There is a lingering corruption in sata_mv that took me several days to

narrow down...  the failure to mask off the lower 16 bits of the S/G

length was leading to corruption, because this stomped on other

important bits in that S/G entry.
This should fix it (stable aka 2.6.23.1 needs this too):
Please pull from 'upstream-fixes' branch of

master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev.git upstream-fixes
to receive the following updates:
 drivers/ata/sata_mv.c |   27 ++++++++++++++++++---------

 1 files changed, 18 insertions(+), 9 deletions(-)
commit 6c08772e49622e90d39903e7ff0be1a0f463ac86

Author: Jeff Garzik <jeff@garzik.org>

Date:   Fri Oct 12 00:16:23 2007 -0400
    [libata] sata_mv: more S/G fixes
    * corruption fix: we only want the lower 16 bits of length (0 == 64kb)
    * ditto: the upper layer sets max-phys-segments to LIBATA_MAX_PRD,

      so we must reset it to own hw-specific length.
    * delete unused mv_fill_sg() return value
    Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
diff --git a/drivers/ata/sata_mv.c b/drivers/ata/sata_mv.c

index cb7dec9..d9832e2 100644

--- a/drivers/ata/sata_mv.c

+++ b/drivers/ata/sata_mv.c

@@ -69,10 +69,11 @@

 #include <linux/device.h>

 #include <scsi/scsi_host.h>

 #include <scsi/scsi_cmnd.h>

+#include <scsi/scsi_device.h>

 #include <linux/libata.h>
 #define DRV_NAME	"sata_mv"

-#define DRV_VERSION	"1.0"

+#define DRV_VERSION	"1.01"
 enum {

 	/* BAR's are enumerated in terms of pci_resource_start() terms */

@@ -420,6 +421,7 @@ static void mv_error_handler(struct ata_port *ap);

 static void mv_post_int_cmd(struct ata_queued_cmd *qc);

 static void mv_eh_freeze(struct ata_port *ap);

 static void mv_eh_thaw(struct ata_port *ap);

+static int mv_slave_config(struct scsi_device *sdev);

 static int mv_init_one(struct pci_dev *pdev, const struct pci_device_id *ent);
 static void mv5_phy_errata(struct mv_host_priv *hpriv, void __iomem ...
[ message continues ]

I'm guessing that this is critical enough to do a 2.6.23.1 with this fix

in it right now.  Any objections to that?
thanks,
greg k-h

-