David,
The following changes since commit
f3073ac76755abd63b1d4c3d145f4c15b65b5355 (i.e., your netdev-2.6):
Guennadi Liakhovetski (1): smc911x: Fix external PHY detection
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/chris/linux-2.6.git for-david
Summary:
a) generic HDLC includes that newer PPP code, and now, with all
drivers ported, syncppp can finally be removed
b) drivers for cards based on Hitachi/Renesas SCA and SCA-II chips are
split and reworked. Means nothing or almost nothing to SCA-based
cards (= ISA etc), and some updates including NAPI to PCI cards
based on SCA-II (their are low speed but NAPI saves them when the
card is being fed with a continuous stream of errors).
c) the relevant part of Wang Chen's dev->priv conversion batch
Posted on netdev, in "next" for months, all reported probles fixed.
Please pull, TIA.
Krzysztof Ha?asa (15):
WAN: split hd6457x.c into hd64570.c and hd64572.c
WAN: remove SCA II support from SCA drivers
WAN: remove SCA support from SCA-II drivers
WAN: convert HD64572-based drivers to NAPI.
WAN: TX-done handler now uses the ownership bit in HD64572 drivers.
WAN: HD64572 already handles TX underruns with DMAC.
WAN: rework HD64572 interrupts a bit.
WAN: Simplify HD64572 status handling.
WAN: don't print HD64572 driver versions anymore.
WAN: Simplify HD64572 drivers.
WAN: HD64572 drivers don't use next_desc() anymore.
WAN: Correct comments in hd6457[02].c
WAN: Simplify sca_init_port() in HD64572 driver.
WAN: new synchronous PPP implementation for generic HDLC.
WAN: syncppp.c is no longer used by any kernel code. Remove it.
Wang Chen (2):
netdevice hdlc: Convert directly reference of netdev->priv to netdev->ml_priv
netdevice pc300: Convert directly reference of netdev->priv to netdev->ml_priv
Documentation/DocBook/Makefile | 2 +-
Documentation/DocBook/networking.tmpl | 3 -
Documentation/DocBook/wanbook.tmpl | 99 --
Documentation/networking/generic-hdlc.txt | 8 +-
drivers/net/wan/Makefile | 2 +-
drivers/net/wan/c101.c | 6 +-
drivers/net/wan/{hd6457x.c => hd64570.c} | 254 ++----
drivers/net/wan/hd64572.c | 641 +++++++++++++
drivers/net/wan/hdlc_fr.c | 10 +-
drivers/net/wan/hdlc_ppp.c | 648 ++++++++++++-
drivers/net/wan/n2.c | 9 +-
drivers/net/wan/pc300_drv.c | 17 +-
drivers/net/wan/pc300too.c | 73 +-
drivers/net/wan/pci200syn.c | 77 +-
drivers/net/wan/syncppp.c | 1479 -----------------------------
include/linux/hdlc.h | 2 +-
include/net/syncppp.h | 102 --
17 files changed, 1372 insertions(+), 2060 deletions(-)
delete Documentation/DocBook/wanbook.tmpl
rename drivers/net/wan/{hd6457x.c => hd64570.c} (78%)
create drivers/net/wan/hd64572.c (hd6457x.c split into these two)
delete drivers/net/wan/syncppp.c
delete include/net/syncppp.h
--
Krzysztof Halasa
From: Krzysztof Halasa <[email protected]>
Date: Sun, 12 Oct 2008 23:12:16 +0200
> David,
>
> The following changes since commit
> f3073ac76755abd63b1d4c3d145f4c15b65b5355 (i.e., your netdev-2.6):
> Guennadi Liakhovetski (1): smc911x: Fix external PHY detection
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/chris/linux-2.6.git for-david
You've missed the networking merge window, as I stated the other
day I'm not taking any more networking changes.
You should have merged this in earlier.
--------------------
Subject: preparing net-next-2.6 for merge...
From: David Miller <[email protected]>
To: [email protected]
CC: [email protected]
Date: Fri, 10 Oct 2008 12:29:37 -0700 (PDT)
X-Mailer: Mew version 6.1 on Emacs 22.1 / Mule 5.0 (SAKAKI)
I'm doing some quick merge, build, and boot tests on the tree as it
stands right now then I'll send a pull request to Linus.
I'm sure there are some things people still wanted to get into the
tree, but let's try to follow the rules this time. Only bug fixes
and corrections at this point forward.
Look at it this way, I have tons of stuff I wasn't able to finish up
in time either, so both the rule and the pain and being applied fairly
:-)
Anyways, just FYI...
David Miller <[email protected]> writes:
> You've missed the networking merge window, as I stated the other
> day I'm not taking any more networking changes.
>
> You should have merged this in earlier.
I was under impression the window is still two weeks long, isn't it?
Anyway, it doesn't matter - I understand I have to resend this request
in two weeks, for "next"? Sure.
--
Krzysztof Halasa
On Sun, Oct 12, 2008 at 02:47:04PM -0700, David Miller wrote:
> From: Krzysztof Halasa <[email protected]>
> Date: Sun, 12 Oct 2008 23:12:16 +0200
>
> > David,
> >
> > The following changes since commit
> > f3073ac76755abd63b1d4c3d145f4c15b65b5355 (i.e., your netdev-2.6):
> > Guennadi Liakhovetski (1): smc911x: Fix external PHY detection
> >
> > are available in the git repository at:
> >
> > git://git.kernel.org/pub/scm/linux/kernel/git/chris/linux-2.6.git for-david
>
> You've missed the networking merge window, as I stated the other
> day I'm not taking any more networking changes.
>
> You should have merged this in earlier.
>...
Ignoring whether it will go into 2.6.28 or 2.6.29 there seems to be a
more fundamental issue:
The over 3 weeks old -next already contains mostly the same as the HDLC
tree contains now.
The thing you two should sort out is why the tree ended up as an own
tree in -next instead of pull requests to David long ago.
E.g. the netdev and wireless trees also don't go into -next, but instead
with frequent pull requests to David.
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
Adrian Bunk <[email protected]> writes:
> The over 3 weeks old -next already contains mostly the same as the HDLC
> tree contains now.
>
> The thing you two should sort out is why the tree ended up as an own
> tree in -next instead of pull requests to David long ago.
>
> E.g. the netdev and wireless trees also don't go into -next, but instead
> with frequent pull requests to David.
I can't see any problem with this, is there any?
I want the HDLC tree tested in next before it's merged upstream
(= David's tree(s)). I don't want to push crap upstream first and
only then try to fix it (though a bug still slips in sometimes).
--
Krzysztof Halasa
From: Krzysztof Halasa <[email protected]>
Date: Mon, 13 Oct 2008 16:53:17 +0200
> Adrian Bunk <[email protected]> writes:
>
> > The over 3 weeks old -next already contains mostly the same as the HDLC
> > tree contains now.
> >
> > The thing you two should sort out is why the tree ended up as an own
> > tree in -next instead of pull requests to David long ago.
> >
> > E.g. the netdev and wireless trees also don't go into -next, but instead
> > with frequent pull requests to David.
>
> I can't see any problem with this, is there any?
>
> I want the HDLC tree tested in next before it's merged upstream
> (= David's tree(s)). I don't want to push crap upstream first and
> only then try to fix it (though a bug still slips in sometimes).
You're supposed to have your changes merged into the relevant
subsystem tree long before Linus's merge window opens up.
That's how it is meant to work.
What I'm doing with networking is nothing special.
So please be punctual and merge early and often to Jeff and
myself, so that you don't miss the merge window in the future.