Thanks for the response, but I'm not convinced its a hardware CRC problem in
the drive. mDMA doesn't use CRC and it also has problems with the 2.2.16
kernel - time outs - see the tail end of the log messages. I'm not a
low-level IDE expert by any stretch but my understanding of CRC checking is
that the OS and the drive both calculate a CRC and then the OS sends its CRC
value to the drive. The drive checks the CRC from the OS against its
calculated CRC and if it doesn't match then the drive does a retry on the
operation. I don't know what you mean by simulate and blow it off, but if
it means that WD drives just believe whatever the OS says is correct, then
unless the OS is broken you would never get CRC errors reported. I'm having
a little trouble believing that my hard drive spontaneously crapped out when
I upgraded kernels. If I get a chance this weekend, I'll reboot with the
2.2.15 kernel and with Windows (its still on the drive) and see if they
still like DMA just to be sure.
- Ralph
----- Original Message -----
From: "Civileme" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, July 21, 2000 8:35 PM
Subject: Re: [expert] 2.2.16 kernel ID DMA problems
> On Fri, 21 Jul 2000, you wrote:
> > I just recently upgraded from the Mandrake 7.1 2.2.15 kernel to the
> > 2.2.16 kernel. Now I get BadCRC errors on my hard drive and DMA gets
> > disabled. If I try to set DMA on for any UDMA mode with hdparm I get
> > the same BadCRC errors. If I try to set any mDMA mode with hdparm I
> > get timeout waiting for DMA messages. DMA works fine on my CD-ROM
> > drive. My hardware is an ASUS P5A-B motherboard with a Cyrix 6x86-200
> > chip. Mach 64 vid, Zoom modem, 2 LNE100TX NIC's. Award BIOS just
> > flashed to 2/2000 level (version 1010) and I still have the problem.
Max
> > UDMA mode is UDMA 2.
> >
> > DMA was working fine in Windows before I converted the box to Linux
> > and was also working fine with the 2.2.15 kernel. Can anyone help me
> > get this working?
> >
> > Thanks - Ralph
> >
> > Here's the messages from the log:
>
> Ummm--don't mean to bust your bubble, but WDC drives don't do CRC--they
> simulate it and blow it off. Whether 2.2.16 has code specific to doing
the
> checking the drive is supposed to do or whether your WD drive allowed you
to
> get by with corrupt data before this is an open question. it could be
that the
> CRC emulator circuitry broke, too.
>
> FInd a nice, deep river. Take your computer box, a scrwedriver, and a
> stopwatch. Remove the WDC drive. With the stopwatch, time how long it
can
> float.
>
> Walk home. Take comfort in the fact that your (U)DMA CDROM is working
fine,
> even in (U)DMA modes.
>
> Civileme
>
> >
> > Jul 16 20:33:28 spinaltap syslog: syslogd startup succeeded
> > Jul 16 20:33:28 spinaltap syslog: klogd startup succeeded
> > Jul 16 20:33:28 spinaltap kernel: klogd 1.3-3, log source = /proc/kmsg
> > started.
> > Jul 16 20:33:28 spinaltap kernel: Inspecting
> > /boot/System.map-2.2.16-9mdksecure
> > Jul 16 20:33:28 spinaltap kernel: Loaded 7766 symbols from
> > /boot/System.map-2.2.16-9mdksecure.
> > Jul 16 20:33:28 spinaltap kernel: Symbols match kernel version 2.2.16.
> > Jul 16 20:33:28 spinaltap kernel: Loaded 96 symbols from 8 modules.
> > Jul 16 20:33:29 spinaltap kernel: Linux version 2.2.16-9mdksecure
> > ([EMAIL PROTECTED]) (gcc version 2.95.3 19991030
> > (prerelease)) #1 SMP Thu Jun 22 15:56:26 PDT 2000
> >
> > .....
> >
> > Jul 16 20:33:29 spinaltap kernel: Uniform Multi-Platform E-IDE driver
> > Revision: 6.30
> > Jul 16 20:33:29 spinaltap kernel: ide: Assuming 33MHz system bus speed
> > for PIO modes; override with idebus=xx
> > Jul 16 20:33:29 spinaltap kernel: ALI15X3: IDE controller on PCI bus
> > 00 dev 78
> > Jul 16 20:33:29 spinaltap kernel: ALI15X3: not 100% native mode: will
> > probe irqs later
> > Jul 16 20:33:29 spinaltap kernel: ide0: BM-DMA at 0xb000-0xb007,
> > BIOS settings: hda:DMA, hdb:pio
> > Jul 16 20:33:29 spinaltap kernel: ide1: BM-DMA at 0xb008-0xb00f,
> > BIOS settings: hdc:pio, hdd:pio
> > Jul 16 20:33:29 spinaltap kernel: hda: WDC AC25100L, ATA DISK drive
> > Jul 16 20:33:29 spinaltap kernel: hdc: ATAPI CDROM, ATAPI CDROM drive
> > Jul 16 20:33:29 spinaltap kernel: hdd: Seagate STT8000A, ATAPI TAPE
> > drive
> > Jul 16 20:33:29 spinaltap kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
> > Jul 16 20:33:29 spinaltap kernel: ide1 at 0x170-0x177,0x376 on irq 15
> > Jul 16 20:33:29 spinaltap kernel: hda: WDC AC25100L, 4924MB w/256kB
> > Cache, CHS=627/255/63, UDMA(33)
> > Jul 16 20:33:29 spinaltap kernel: hdc: ATAPI 32X CD-ROM drive, 128kB
> > Cache, (U)DMA
> > Jul 16 20:33:29 spinaltap kernel: Uniform CD-ROM driver Revision: 3.10
> >
> > Jul 16 20:33:29 spinaltap kernel: Floppy drive(s): fd0 is 1.44M
> > Jul 16 20:33:29 spinaltap kernel: FDC 0 is a post-1991 82077
> > Jul 16 20:33:29 spinaltap kernel: md driver 0.90.0 MAX_MD_DEVS=256,
> > MAX_REAL=12
> > Jul 16 20:33:29 spinaltap kernel: raid5: measuring checksumming speed
> > Jul 16 20:33:29 spinaltap kernel: raid5: MMX detected, trying
> > high-speed MMX checksum routines
> > Jul 16 20:33:29 spinaltap kernel: pII_mmx : 243.840 MB/sec
> > Jul 16 20:33:29 spinaltap kernel: p5_mmx : 285.750 MB/sec
> > Jul 16 20:33:29 spinaltap kernel: 8regs : 256.032 MB/sec
> > Jul 16 20:33:29 spinaltap kernel: 32regs : 176.784 MB/sec
> > Jul 16 20:33:29 spinaltap kernel: using fastest function: p5_mmx
> > (285.750 MB/sec)
> > Jul 16 20:33:29 spinaltap kernel: scsi : 0 hosts.
> > Jul 16 20:33:29 spinaltap kernel: scsi : detected total.
> > Jul 16 20:33:29 spinaltap kernel: md.c: sizeof(mdp_super_t) = 4096
> > Jul 16 20:33:29 spinaltap kernel: Partition check:
> > Jul 16 20:33:29 spinaltap kernel: hda:hda: dma_intr: status=0x51 {
> > DriveReady SeekComplete Error }
> > Jul 16 20:33:29 spinaltap kernel: hda: dma_intr: error=0x84 {
> > DriveStatusError BadCRC }
> > Jul 16 20:33:29 spinaltap kernel: hda: dma_intr: status=0x51 {
> > DriveReady SeekComplete Error }
> > Jul 16 20:33:29 spinaltap kernel: hda: dma_intr: error=0x84 {
> > DriveStatusError BadCRC }
> > Jul 16 20:33:29 spinaltap kernel: hda: dma_intr: status=0x51 {
> > DriveReady SeekComplete Error }
> > Jul 16 20:33:29 spinaltap kernel: hda: dma_intr: error=0x84 {
> > DriveStatusError BadCRC }
> > Jul 16 20:33:29 spinaltap kernel: hda: dma_intr: status=0x51 {
> > DriveReady SeekComplete Error }
> > Jul 16 20:33:29 spinaltap kernel: hda: dma_intr: error=0x84 {
> > DriveStatusError BadCRC }
> > Jul 16 20:33:29 spinaltap kernel: hda: DMA disabled
> > Jul 16 20:33:29 spinaltap kernel: ide0: reset: success
> > Jul 16 20:33:29 spinaltap kernel: hda1 < hda5 hda6 hda7 > hda4
> > Jul 16 20:33:29 spinaltap kernel: autodetecting RAID arrays
> > Jul 16 20:33:29 spinaltap kernel: autorun ...
> > Jul 16 20:33:29 spinaltap kernel: ... autorun DONE.
> > Jul 16 20:33:29 spinaltap kernel: VFS: Mounted root (ext2 filesystem)
> > readonly.
> >
> >
> > When I try mDMA.....
> >
> > Jul 19 20:01:10 spinaltap kernel: hda: timeout waiting for DMA
> > Jul 19 20:01:10 spinaltap kernel: ide_dmaproc: chipset supported
> > ide_dma_timeout func only: 12
> > Jul 19 20:01:10 spinaltap kernel: hda: irq timeout: status=0x58 {
> > DriveReady SeekComplete DataRequest }
> > Jul 19 20:01:20 spinaltap kernel: hda: timeout waiting for DMA
> > Jul 19 20:01:20 spinaltap kernel: ide_dmaproc: chipset supported
> > ide_dma_timeout func only: 12
> > Jul 19 20:01:20 spinaltap kernel: hda: irq timeout: status=0x58 {
> > DriveReady SeekComplete DataRequest }
> > Jul 19 20:01:30 spinaltap kernel: hda: timeout waiting for DMA
> > Jul 19 20:01:30 spinaltap kernel: ide_dmaproc: chipset supported
> > ide_dma_timeout func only: 12
> > Jul 19 20:01:30 spinaltap kernel: hda: irq timeout: status=0x58 {
> > DriveReady SeekComplete DataRequest }
> > Jul 19 20:01:40 spinaltap kernel: hda: timeout waiting for DMA
> > Jul 19 20:01:40 spinaltap kernel: ide_dmaproc: chipset supported
> > ide_dma_timeout func only: 12
> > Jul 19 20:01:40 spinaltap kernel: hda: irq timeout: status=0x58 {
> > DriveReady SeekComplete DataRequest }
> > Jul 19 20:01:41 spinaltap kernel: hda: DMA disabled
> > Jul 19 20:01:41 spinaltap kernel: ide0: reset: success
