On 04/30/2015 11:21 AM, Christoph Hellwig wrote:
On Thu, Apr 30, 2015 at 11:14:27AM -0700, Andy Grover wrote:
TCMU requires more work to correctly handle both user handlers that want
all SCSI commands (pass_level=0) for a se_device, and also handlers that
just want I/O commands and let the others be emulated by the kernel
(pass_level=1).
Right now, TCMU is doing the right thing for level=1, but attributes and
transport_type are wrong for level=0. Do not allow it to be used for now.
We can resolve the issues and then enable it when it's ready.
The ABI break for TCMU v2 is already applied for this release, so it's
best to include this now to avoid another ABI break in the future.
Should we just remove the pass_level option for tcmuv2 entirely?
As mentioned in the earlier analysis I don't think this can be made
to work without a second se_subsystem_api regiration, and not selected
on an already existing configfs object.
If for some reason we will be able to add this as a runtime attribute
I'd prefer to have a more descriptive name than the rather nondescript
pass_level.
OK, I finally get it. We can't use a configuration value to
differentiate between the two modes, they need to look like completely
unrelated se_subsystems, and userspace picks ones or the other based on
the hba name it mkdirs, not a dev_param. Yes pass_level goes away.
I'll now work on what will be a slightly larger patch.
Thanks -- Andy
--
To unsubscribe from this list: send the line "unsubscribe target-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
