Re: [PATCH v3] net/9p/fd: support ipv6 for trans=tcp

[asmadeus@codewreck.org]

asmadeus@codewreck.org wrote on Mon, Jan 20, 2025 at 05:21:39PM +0900:
> Meanwhile I should have tried earlier but checkpatch has a few qualms
> about style; if you send a v4 for any other reason please try running it
> yourself.
> [..]

There's no value in having you send a v4 just for this so I went ahead
and just removed a couple of spaces in my tree;
you can find the commit here if you want to check:
https://github.com/martinetd/linux/commits/9p-next

> I can't make any promise about testing but will try to find time around
> next weekend, after which the patch will slowly make its way in next
> cycle.

And I found time for this sooner than I thought I would: mount worked
just fine, both in v4 and v6.

Since there is no obvious regression I've pushed the patch to my "next"
branch, which will be merged into linux-next shortly and I'll submit to
Linus next cycle -- we're in a merge window right now but I don't have
any other patch ready and this doesn't feel like it needs to be rushed
so it'll be in a couple of months.


FWIW, the "current" workaround that also works with any released kernel
is to mount as follow, with shells like bash that allow /dev/tcp this is
actually easy to script:
```
exec 3<>"/dev/tcp/$TCP_SERVER/$TCP_PORT"
mount -t 9p -o trans=fd,aname="$TCP_ANAME",rfdno=3,wfdno=3 none /mnt
exec 3<&-
```


Thank you again, this has been on my todo list for almost as long as I've
been cattering to this 9p client!

(and I'll conveniently ignore the fact that trans_rdma has the very same
problem, I haven't had access to hardware to test RDMA for years so
unless someone cares it's probably better to leave it untouched..)

-- 
Dominique Martinet | Asmadeus