Description of problem:

Using the device mapper on a SATA disk with sata_sil24 seems to cause data
corruption.

Version-Release number of selected component (if applicable):

kernel-2.6.21-1.3228.fc7
device-mapper-libs-1.02.17-7.fc7
device-mapper-1.02.17-7.fc7

How reproducible:

Create two DM linear mappings such that they both live on the same disk.  mkfs
them.  Write to them.  fsck them and see corruption.

Steps to Reproduce:
1.  echo "0 482344960 linear 8:32 0" | dmsetup create one
    echo "0 209715200 linear 8:32 482345000" | dmsetup create two
    (change 8:32 as needed)
2.  mke2fs -j -m0 /dev/mapper/one
    mke2fs -j -m0 /dev/mapper/two
    mount /dev/mapper/one /one
    mount /dev/mapper/two /two

3.  cd /one ; \
    for i in `seq 0 3`; do dd if=/dev/zero bs=4K count=1M of=$i; done ; \
    cd ; \
    umount /one

4.  cd /two ; \
    for i in `seq 0 3`; do dd if=/dev/zero bs=4K count=1M of=$i; done ; \
    cd ; \
    umount /two

5.  fsck -f /dev/mapper/one
    fsck -f /dev/mapper/two

Actual results:

Many errors on fsck.

Expected results:

No errors on fsck.

Additional info:

The disk for me is a a Seagate 500GB SATA drive (ST3500630AS, firmware 3.AAE)
It is attached to the machine via an Addonics ADS3GX4R5-E eSATA card (which
appears to be a Silicon Image 3124 OEM).  The driver, in any event, is sata_sil24.

I debated whether to file this against LVM/device-mapper or the kernel.  kernel
won because given the libata changes in f7, it seemed a more likely place for a
bug to pop up.  I could be wrong.