linux-kernel - Re: [PATCH RFC net-next 7/7] netdev_features: convert NETIF_F_FCOE_MTU to IFF_FCOE

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, 16 Apr 2024 14:45:42 +0200
From: Alexander Lobakin <aleksander.lobakin@...el.com>
To: Jakub Kicinski <kuba@...nel.org>
CC: "David S. Miller" <davem@...emloft.net>, Eric Dumazet
	<edumazet@...gle.com>, Paolo Abeni <pabeni@...hat.com>, David Ahern
	<dsahern@...nel.org>, Xuan Zhuo <xuanzhuo@...ux.alibaba.com>, Andrew Lunn
	<andrew@...n.ch>, <nex.sw.ncis.osdt.itp.upstreaming@...el.com>,
	<netdev@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH RFC net-next 7/7] netdev_features: convert
 NETIF_F_FCOE_MTU to IFF_FCOE_MTU

From: Jakub Kicinski <kuba@...nel.org>
Date: Thu, 11 Apr 2024 07:47:09 -0700

> On Thu, 11 Apr 2024 12:28:08 +0200 Alexander Lobakin wrote:
>>> Any reason not to make it a bitfield? I haven't looked at the longer
>>> patches but this one seems to be used like a basic bool.  
>>
>> This whole enum could be made as bitfields, should we convert it? Would
>> be a big patch tho ._.
> 
> As always, I haven't investigated closely :) But my thinking was -
> we are at 34 bits in priv. We just need to convert 2 of them to
> a bitfield, pick two with fewest uses. Then we can downgrade 
> the field to u32 from ulonglong, and we can carry on adding bitfields?

Hmm, sounds good. You mean convert the existing bits which don't fit
into u32 to bitfields and then add new priv flags as bitfields?

> 
>>> But this definitely _is_ a uAPI change, right?  
>>
>> Why?
> 
> It will be user visible, ethtool -k is losing a field.
> Whether that's actually going to break anything depends on how silly
> user space is.

ethtool -{k,K} output would definitely change, but it's not an ABI.
Rather some corpo scripts may break, OTOH those "features" never were
available for toggling.

> 
> As Andrew pointed out, definitely something that should be called out
> in the commit message.

Sure, good point.

Thanks,
Olek