Well, well, well - Three holes in the ground.

My bridge boards *are* choking on > 200 sector transfers.

I patched sata_via in the same way sata_sil was patched to tell the controller to use 200 as 
max_sectors the same way sata_sil uses 15 as max_sectors on the seagates.

The problem is that ata_scsi_slave_config in libata-scsi.c is getting called *after* .dev_config in 
sata_via and resetting my value of 200 to 2048.

If I hard code 200 into ata_scsi_slave_config then this issue never raises it's ugly head and I 
can't get the bridge boards to crash.

I wonder if this might be happening in sata_sil also?

I added a simple printk in ata_scsi_slave_config to print "Slave Config!!!"
See attached log. I have also attached my patch to sata_via for sanity checking, in case I did 
something obviously wrong.

ata9: no device found (phy stat 00000000)
scsi8 : sata_promise
ata10: no device found (phy stat 00000000)
scsi9 : sata_promise
ata11: no device found (phy stat 00000000)
scsi10 : sata_promise
ata12: no device found (phy stat 00000000)
scsi11 : sata_promise
ACPI: PCI interrupt 0000:00:0f.0[B] -> GSI 20 (level, low) -> IRQ 20
sata_via(0000:00:0f.0): routed to hard irq line 0
ata13: SATA max UDMA/133 cmd 0x8800 ctl 0x8402 bmdma 0x7400 irq 20
ata14: SATA max UDMA/133 cmd 0x8000 ctl 0x7802 bmdma 0x7408 irq 20
ata13: dev 0 ATA, max UDMA/100, 390622887 sectors: lba48

ata13(0): applying WD errata fix

ata13: dev 0 configured for UDMA/100
scsi12 : sata_via
ata14: no device found (phy stat 00000000)
scsi13 : sata_via
   Vendor: ATA       Model: WDC WD2000JB-00D  Rev: 02.1
   Type:   Direct-Access                      ANSI SCSI revision: 05

Slave Config!!!

SCSI device sda: 390622887 512-byte hdwr sectors (199999 MB)
SCSI device sda: drive cache: write back
  /dev/scsi/host12/bus0/target0/lun0: unknown partition table
Attached scsi disk sda at scsi12, channel 0, id 0, lun 0
Attached scsi generic sg0 at scsi12, channel 0, id 0, lun 0,  type 0

Regards,
Brad