As suggested by Keith I've merged the new flag for nvme-fabrics with the last
patch. I also added some information why nvme-fc is handling the initial connect
attempt differently as requested by Christoph.

The new flag is called connect_async and the default is false/0. I've tested
with a patched nvme-cli/libnvme version and all looks good. When this series is
accepted I'll update nvme-cli/libnvme accordingly. Note, that even with an older
nvme-cli the blktests (especially nvme/041) will pass with a newer kernel.

(nvme/048 is still fails but this is a different problem. On my TODO list)

changes:
v2:
 - renamed flag to connect_async, default false
 - add info to commit message why nvme-fc is different
 - merged connect_async with 'nvme-fc: wait for
   initial connect attempt to finish'

v1:
 - renamed 'nvme-fc: redesign locking and refcounting'
   to 'nvme-fc: reorder ctrl ref counting and cleanup code path'
 - testing with scsi/nvme dev_loss_tmo on real hw
 - removed rport ref counting part
 - collected RB tags
 - https://lore.kernel.org/linux-nvme/20240219131531.15134-1-dwagner@suse.de/
 
v0:
 - initial version
 - https://lore.kernel.org/linux-nvme/20240216084526.14133-1-dwagner@suse.de/

Daniel Wagner (5):
  nvme-fc: rename free_ctrl callback to match name pattern
  nvme-fc: do not retry when auth fails or connection is refused
  nvme-fabrics: introduce ref counting for nvmf_ctrl_options
  nvme-fc: reorder ctrl ref counting and cleanup code path
  nvme-fc: wait for initial connect attempt to finish

 drivers/nvme/host/fabrics.c |  39 +++++++++-
 drivers/nvme/host/fabrics.h |   9 ++-
 drivers/nvme/host/fc.c      | 145 +++++++++++++++++-------------------
 drivers/nvme/host/rdma.c    |  18 +++--
 drivers/nvme/host/tcp.c     |  21 ++++--
 drivers/nvme/target/loop.c  |  19 +++--
 6 files changed, 152 insertions(+), 99 deletions(-)

-- 
2.43.1

On Wed, Feb 21, 2024 at 02:23:59PM +0100, Daniel Wagner wrote:
> As suggested by Keith I've merged the new flag for nvme-fabrics with the last
> patch. I also added some information why nvme-fc is handling the initial connect
> attempt differently as requested by Christoph.
> 
> The new flag is called connect_async and the default is false/0. I've tested
> with a patched nvme-cli/libnvme version and all looks good. When this series is
> accepted I'll update nvme-cli/libnvme accordingly. Note, that even with an older
> nvme-cli the blktests (especially nvme/041) will pass with a newer kernel.
> 
> (nvme/048 is still fails but this is a different problem. On my TODO list)

Series looks good to me. Sounds like it's not fixing any regressions, so
I think this goes to the 6.9 branch. I'll wait till end of week for any
other feedback before pushing anything out.

On Wed, Feb 21, 2024 at 11:04:13AM -0700, Keith Busch wrote:
> On Wed, Feb 21, 2024 at 02:23:59PM +0100, Daniel Wagner wrote:
> > As suggested by Keith I've merged the new flag for nvme-fabrics with the last
> > patch. I also added some information why nvme-fc is handling the initial connect
> > attempt differently as requested by Christoph.
> > 
> > The new flag is called connect_async and the default is false/0. I've tested
> > with a patched nvme-cli/libnvme version and all looks good. When this series is
> > accepted I'll update nvme-cli/libnvme accordingly. Note, that even with an older
> > nvme-cli the blktests (especially nvme/041) will pass with a newer kernel.
> > 
> > (nvme/048 is still fails but this is a different problem. On my TODO list)
> 
> Series looks good to me. Sounds like it's not fixing any regressions, so
> I think this goes to the 6.9 branch. I'll wait till end of week for any
> other feedback before pushing anything out.

The second patch should be dropped in this series and replaced with the
patches from Hannes '[PATCH 0/4] nvme: fixes for authentication errors'

https://lore.kernel.org/linux-nvme/20240301112823.132570-1-hare@kernel.org/

Should I send out new version or do you take it from here?