMultSect=16, MultSect=16
      CurCHS=4047/16/255, CurSects=16511760, LBA=yes, LBAsects=120064896
      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
      AdvancedPM=yes: disabled (255) WriteCache=enabled
      Drive conforms to: ATA/ATAPI-6 T13 1410D revision 0:
      * signifies the current active mode

    /dev/hde:
      Model=WDC WD1200BB-60CJA0, FwRev=16.06V16, SerialNo=WD-WMAA91061045
      Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
      RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=40
      BuffType=DualPortCache, BuffSize=2048kB, MaxMultSect=16, MultSect=16
      CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=234441648
      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
      AdvancedPM=no WriteCache=enabled
      Drive conforms to: device does not report version:
      * signifies the current active mode

    /dev/hdg:
      Model=Maxtor 96147H6, FwRev=ZAH814Y0, SerialNo=V60C42GC
      Config={ Fixed }
      RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57
      BuffType=DualPortCache, BuffSize=2048kB, MaxMultSect=16, MultSect=16
      CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=120064896
      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
      AdvancedPM=yes: disabled (255) WriteCache=enabled
      Drive conforms to: ATA/ATAPI-6 T13 1410D revision 0:
      * signifies the current active mode

    # hdparm -d1 /dev/hd[abeg]
    /dev/hda:
      setting using_dma to 1 (on)
      HDIO_SET_DMA failed: Operation not permitted
      using_dma = 0 (off)
    /dev/hdb:
      setting using_dma to 1 (on)
      HDIO_SET_DMA failed: Operation not permitted
      using_dma = 0 (off)
    /dev/hde:
      setting using_dma to 1 (on)
      using_dma = 1 (on)
    /dev/hdg:
      setting using_dma to 1 (on)
      using_dma = 1 (on)

    # hdparm -tT /dev/hd[abeg]
    /dev/hda:
      Timing cached reads: 924 MB in 2.00 seconds = 461.15 MB/sec
      Timing buffered disk reads: 12 MB in 3.27 seconds = 3.67 MB/sec
    /dev/hdb:
      Timing cached reads: 1012 MB in 2.01 seconds = 504.06 MB/sec
      Timing buffered disk reads: 10 MB in 3.22 seconds = 3.11 MB/sec
    /dev/hde:
      Timing cached reads: 924 MB in 2.01 seconds = 460.69 MB/sec
      Timing buffered disk reads: 134 MB in 3.02 seconds = 44.33 MB/sec
    /dev/hdg:
      Timing cached reads: 980 MB in 2.00 seconds = 488.85 MB/sec
      Timing buffered disk reads: 74 MB in 3.00 seconds = 24.65 MB/sec

    I'm running RHF4 (+extras and updates) 2.46.12-on a ABit KX7-333R
    motherboard, with 1GB RAM, and the following drives Western Digital
    WD1600JB-00FUA0 160GB ATA100 (hda), 2 Maxtor 96147H6 60GB ATA100's (hdb,
    hdg), and Western Digital WD1200BB-60CJA0 120GB ATA100 (hde).

    Please help.


  • Next message: John Thompson: "Re: ultraportable?"

    Relevant Pages

    • Re: drive has continual errors found.....
      ... Every option on the first page of CMOS about drives is alreadyset for ... the printer there was an option about DMA modes. ... Double-click the controller for which you want to restore the typical DMA transfer mode. ... dont use one of those stupid round cables. ...
      (alt.comp.hardware.pc-homebuilt)
    • Re: dma problems with 2.6 kernel and vt8233A southbridge
      ... I cannot get any dma through my southbridge though. ... The HighPoint controller works fine with the drive, ... > * signifies the current active mode ...
      (comp.os.linux.hardware)
    • Re: Whats going on with my IDE devices ?
      ... >> It's disabled on your controller, ... since the dma in question is ... Then what on earth is their ide ... * for drives which already have DMA enabled, ...
      (alt.os.linux)
    • Re: Do I have a drive problem with one of a matched pair?
      ... | Beemer wrote ... Run WD's diagnostic on both drives. ... then check if DMA is enabled and being used for both drives. ... DMA 2 - standard floppy controller ...
      (comp.sys.ibm.pc.hardware.storage)
    • More ATAng probing issues
      ... MB with a nVidia controller. ... Suddenly my drives don't show up on ... They do on a 8/28 kernel. ... simplex device, DMA on primary only ...
      (freebsd-current)