netdev
[Top] [All Lists]

Re: [patches] 2.6.0-test9 - r8169 DMA API conversion

To: Brad House <brad@xxxxxxxx>
Subject: Re: [patches] 2.6.0-test9 - r8169 DMA API conversion
From: Brad House <brad@xxxxxxxx>
Date: Wed, 19 Nov 2003 20:36:52 -0500
Cc: Andre Tomt <andre@xxxxxxxx>, netdev@xxxxxxxxxxx, Francois Romieu <romieu@xxxxxxxxxxxxx>, Jeff Garzik <jgarzik@xxxxxxxxx>, Brad House <brad_mssw@xxxxxxxxxx>
In-reply-to: <3FBC1A72.9020101@mcve.com>
References: <47973.68.105.173.45.1069042089.squirrel@mail.mainstreetsoftworks.com> <3FB9A277.70309@pobox.com> <20031118135848.A2451@electric-eye.fr.zoreil.com> <3FBBA76B.4070606@pobox.com> <20031120010056.A19444@electric-eye.fr.zoreil.com> <3FBC05FB.2020409@mcve.com> <20031120014550.A22377@electric-eye.fr.zoreil.com> <3FBC11F2.5080805@mcve.com> <1069292052.3241.36.camel@slurv.pasop.tomt.net> <3FBC1A72.9020101@mcve.com>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030921
ok, misunderstood again ...
geez .... now I get it ...
he'll apply the 1.6 patch this weekend ... _sigh_
I really must be brain dead today.

-Brad

Brad House wrote:
yes, I definately misunderstood then ...
sorry ;)
I'll just apply the patches to my version and see what happens ;)

-Brad

Andre Tomt wrote:

On Thu, 2003-11-20 at 01:59, Brad House wrote:

any reason why it's not against the v1.6 ?



I think you misunderstood. If I intepret the list correctly, the 1.6 changes are in that list (as "your changes (week-end)".)


the current driver is missing support for the
8110S chip (or at least it doesn't seem to work at all),
so you're cutting out all the chips on the mobos out there.
Haven't looked to see if it's just a PCI ID or other functionality,
but since the V1.2 is from realtek that's in the official
kernel, it would make sense to use the 1.6 patch....



IIRC, the realtek one got cleaned up a great deal before entering mainline, both bug-fixes and coding style cleanups. Your patch may revert a lot of that work (I havn't looked at it closely.) Also the changelog differs for the 1.2 entry between the two versions, indicating just this.

Lets take one step at a time ;-)


It is against the original r8169. Once finished, it should look like:
- first cut at dma api conversion (today)
- remaining bits of dma api conversion (tomorrow)
- rx_copybreak (tomorrow)
- your changes (week-end)
- big-endian fixes (week-end)

Each part divided in reasonably small patches to ease the test/review process.









<Prev in Thread] Current Thread [Next in Thread>