ctld: Add abstractions to support multiple target protocols

[jhb]

Summary

This is a prerequisite for adding NVMe over Fabrics support.

Convert portal_group, portal_group_port, and target into abstract
classes with virtual methods to support protocol-specific methods.

Add new iscsi_portal_group, iscsi_port, iscsi_portal and iscsi_target
subclasses in a new iscsi.cc file and move some iSCSI-specific logic
there. Rename ctld_connection to iscsi_connection and move it to a
new iscsi.hh header. Move iscsi_connection methods out of ctld.cc and
kernel.cc into iscsi.cc.

Sponsored by: Chelsio Communications

[jhb]

Rebase after fixing previous patch

[asomers]

usr.sbin/ctld/ctld.h

116 ↗

(On Diff #150367)

Is there any reason why this couldn't be an enum?

139 ↗

(On Diff #150367)

Since it's actually allocating and returning a different string, this function isn't really a validator. How about "normalize_target_name" instead?

usr.sbin/ctld/kernel.c

572 ↗

(On Diff #150367)

Frankly, I think that all of these commented-out sections should just be deleted. It's highly likely that they don't all compile anymore.

[jhb]

usr.sbin/ctld/ctld.h

116 ↗

(On Diff #150367)

I guess it could be. I was just following the existing convention.

[jhb]

Rebase after C++ prep commits

[jhb]

Rewrite after C++ conversion

[asomers]

usr.sbin/ctld/Makefile

28

Why does Phabricator's web interface show that you removed this MK_ISCSI section, but when I download the code with "arc patch", I see no changes to this section?

usr.sbin/ctld/ctld.hh

59

Here too, Phabricator's web UI shows that you removed a line. But I don't see that when I download the code locally.

[jhb]

You can find the commits directly at https://github.com/freebsd/freebsd-src/compare/main...bsdjhb:freebsd:ctld_nvmf_cxx if that is easier (or you can just pull that branch to look at the changes locally)

usr.sbin/ctld/Makefile

28

Hmm, maybe arc patch is confused? I did remove it. It didn't seem worth keeping. At one point while working on this patch I had made iscsi.cc and discovery.cc conditional on MK_ISCSI, but I'd still need some sort of #ifdef's and it didn't seem worth the effort.

[imp]

I didn't read this like a hawk, but I spot checked and didn't see anything weird or abnormal and the couple of areas I spot checked were good.