From: "Toke Høiland-Jørgensen" <toke@redhat.com>
To: Felix Fietkau <nbd@nbd.name>, linux-wireless@vger.kernel.org
Cc: johannes@sipsolutions.net
Subject: Re: [PATCH 2/7] mac80211: force calculation of software hash for tx fair queueing
Date: Thu, 17 Dec 2020 14:01:03 +0100 [thread overview]
Message-ID: <871rfobn8w.fsf@toke.dk> (raw)
In-Reply-To: <add3d1fa-c2ad-5aaf-83c7-31e919129bbf@nbd.name>
Felix Fietkau <nbd@nbd.name> writes:
> On 2020-12-17 12:54, Toke Høiland-Jørgensen wrote:
>> Felix Fietkau <nbd@nbd.name> writes:
>>
>>> Depending on the source, a hardware calculated hash may not provide the
>>> same level of collision resistance.
>>
>> This seems like it would have performance implications?
>>
>> Also, this can potentially discard information from tunnels that
>> preserve the hash before encapsulation (we added support for this to
>> Wireguard which had some nice effects on queueing of encapsulated
>> traffic).
> If the hash was calculated in software using the flow dissector, it will
> be preserved, even if it went through a few virtual interfaces.
> The only hashes discarded are hardware generated ones.
Yeah, but I was thinking something like:
Packet comes in with HW hash -> gets encapsulated (preserving the hash)
-> gets to mac80211 which discards the HW hash. So now you're replacing
a (possibly bad-quality) HW hash with a software hash of the *outer*
encapsulation header...
-Toke
next prev parent reply other threads:[~2020-12-17 13:02 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-16 20:43 [PATCH 1/7] net/fq_impl: bulk-free packets from a flow on overmemory Felix Fietkau
2020-12-16 20:43 ` [PATCH 2/7] mac80211: force calculation of software hash for tx fair queueing Felix Fietkau
2020-12-17 11:54 ` Toke Høiland-Jørgensen
2020-12-17 12:20 ` Felix Fietkau
2020-12-17 13:01 ` Toke Høiland-Jørgensen [this message]
2020-12-17 15:48 ` Felix Fietkau
2020-12-17 17:26 ` Toke Høiland-Jørgensen
2020-12-17 19:07 ` Felix Fietkau
2020-12-18 12:41 ` Toke Høiland-Jørgensen
2020-12-18 13:40 ` Felix Fietkau
2020-12-18 15:49 ` Toke Høiland-Jørgensen
2020-12-16 20:43 ` [PATCH 3/7] net/fq_impl: drop get_default_func, move default flow to fq_tin Felix Fietkau
2020-12-17 11:55 ` Toke Høiland-Jørgensen
2020-12-16 20:43 ` [PATCH 4/7] net/fq_impl: do not maintain a backlog-sorted list of flows Felix Fietkau
2020-12-16 20:59 ` Johannes Berg
2020-12-17 12:40 ` Felix Fietkau
2020-12-16 20:43 ` [PATCH 5/7] mac80211: fix encryption key selection for 802.3 xmit Felix Fietkau
2020-12-16 20:43 ` [PATCH 6/7] mac80211: fix fast-rx encryption check Felix Fietkau
2020-12-16 20:43 ` [PATCH 7/7] mac80211: add rx decapsulation offload support Felix Fietkau
2020-12-16 21:03 ` Johannes Berg
2020-12-16 21:19 ` Felix Fietkau
2020-12-17 8:08 ` Johannes Berg
2020-12-16 21:04 ` Johannes Berg
2020-12-16 21:06 ` Felix Fietkau
2020-12-16 20:54 ` [PATCH 1/7] net/fq_impl: bulk-free packets from a flow on overmemory Johannes Berg
2020-12-16 21:28 ` Felix Fietkau
2020-12-17 12:09 ` Toke Høiland-Jørgensen
2020-12-17 11:51 ` Toke Høiland-Jørgensen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=871rfobn8w.fsf@toke.dk \
--to=toke@redhat.com \
--cc=johannes@sipsolutions.net \
--cc=linux-wireless@vger.kernel.org \
--cc=nbd@nbd.name \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).