Re: filesystem corruption

[Marc Horowitz <marc@mit.edu>]

"David S. Miller" <davem@redhat.com> writes:

>> On 12 Feb 2004 21:10:29 -0500
>> Marc Horowitz <marc@mit.edu> wrote:
>> 
>> > Briefly: each time an IDE channel is reset, this code reads pci config
>> > register 0x58, and saves it.  For channel 0, it clears bit 0x04,
>> > writes it back to the register, waits a millisecond, then writes the
>> > original saved value back.  For channel 1, it does the same with bit
>> > 0x08.
>> 
>> Register 0x58 specifies the primary channel address setup timing.
>> 
>> Only bits 0-2 are supposed to mean anything, the rest are reserved.
>> This workaround you describe, therefore, doesn't make much sense.

I went back to my source, and he told me I was incorrect in my earlier
description.  The register which is modified is pci config register
0x58 on the PCI-ISA bridge, not the ide controller.

>> > I'm not sure what this register does, but the algorithm is easy enough
>> > to implement.  However the kernel build is blowing out early, so I
>> > can't test it out:
>> 
>> Need to update your build tools probably.  I never use the make-deb script
>> myself, I always just build a vanilla tree to do my work.

It turns out I had to set MAKEFLAGS=CC=sparc64-linux-gcc.egcs64 in the
environment before running make-kpkg.  Perhaps the instructions at
http://www.debian.org/ports/sparc/ should be updated, since the
--subarch=sun4u flag also causes package generation to fail.  For the
benefit of the archives, I used this command to build a sparc64 kernel
deb from kernel-source-2.4.21:

# MAKEFLAGS=CC=sparc64-linux-gcc.egcs64 make-kpkg --revision=marc.1 kernel_image

This kernel seems to work, but I haven't tested it heavily, nor have I
tried adding the pci config magic yet.

Going back to your earlier suggestion that I boot with ide=nodma, that
made the system reliable.  However, surprisingly, what didn't work was
using hdparm to disable dma.  The system wouldn't even boot:

The output from /etc/init.d/hdparm:

    Setting disc parameters:hda: DMA disabled
     /dev/hda,hdc: DMA disabled
     /dev/hdc, done.

Then immediately after, a bunch of status errors:

    hdc: status error: status=0x58 { DriveReady SeekComplete DataRequest }

    hdc: drive not ready for command
    /etc/init.d/rcS:hdc: status error: status=0x58 line 54: /etc/r { cS.d/S20modutilsDriveReady : Permission denSeekComplete ied
    DataRequest }

    hdc: drive not ready for command
    /etc/init.d/rcS:hdc: status error: status=0x58 line 54: /etc/r { cS.d/S25raid2: PDriveReady ermission deniedSeekComplete 
    DataRequest }

    hdc: drive not ready for command
    /etc/init.d/rcS:hdc: status error: status=0x58 xmalloc: cannot {  allocate 154730DriveReady 497 bytes (0 bytSeekComplete es allocated)
    DataRequest }

Then a bunch of these: (the disk is an 80gb disk)

    attempt to access beyond end of device
    16:01: rw=0, want=774766593, limit=522112
    attempt to access beyond end of device
    16:01: rw=0, want=1768843568, limit=522112
    attempt to access beyond end of device
    16:01: rw=0, want=851982, limit=522112

Finally, this, with the "failed to exec..." repeating forever:

    kmod: runaway modprobe loop assumed and stopped
    kmod: runaway modprobe loop assumed and stopped
    kmod: failed to exec /sbin/modprobe -s -k binfmt-7f45, errno = 8
    kmod: runaway modprobe loop assumed and stopped
    kmod: runaway modprobe loop assumed and stopped
    kmod: failed to exec /sbin/modprobe -s -k binfmt-7f45, errno = 8
    kmod: failed to exec /sbin/modprobe -s -k binfmt-7f45, errno = 8
    kmod: runaway modprobe loop assumed and stopped
    kmod: failed to exec /sbin/modprobe -s -k binfmt-7f45, errno = 8
    kmod: failed to exec /sbin/modprobe -s -k binfmt-7f45, errno = 8
    kmod: failed to exec /sbin/modprobe -s -k binfmt-7f45, errno = 8
    kmod: failed to exec /sbin/modprobe -s -k binfmt-7f45, errno = 8

                Marc