mbox series

[0/4] scsi_debug improvements

Message ID 20250224115517.495899-1-john.g.garry@oracle.com (mailing list archive)
Headers show
Series scsi_debug improvements | expand

Message

John Garry Feb. 24, 2025, 11:55 a.m. UTC 
This is a re-post with modifications of Bart's series at
https://lore.kernel.org/linux-scsi/20240307203015.870254-1-bvanassche@acm.org/

I re-ordered the patches solving the atomic context sleeping and
simplifying command handling. This makes it easier to solve the atomic
context sleeping issue. However, it is harder to backport. Personally I
don't think that backporting is so important as the atomic context
sleeping is so difficult to recreate.

The main change in this series is to solve the atomic context sleeping
issue by now reporting errors back from scsi_debug_abort() to the SCSI
midlayer when we cannot guarantee if the command was aborted.

Based on 1d67c48947c6 (origin/staging, origin/for-next,
origin/6.15/scsi-staging) Merge patch series "scsi: scsi_debug: Add
more tape support"

Bart Van Assche (3):
  scsi: scsi_debug: Remove a reference to in_use_bm
  scsi: scsi_debug: Simplify command handling
  scsi: scsi_debug: Do not sleep in atomic sections

John Garry (1):
  scsi: scsi_debug: Remove sdebug_device_access_info

 drivers/scsi/scsi_debug.c | 153 ++++++++------------------------------
 1 file changed, 30 insertions(+), 123 deletions(-)

Comments

Martin K. Petersen Feb. 25, 2025, 1:57 a.m. UTC | #1 
John,

> The main change in this series is to solve the atomic context sleeping
> issue by now reporting errors back from scsi_debug_abort() to the SCSI
> midlayer when we cannot guarantee if the command was aborted.

Applied to 6.15/scsi-staging, thanks!