netdev - Re: [PATCH net] ibmveth: use net_err_ratelimited when set_multicast

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190802095300.GU18865@dhcp-12-139.nay.redhat.com>
Date:   Fri, 2 Aug 2019 17:53:01 +0800
From:   Hangbin Liu <liuhangbin@...il.com>
To:     David Miller <davem@...emloft.net>
Cc:     netdev@...r.kernel.org, tlfalcon@...ux.ibm.com
Subject: Re: [PATCH net] ibmveth: use net_err_ratelimited when
 set_multicast_list

On Thu, Aug 01, 2019 at 12:51:14PM -0400, David Miller wrote:
> From: Hangbin Liu <liuhangbin@...il.com>
> Date: Thu,  1 Aug 2019 17:03:47 +0800
> 
> > When setting lots of multicast list on ibmveth, e.g. add 3000 membership on a
> > multicast group, the following error message flushes our log file
> > 
> > 8507    [  901.478251] ibmveth 30000003 env3: h_multicast_ctrl rc=4 when adding an entry to the filter table
> > ...
> > 1718386 [ 5636.808658] ibmveth 30000003 env3: h_multicast_ctrl rc=4 when adding an entry to the filter table
> > 
> > We got 1.5 million lines of messages in 1.3h. Let's replace netdev_err() by
> > net_err_ratelimited() to avoid this issue. I don't use netdev_err_once() in
> > case h_multicast_ctrl() return different lpar_rc types.
> > 
> > Signed-off-by: Hangbin Liu <liuhangbin@...il.com>
> 
> Let's work on fixing what causes this problem, or adding a retry
> mechanism, rather than making the message less painful.

Yes, the multicast issue should also be fixed. It looks more like a
driver issue as I haven't seen it on other drivers.

I think these should be two different fixes.

Thanks
Hangbin
> 
> What is worse is that these failures are not in any way communicated
> back up the callchain to show that the operation didn't complete
> sucessfully.
> 
> This is a real mess in behavior and error handling, don't make it
> worse please.