I managed to fix the iSER endpoint issue by making endpoints created
without a host valid again. Once I had iSER working, I went ahead and
made it network namespace aware as well.
Only tested with software roce (rxe) against the kernel target.
I think the net_exit code might need to do a bit
iSER creates endpoints before hosts and sessions, so we need to keep
compatibility with virtual device paths and not always expect a parent
host.
Signed-off-by: Chris Leech
---
drivers/scsi/scsi_transport_iscsi.c | 9 ++---
include/scsi/scsi_transport_iscsi.h | 3 ++-
2 files changed, 8
Split endpoint creation into host-bound and with a specified namespace,
for iSER's unique use of enpoint objects + virtual host-per-session.
This is much like was done with sessions already for iscsi_tcp.
Signed-off-by: Chris Leech
---
drivers/infiniband/ulp/iser/iscsi_iser.c | 6 ++--
Move the logic to store the network namespace during virtual host
creation (due to the way transport-class object setup callbacks
function) from iscsi_tcp into libiscsi, and share it with iser.
Signed-off-by: Chris Leech
---
drivers/infiniband/ulp/iser/iscsi_iser.c | 7 ++-
On Wed, Feb 08, 2023 at 09:40:50AM -0800, Lee Duncan wrote:
> Right now the iscsi_endpoint is only linked to a connection once that
> connection has been established. For net namespace filtering of the
> sysfs objects, associate an endpoint with the host that it was
> allocated for when it is