netdev - Re: [PATCH 0/5] Rust abstractions for network device drivers

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: <838805e5-c2a6-e3f3-d2e7-d435f07b9bda@ryhl.io>
Date: Fri, 16 Jun 2023 15:24:43 +0200
From: Alice Ryhl <alice@...l.io>
To: Andrew Lunn <andrew@...n.ch>
Cc: Jakub Kicinski <kuba@...nel.org>, netdev@...r.kernel.org,
 rust-for-linux@...r.kernel.org, aliceryhl@...gle.com,
 miguel.ojeda.sandonis@...il.com, FUJITA Tomonori <fujita.tomonori@...il.com>
Subject: Re: [PATCH 0/5] Rust abstractions for network device drivers

On 6/16/23 15:20, Andrew Lunn wrote:
>> As for this being a single function rather than four functions, that's
>> definitely a debatable decision. You would only do that if it makes sense to
>> merge them together and if you would always assign all of them together. I
>> don't know enough about these fields to say whether it makes sense here.
> 
> It can actually make sense to do them all together, because the source
> of these is likely to be a per CPU data structure protected by a per
> CPU sequence lock. You iterate over all CPUs, doing a transaction,
> taking the sequence lock, copy the values, and then releasing the
> lock. Taking and releases the lock per value is unnecessary expense.

It can probably be split into several methods without introducing a lock 
call for each one, if the API is designed right.

Alice