netdev - Re: [RFC PATCH bpf-next 2/3] bpf: udp: Avoid socket skips and repeats during iteration

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: <CABi4-oiXH+H=6=LaajcQK5faqDn20tUQ86cTJXF0Om-zcxNSUQ@mail.gmail.com>
Date: Mon, 7 Apr 2025 16:39:25 -0700
From: Jordan Rife <jordan@...fe.io>
To: Martin KaFai Lau <martin.lau@...ux.dev>
Cc: netdev@...r.kernel.org, bpf@...r.kernel.org, 
	Aditi Ghag <aditi.ghag@...valent.com>, Daniel Borkmann <daniel@...earbox.net>, 
	Willem de Bruijn <willemdebruijn.kernel@...il.com>
Subject: Re: [RFC PATCH bpf-next 2/3] bpf: udp: Avoid socket skips and repeats
 during iteration

> nit. It is to get the first entry? May be directly do
> hlist_entry_safe(hslot2->head.first, ... ) instead.

Sure, I can change this and drop the RFC tag for the next iteration of
this series.

> My understanding is that it may or may not batch something newer than the last
> stop(). This behavior should be similar to the current offset approach also. I
> think it is fine. The similar situation is true for the next bucket anyway.

Assuming it's rare that the first unvisited socket disappears between
stop and start, which seems like a reasonable assumption, you should
generally only need to scan through the list once to find that socket
(similar amount of work to offset). Worst case is if every socket from
last time is no longer there. Then you'd end up scanning through the
full list end_cookie - find_cookie times. And yeah, I think the
iterator shouldn't really care if new sockets are seen or not as long
as you see all sockets that were there when you started iterating.

-Jordan