> > The patch adds entries for 0xFFFF (is that really right) as a > bridge and > 0xC409 - I see nothing for 0x0571. PCI_DEVICE_ID_VIA_ANON is used to stand for Unknown South Bridge for VX855 and future chips, since the pata_via has a southbridge check, which is redundant. If we may VIA has a new SB with new bridge ID which in not in the check list, whose IDE controller is 0x0571. Linux default pata_via can not drive IDE controller of 402, since the southbridge check in pata_via can not pass. So we add PCI_DEVICE_ID_VIA_ANON to stand for our future southbridge, and this will help to pass southbridge check. Is it possible to do this way? And can it be acceptable in kernel policy? > Also all the second port stuff is unneeded: Just do > > if (has_one_port) > ppi[1] = &ata_dummy_port_info; > > and only the first port will be probed. > We will check it out. Thanks. :) ��.n��������+%������w��{.n�����{��'^�)��jg��������ݢj����G�������j:+v���w�m������w�������h�����٥