ata fixes for 6.6-final

Linus,

The following changes since commit 626b13f015e080e434b1dee9a0c116ddbf4fb695:

scsi: Do not rescan devices with a suspended queue (2023-10-10 14:42:22 +0900)

are available in the Git repository at:

ssh://[email protected]/pub/scm/linux/kernel/git/dlemoal/libata tags/ata-6.6-final

for you to fetch changes up to 24eca2dce0f8d19db808c972b0281298d0bafe99:

scsi: sd: Introduce manage_shutdown device flag (2023-10-27 10:00:19 +0900)

----------------------------------------------------------------
ATA fixes for 6.6-final

A single patch to fix a regression introduced by the recent suspend/resume fixes. The regression is that ATA disks are not stopped on system shutdown, which is not recommended and increases the disks SMART counters for unclean power off events. This patch fixes this by refining the recent rework of the scsi device manage_xxx flags.

----------------------------------------------------------------
Damien Le Moal (1):
scsi: sd: Introduce manage_shutdown device flag

 drivers/ata/libata-scsi.c  |  5 +++--
 drivers/firewire/sbp2.c    |  1 +
 drivers/scsi/sd.c          | 39 ++++++++++++++++++++++++++++++++++++---
 include/scsi/scsi_device.h | 20 ++++++++++++++++++--
 4 files changed, 58 insertions(+), 7 deletions(-)

Side note: I think 'bool' ends up always using a whole byte in 'struct scsi_device' (because you can't take an address of a single bit).

And while it might all end up being unnoticeable due to padding, those bools are right next to bitfields.

Which just makes me go "those should all have been just 'unsigned:1' fields".

Of course, maybe they actually have their address taken, and it's all very intentional.

Linus

I will check. Thanks.

Damien Le Moal
Western Digital Research