Commit
de8ef4c09faf6cd3a added HFLAG_YES_FBS workaround for 88SE9128
(1b4b:9123).
However, that change inadvertently caused the legacy IDE interface of
the controller (with the same pci id) to become associated with the AHCI
driver as well, causing the driver to try to bring the interface up in
vain.
Fix that by matching against class as well.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
{ PCI_VDEVICE(MARVELL, 0x6145), board_ahci_mv }, /* 6145 */
{ PCI_VDEVICE(MARVELL, 0x6121), board_ahci_mv }, /* 6121 */
{ PCI_DEVICE(0x1b4b, 0x9123),
+ .class = PCI_CLASS_STORAGE_SATA_AHCI,
+ .class_mask = 0xffffff,
.driver_data = board_ahci_yes_fbs }, /* 88se9128 */
/* Promise */