Re: [PATCH v2 2/2] scsi: sd: Rework asynchronous resume support

[Bart Van Assche]

On 8/17/22 12:07, Vlastimil Babka wrote:
> In my case it's
> /sys/devices/pci0000:00/0000:00:17.0/ata1/host0/scsi_host/host0/proc_name:ahci
> /sys/devices/pci0000:00/0000:00:17.0/ata2/host1/scsi_host/host1/proc_name:ahci
>
> Some more details from dmesg
>
> [    0.849373] ata2: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
> [    0.852849] ata2.00: ACPI cmd f5/00:00:00:00:00:a0(SECURITY FREEZE LOCK) filtered out
> [    0.854671] ata2.00: supports DRM functions and may not be fully accessible
> [    0.856181] ata2.00: ATA-9: SAMSUNG MZ7LN512HMJP-000L7, MAV01L6Q, max UDMA/133
> [    0.858115] ata2.00: 1000215216 sectors, multi 1: LBA48 NCQ (depth 32), AA
> [    0.861584] ata2.00: Features: Trust Dev-Sleep NCQ-sndrcv
> [    0.863749] ata2.00: ACPI cmd f5/00:00:00:00:00:a0(SECURITY FREEZE LOCK) filtered out
> [    0.865481] ata2.00: supports DRM functions and may not be fully accessible
> [    0.870043] ata2.00: configured for UDMA/133
> [    0.871871] scsi 1:0:0:0: Direct-Access     ATA      SAMSUNG MZ7LN512 1L6Q PQ: 0 ANSI: 5
>
> Please Cc me on further questions/steps to try/patches to test.

Hi Vlastimil,

Thank you for having provided the above information. The root cause of 
the hang is not yet clear to me. I was wondering whether the hang 
perhaps would be triggered by controllers that only support queue depth 
1. However, in the above output I see "depth 32".

As already reported in this email thread a revert for commit 
88f1669019bd62b3 has been posted on the linux-scsi mailing list. 
Additionally, Greg KH has been asked to drop that patch from the stable 
trees.

Thanks,

Bart.