In the patch:
+ /*
+ * Samba server e.g. can return an empty interface list in some cases,
+ * which would only be a problem if we were requesting multichannel
+ */
+ if (bytes_left == 0) {
+ /* avoid spamming logs every 10 minutes, so log only in mount */
+ if ((ses->chan_max > 1) && in_mount)
+ cifs_dbg(VFS,
+ "empty network interface list returned by server %s\n",
+ ses->server->hostname);
+ rc = -EINVAL;
+ goto out;
+ }
This logs the server name, but it might be confusing to the
admin since the mount does not actually fail. Perhaps add some
words to the effect of "multichannel not available"?
Acked-by: Tom Talpey <tom@xxxxxxxxxx>
On 10/3/2022 6:36 PM, Steve French wrote:
attached wrong patch - resending
On Mon, Oct 3, 2022 at 5:32 PM Steve French <smfrench@xxxxxxxxx> wrote:
updated patch to:
1) log the server name for this message
2) only log on mount (not every ten minutes)
See attached
On Mon, Oct 3, 2022 at 9:21 AM Tom Talpey <tom@xxxxxxxxxx> wrote:
On 10/3/2022 12:38 AM, Steve French wrote:
On Sat, Oct 1, 2022 at 6:22 PM Tom Talpey <tom@xxxxxxxxxx> wrote:
On 10/1/2022 12:54 PM, Steve French wrote:
Some servers can return an empty network interface list so, unless
multichannel is requested, no need to log an error for this, and
when multichannel is requested on mount but no interfaces, log
something less confusing. For this case change
parse_server_interfaces: malformed interface info
to
empty network interface list returned by server
Will this spam the log if it happens on every MC refresh (10 mins)?
It might be helpful to identify the servername, too.
Yes - I just noticed that in this case (multichannel mount to Samba
where no valid interfaces) we log it every ten minutes.
Maybe best way to fix this is to change it to a log once error (with
server name is fine with me) since it is probably legal to return an
empty list (so not serious enough to be worth logging every ten
minutes) and in theory server could fix its interfaces later.
Ten minutes is the default recommended polling interval in the doc.
While it's odd, it's not prevented by the protocol. I could guess
that a server running in a namespace might return strange things
as devices came and went, for example.
It's not an error, so the message is purely informational. It is
useful though. Is it possible to suppress the logging if the
message *doesn't* change, but otherwise emit new ones? That might
require some per-server fiddling to avoid multiple servers flipping
the message.
A boolean or bit in the server struct? A little ugly for the purpose,
but surfacing multichannel events - especially ones that prevent it
from happening - seems worthwhile.
Tom.
Tom.
Cc: <stable@xxxxxxxxxxxxxxx>
Signed-off-by: Steve French <stfrench@xxxxxxxxxxxxx>
See attached patch
--
Thanks,
Steve
