netdev - Re: Question on Lockless TX

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] [day] [month] [year] [list]

Message-ID: <CAJzFV351AoqC6_i0fz1yMNOqCS146issDhcN5D7Gkd4zYC3riA@mail.gmail.com>
Date:	Fri, 18 Apr 2014 14:32:53 -0600
From:	Sharat Masetty <sharat04@...il.com>
To:	Tom Herbert <therbert@...gle.com>
Cc:	Linux Netdev List <netdev@...r.kernel.org>
Subject: Re: Question on Lockless TX

I think this flag will work for our case, but in the documentation
section it says
"Note that the use of NETIF_F_LLTX is deprecated.  Don't use it for
new drivers." Trying to dig into this a little more but not sure why
this deprecated and should not be used.. can you shed some light here?
Are there any known issues with this flag?

I did see this thread from a while back:
https://lwn.net/Articles/121566/ but issue seems to be fairly
contained..

Regards,
Sharat


On Wed, Apr 16, 2014 at 10:51 AM, Tom Herbert <therbert@...gle.com> wrote:
> On Tue, Apr 15, 2014 at 9:29 PM, Sharat Masetty <sharat04@...il.com> wrote:
>> Can someone explain to me, why would a driver choose to implement
>> Lockless TX? What are the advantages/disadvantages of choosing to do
>> so?Any performance benefits? A reference to a sample driver
>> implementation would be helpful to understand.
>>
> Locks in the TX path are needed provide mutual exclusion to the
> underlying HW queue and other structures. Some drivers either don't
> need any form of locking because their is no queue to protect (often
> the case with virtual drivers like loopback and bonding), others might
> provide their own locking at higher granularity or for other purposes.
> Generally, taking fewer locks can reduce lock contention, atomic
> operations, and cache line misses. If you grep for NETIF_F_LLTX under
> drivers/net you can find examples of drivers that support this. Also,
> please look at Documentation/networking/netdevices.txt.
>
> Tom
>
>> Thanks
>> Sharat
>> --
>> To unsubscribe from this list: send the line "unsubscribe netdev" in
>> the body of a message to majordomo@...r.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html