netdev - Re: [PATCH v2 net-next 2/5] net-smnp: reorganize SNMP fast path variables

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]

Date: Tue, 17 Oct 2023 21:09:57 +0200
From: Andrew Lunn <andrew@...n.ch>
To: Eric Dumazet <edumazet@...gle.com>
Cc: Coco Li <lixiaoyan@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
	Neal Cardwell <ncardwell@...gle.com>,
	Mubashir Adnan Qureshi <mubashirq@...gle.com>,
	Paolo Abeni <pabeni@...hat.com>, netdev@...r.kernel.org,
	Chao Wu <wwchao@...gle.com>, Wei Wang <weiwan@...gle.com>,
	David Ahern <dsahern@...nel.org>
Subject: Re: [PATCH v2 net-next 2/5] net-smnp: reorganize SNMP fast path
 variables

On Tue, Oct 17, 2023 at 08:10:21PM +0200, Eric Dumazet wrote:
> On Tue, Oct 17, 2023 at 3:57 AM Andrew Lunn <andrew@...n.ch> wrote:
> >
> > On Tue, Oct 17, 2023 at 01:47:13AM +0000, Coco Li wrote:
> > > From: Chao Wu <wwchao@...gle.com>
> > >
> > > Reorganize fast path variables on tx-txrx-rx order.
> > > Fast path cacheline ends afer LINUX_MIB_DELAYEDACKLOCKED.
> > > There are only read-write variables here.
> > >
> > > Below data generated with pahole on x86 architecture.
> > >
> > > Fast path variables span cache lines before change: 12
> > > Fast path variables span cache lines after change: 2
> >
> > As i pointed out for the first version, this is a UAPI file.
> >
> > Please could you add some justification that this does not cause any
> > UAPI changes. Will old user space binaries still work after this?
> >
> > Thanks
> >         Andrew
> 
> I do not think the particular order is really UAPI. Not sure why they
> were pushed in uapi in the first place.
> 
> Kernel exports these counters with a leading line with the names of the metrics.
> 
> We already in the past added fields and nothing broke.
> 
> So the answer is : user space binaries not ignoring the names of the
> metrics will work as before.
> 
> nstat is one of the standard binary.

This is the sort of thing which i think should be in the commit
message. It makes it clear somebody has thought about this, and they
think the risk is minimal. Without such a comment, somebody will ask
if changing to a uapi file is safe.

   Andrew