netdev - Re: [PATCH net-next v5 1/2] ptp: introduce Alibaba CIPU PHC driver

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: <f2afb292-287e-4f2f-b131-50a1650bbb1d@linux.alibaba.com>
Date: Fri, 28 Nov 2025 14:22:21 +0800
From: Wen Gu <guwen@...ux.alibaba.com>
To: Jakub Kicinski <kuba@...nel.org>,
 Richard Cochran <richardcochran@...il.com>
Cc: andrew+netdev@...n.ch, davem@...emloft.net, edumazet@...gle.com,
 pabeni@...hat.com, xuanzhuo@...ux.alibaba.com, dust.li@...ux.alibaba.com,
 netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next v5 1/2] ptp: introduce Alibaba CIPU PHC driver

On 2025/11/28 00:36, Jakub Kicinski wrote:
> On Thu, 27 Nov 2025 13:48:47 +0800 Wen Gu wrote:
>>> We can't delete existing drivers. It used to be far less annoying
>>> until every cloud vendor under the sun decided to hack up their own
>>> implementation of something as simple as the clock.
>>
>> So what kind of drivers do you think are qualified to be placed in the
>> drivers/ptp? I checked some docs, e.g.[1], and codes in drivers/ptp,
>> but I am not sure what the deciding factor is, assuming that exposing
>> a PTP character device is not sufficient.
>>
>> [1] https://docs.kernel.org/driver-api/ptp.html
> 
> Networking ones? I don't have a great answer. My point is basically
> that we are networking maintainers. I have a good understanding of PTP
> (the actual protocol) and TSN as these are networking technologies.
> But I don't feel qualified to review purely time / clock related code.
> I don't even know the UNIX/Linux clock API very well.
> 

Searching the codebase for ptp_clock_register() shows that the PTP
implementations that related to the networking (IEEE 1588) are directly
embedded in the NIC drivers (drivers/net), e.g. ena_phc.c, ice_ptp.c,
igb_ptp.c, instead of being in drivers/ptp.

Most drivers under drivers/ptp are not directly related to networking
and operate independently of NICs. They maintain accuracy through their
own mechanisms.

> Sorry to put you in this position but the VM clocks should have some
> other tree. Or at the very least some clock expert needs to review them.
> 

I noticed you emphasized 'VM clock', but I don’t think it really matters
whether it’s a VM/cloud scenario or not. IMHO whether the PHC is provided
by a chip, FPGA, hypervisor or CIPU makes no fundamental difference.

> Could you go complain to clock people? Or virtualization people?

I understand that the PTP implementations in drivers/ptp aren't closely
related to networking though drivers/ptp is included in NETWORKING DRIVERS
in the MAINTAINER file.

I noticed that drivers/ptp/* is also inclued in PTP HARDWARE CLOCK SUPPORT.
This attribution seems more about 'clock'.

Hi @Richard Cochran, could you please review this? Thanks! :)

Regards.