Problem: 2.4.20, 2.5.66 have different IDE channel order
Quoted 5 lines
For compiled in IDE the order is defined by the order on the PCI bus scan, not by the IDE layer
(See ide/setup-pci.c:ide_scan_pcibus)
Well, that matches what 2.4 does:
00:0d.1 IDE interface: Intel Corp. 82371SB PIIX3 IDE [Natoma/Triton II] 00:10.0 Unknown mass storage controller: Promise Technology, Inc. 20268 (rev 02) 01:0b.0 Unknown mass storage controller: Triones Technologies, Inc. HPT366 / HPT370 (rev 03)
2.5 nonmodular seems to be doing it in BIOS order -- the HPT370 BIOS
initializes before the Promise (and won't let it boot but I can deal with that.) I'll probably replace it with a PDC20262 before looking at any code...
--
Chuck
Quoted 11 lines · Chuck Ebbert
Well, that matches what 2.4 does:
00:0d.1 IDE interface: Intel Corp. 82371SB PIIX3 IDE [Natoma/Triton II] 00:10.0 Unknown mass storage controller: Promise Technology, Inc. 20268 (rev 02) 01:0b.0 Unknown mass storage controller: Triones Technologies, Inc. HPT366 / HPT370 (rev 03)
2.5 nonmodular seems to be doing it in BIOS order -- the HPT370 BIOS
initializes before the Promise (and won't let it boot but I can deal with that.) I'll probably replace it with a PDC20262 before looking
Im a bit puzzled by this because it does look like a bug. Our pci scan code hasnt changed that materially. I assume the promise and hpt are both plug in cards >
On 15 Apr 2003 21:43:02 +0100 Alan Cox <[email protected]> wrote:
| On Maw, 2003-04-15 at 19:33, Chuck Ebbert wrote: | > Well, that matches what 2.4 does: | > | > | > 00:0d.1 IDE interface: Intel Corp. 82371SB PIIX3 IDE [Natoma/Triton II] | > 00:10.0 Unknown mass storage controller: Promise Technology, Inc. 20268 (rev 02) | > 01:0b.0 Unknown mass storage controller: Triones Technologies, Inc. HPT366 / HPT370 (rev 03) | > | > | > 2.5 nonmodular seems to be doing it in BIOS order -- the HPT370 BIOS | > initializes before the Promise (and won't let it boot but I can deal | > with that.) I'll probably replace it with a PDC20262 before looking | | Im a bit puzzled by this because it does look like a bug. Our pci scan code hasnt changed | that materially. I assume the promise and hpt are both plug in cards >
but something also changed NIC interface ordering (according to davej et al)... so maybe it's deep inside PCI bus scanning.
--
~Randy