netdev - Re: [PATCH net-next] bnxt_en: support PPS in/out on all pins

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: <f1c42ecd-57c0-4cea-906e-aebcd583944a@linux.dev>
Date: Fri, 17 Oct 2025 11:40:05 +0100
From: Vadim Fedorenko <vadim.fedorenko@...ux.dev>
To: Pavan Chebbi <pavan.chebbi@...adcom.com>
Cc: Michael Chan <michael.chan@...adcom.com>, Jakub Kicinski
 <kuba@...nel.org>, Andrew Lunn <andrew+netdev@...n.ch>,
 "David S. Miller" <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
 Paolo Abeni <pabeni@...hat.com>, Richard Cochran <richardcochran@...il.com>,
 netdev@...r.kernel.org
Subject: Re: [PATCH net-next] bnxt_en: support PPS in/out on all pins

On 17/10/2025 10:15, Pavan Chebbi wrote:
> On Fri, Oct 17, 2025 at 2:21 PM Vadim Fedorenko
> <vadim.fedorenko@...ux.dev> wrote:
>>
>> On 17.10.2025 04:45, Pavan Chebbi wrote:
>>> On Fri, Oct 17, 2025 at 3:54 AM Vadim Fedorenko
>>> <vadim.fedorenko@...ux.dev> wrote:
>>>>
>>>> n_ext_ts and n_per_out from ptp_clock_caps are checked as a max number
>>>> of pins rather than max number of active pins.
>>>
>>> I am not 100pc sure. How is n_pins going to be different then?
>>> https://elixir.bootlin.com/linux/v6.17/source/include/linux/ptp_clock_kernel.h#L69
>>
>> So in general it's more for the case where HW has pins connected through mux to
>> the DPLL channels. According to the bnxt_ptp_cfg_pin() bnxt HW has pins
>> hardwired to channels and NVM has pre-defined configuration of pins' functions.
>>
>> [host ~]# ./testptp -d /dev/ptp2 -l
>> name bnxt_pps0 index 0 func 0 chan 0
>> name bnxt_pps1 index 1 func 0 chan 1
>> name bnxt_pps2 index 2 func 0 chan 2
>> name bnxt_pps3 index 3 func 0 chan 3
>>
>> without the change user cannot configure EXTTS or PEROUT function on pins
>> 1-3 preserving channels 1-3 on them.
>>
>> The user can actually use channel 0 on every pin because bnxt driver doesn't
>> care about channels at all, but it's a bit confusing that it sets up different
>> channels during init phase.
> 
> You are right that we don't care about the channels. So I think
> ideally it should have been set to 0 for all the pins.
> Does that not make a better fix? Meaning to say, we don't care about
> the channel but/therefore please use 0 for all pins.
> What I am not sure about the proposed change in your patch is that it
> may be overriding the definition of the n_ext_ts and n_per_out in
> order to provide flexibility to users to change channels, no?

Well, yeah, the overriding exists, but that's mostly the artifact of not
so flexible API. But I agree, we can improve init part to make it clear.
But one more thing has just come to my mind - is it
really possible to configure PPS-in/PPS-out on pins 0-1?
AFAIU, there are functions assigned to each pin, which can only be
enabled or disabled by the user-space app, and in this case
bnxt_ptp_verify() should be improved to validate function per pin,
right?