On 17/09/2021 13:21, Damien Le Moal wrote:

On 2021/09/17 19:24, Krzysztof Kozlowski wrote:

quoted

Intel devices 0x2822, 0x2823, 0x2826 and 0x2827 are already on the list
as Lewisburg AHCI/RAID.  They use same configuration except 0x2822 which
has board_ahci_nosntf (for ICH8).

Signed-off-by: Krzysztof Kozlowski <redacted>

---

Not tested.

I cannot test this either. I do not have this hardware. No chance of getting
this tested by someone ?

Changes for 0x2823, 0x2826 and 0x2827 seem OK, but 0x2822 is weird.
Since it is replicated, I think that the second entry was in fact never used as
the first one would be a hit before the second. So shouldn't we keep the first
entry with board_ahci_nosntf and remove the second one with board_ahci ?

The second 0x2822 entry was added by Alexandra Yates
[off-list ref].

However first entry was added with much bigger explanation by Shaohua Li
[off-list ref], where he/she tested it.

Indeed it points to using board_ahci_nosntf.

Best regards,
Krzysztof