On Wed, Jan 29, 2025 at 09:39:42AM +0100, Christian Kühnke wrote:

Hi all,

I have been sent here from the linux-ide mailing list. Over there, I have reported an issue with the 6.6 stable series of the kernel, starting with 6.6.51. The root cause is as follows:

On 12.09.2024, commit 872f86e1757bbb0a334ee739b824e47c448f5ebc ("ata: libata-scsi: Check ATA_QCFLAG_RTF_FILLED before using result_tf") was applied to 6.6, adding checks of ATA_QCFLAG_RTF_FILLED to libata_scsi. The patch seen in baseline commit 18676c6aab0863618eb35443e7b8615eea3535a9 ("ata: libata-core: Set ATA_QCFLAG_RTF_FILLED in fill_result_tf()") should have gone together with this.

Without it, I receive errors retrieving SMART data from SATA disks via a C602 SAS controller, apparently because in this situation ATA_QCFLAG_RTF_FILLED is not set.

I applied 18676c6aab0863618eb35443e7b8615eea3535a9 from baseline to 6.6.74 and the problem went away.

If you need any further information, do not hesitate to contact me (linux user since 0.99.x, but only debugging it once every few years or so...).

Now queued up, thanks.

greg k-h