[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YYHfHIcVSDOSuJnx@codewreck.org>
Date: Wed, 3 Nov 2021 10:00:12 +0900
From: Dominique Martinet <asmadeus@...ewreck.org>
To: Thomas Weißschuh <thomas@...ch.de>
Cc: Thomas Weißschuh <linux@...ssschuh.net>,
Eric Van Hensbergen <ericvh@...il.com>,
Latchesar Ionkov <lucho@...kov.net>,
"David S. Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>,
v9fs-developer@...ts.sourceforge.net, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] net/9p: autoload transport modules
Dominique Martinet wrote on Wed, Nov 03, 2021 at 09:26:32AM +0900:
> Feel free to resend in a proper form though, I could make up a commit
> message but it might as well be your words!
Ah, just a couple more things:
* make with W=1 complains about missing prototypes:
net/9p/trans_fd.c:1155:5: warning: no previous prototype for ‘p9_trans_fd_init’ [-Wmissing-prototypes]
1155 | int p9_trans_fd_init(void)
| ^~~~~~~~~~~~~~~~
net/9p/trans_fd.c:1164:6: warning: no previous prototype for ‘p9_trans_fd_exit’ [-Wmissing-prototypes]
1164 | void p9_trans_fd_exit(void)
| ^~~~~~~~~~~~~~~~
* This actually break the 'no trans=tcp' specified case when no extra
module is loaded, but I'm not sure how impactful that is.
See v9fs_get_default_trans(), they iterate through loaded transports
(through register_trans()), we might want to bake in a list that
additionally tries to load modules if no module is loaded at all
(in my opinion virtio makes sense before tcp, then fd, unix, xen, rdma?)
Well, that can probably come later.
--
Dominique
Powered by blists - more mailing lists