netdev - [PATCH] [sis900] convert to NAPI, WAS Re: pktgen terminating condition

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: <20070904032036.GA11153@ludhiana>
Date:	Mon, 3 Sep 2007 20:20:36 -0700
From:	Mandeep Singh Baines <mandeep.baines@...il.com>
To:	Daniele Venzano <venza@...wnhat.org>
Cc:	hadi@...erus.ca, Mandeep Baines <mandeep.baines@...il.com>,
	davem@...emloft.net, rick.jones2@...com, msb@...gle.com,
	netdev@...r.kernel.org, grundler@...gle.com,
	robert.olsson@....uu.se, jeff@...zik.org, nhorman@...driver.com
Subject: [PATCH] [sis900] convert to NAPI, WAS Re: pktgen terminating
	condition

Hi Daniele,

Attached is a patch for converting the sis900 driver to NAPI. Please take a
look at let me know what you think. I'm not 100% sure if I'm handling the 
rotting packet issue correctly or that I have the locking right in tx_timeout.
These might be areas to look at closely.

I didn't see much saving in interrupts on my machine (too fast, I guess). But
I still think its beneficial: pushing work out of the interrupt handler into
a bottom half is a good thing and we no longer need to disable interrupts 
in start_xmit. 

I did see a significant boost to tx performance by optimizing start_xmit: more
than double pps in pktgen.

I'm also attaching some test results for various iterations of development.

Regards,
Mandeep

Daniele Venzano (venza@...wnhat.org) wrote:
> ----- Message d'origine -----
> De: jamal <hadi@...erus.ca>
> Date: Fri, 31 Aug 2007 09:50:03 -0400
> Sujet: Re: Re: pktgen terminating condition
> 
> >I dont know if you followed the discussion - by defering the freeing of
> >skbs, you will be slowing down socket apps sending from the local
> >machine. It may be ok if the socket buffers were huge, but that comes at
> >the cost of system memory (which may not be a big deal)
> >Do you by any chance recall why you used the idle interupt instead of
> >txok to kick the prunning of tx descriptors?
> 
> That should be asked to the original author of the driver, that I wasn't able to contact when I took over the maintainership several years ago.
> I think that since this chip is usually used on cheap/low performance (relatively speaking) devices it was felt that generating an interrupt for each transmitted packet was going to affect the performance of the system too much. At the start, if I remember correctly, this was a chip thought for consumer use, where transmissions won't happen continuously for long periods of time. Then the sis900 started to get used everywhere, from embedded systems to (cheap) server motherboards and the usage scenario changed.
> 
> 
> --
> Daniele Venzano
> venza@...wnhat.org

View attachment "sis900_testing.txt" of type "text/plain" (1296 bytes)

View attachment "sis900_napi.patch" of type "text/plain" (19791 bytes)