netdev - Re: [PATCH net-next v3 2/5] eth: fbnic: add initial PHC support

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: <20241007181604.32b5a330@kernel.org>
Date: Mon, 7 Oct 2024 18:16:04 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Jacob Keller <jacob.e.keller@...el.com>
Cc: Vadim Fedorenko <vadim.fedorenko@...ux.dev>, Vadim Fedorenko
 <vadfed@...a.com>, David Ahern <dsahern@...nel.org>, "Paolo Abeni"
 <pabeni@...hat.com>, "David S. Miller" <davem@...emloft.net>, "Alexander
 Duyck" <alexanderduyck@...com>, <netdev@...r.kernel.org>, Richard Cochran
 <richardcochran@...il.com>
Subject: Re: [PATCH net-next v3 2/5] eth: fbnic: add initial PHC support

On Mon, 7 Oct 2024 16:49:45 -0700 Jacob Keller wrote:
> >> Thanks for pointing this out, I'll make it with timecounter/cyclecounter  
> > 
> > Please don't, the clock is synthonized, we only do simple offsetting.
> >   
> I still think it makes sense to re-use the logic for converting cycles
> to full 64bit time values if possible.
> 
> If you're already doing offset adjustment, you still have to apply the
> same logic to every timestamp, which is exactly what a timecounter does
> for you.

"exactly what a timecounter does for you" is an overstatement.
timecounter tracks the overflows by itself and synthonizes. 
We have the 64b value, just need to combine the top bits.

> You can even use a timecounter and cyclecounter without using its
> ability to do syntonizing, by just setting the cyclecounter values
> appropriately, and leaving the syntonizing to the hardware mechanism.
> 
> I think the result is much easier to understand and follow than
> re-implementing a different mechanism for offset correction with bespoke
> code.

This is fast path code, I hope we can get a simple addition and
an overflow check right.