Re: [PATCH net-next v2 3/3] genetlink: fit NLMSG_DONE into same read() as families

From: Gal Pressman <gal.pressman@linux.dev>
To: Jakub Kicinski <kuba@kernel.org>, Eric Dumazet <edumazet@google.com>
Cc: Stefano Brivio <sbrivio@redhat.com>,
	davem@davemloft.net, netdev@vger.kernel.org, pabeni@redhat.com,
	jiri@resnulli.us, idosch@idosch.org, johannes@sipsolutions.net,
	fw@strlen.de, pablo@netfilter.org, Martin Pitt <mpitt@redhat.com>,
	Paul Holzinger <pholzing@redhat.com>,
	David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [PATCH net-next v2 3/3] genetlink: fit NLMSG_DONE into same read() as families
Date: Thu, 21 Mar 2024 14:56:41 +0200	[thread overview]
Message-ID: <7e261328-42eb-411d-b1b4-ad884eeaae4d@linux.dev> (raw)
In-Reply-To: <20240319104046.203df045@kernel.org>

On 19/03/2024 19:40, Jakub Kicinski wrote:
> On Tue, 19 Mar 2024 18:17:47 +0100 Eric Dumazet wrote:
>>> Hi Stefano! I was worried this may happen :( I think we should revert
>>> offending commits, but I'd like to take it on case by case basis.
>>> I'd imagine majority of netlink is only exercised by iproute2 and
>>> libmnl-based tools. Does passt hang specifically on genetlink family
>>> dump? Your commit also mentions RTM_GETROUTE. This is not the only
>>> commit which removed DONE:
>>>
>>> $ git log --since='1 month ago' --grep=NLMSG_DONE --no-merges  --oneline
>>>
>>> 9cc4cc329d30 ipv6: use xa_array iterator to implement inet6_dump_addr()
>>> 87d381973e49 genetlink: fit NLMSG_DONE into same read() as families
>>> 4ce5dc9316de inet: switch inet_dump_fib() to RCU protection
>>> 6647b338fc5c netlink: fix netlink_diag_dump() return value  
>>
>> Lets not bring back more RTNL locking please for the handlers that
>> still require it.
> 
> Definitely. My git log copy/paste is pretty inaccurate, these two are
> better examples:
> 
> 5d9b7cb383bb nexthop: Simplify dump error handling
> 02e24903e5a4 netlink: let core handle error cases in dump operations
> 
> I was trying to point out that we merged a handful of DONE "coalescing"
> patches, and if we need to revert - let's only do that for the exact
> commands needed. The comment was raised on my genetlink patch while
> the discussion in the link points to RTM_GETROUTE.
> 
>> The core can generate an NLMSG_DONE by itself, if we decide this needs
>> to be done.
> 
> Exactly.
> 

We've encountered a new issue recently which I believe is related to
this discussion.

Following Eric's patch:
9cc4cc329d30 ("ipv6: use xa_array iterator to implement inet6_dump_addr()")

Setting the interface mtu to < 1280 results in 'ip addr show eth2'
returning an error, because the ipv6 dump fails. This is a degradation
from the user's perspective.

# ip addr show eth2
4: eth2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group
default qlen 1000
    link/ether 24:42:53:21:52:44 brd ff:ff:ff:ff:ff:ff
    altname enp6s0f0np0
# ip link set dev eth2 mtu 1000
# ip addr show eth2
RTNETLINK answers: No such device
Dump terminated