AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:
CFG:
PT:0
S:andy.sharp@lsi.com
RQ:
SSV:mhbs.lsil.com
NSV:
SSH:
R:<Bill.Fisher@lsi.com>,<Maxim.Kozlovsky@lsi.com>,<Larry.Scheer@lsi.com>,<Rendell.Fong@lsi.com>,<Brian.Stark@lsi.com>
MAID:2
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/LSI/INBOX	0	4B295BF5.6090202@lsi.com
X-Sylpheed-End-Special-Headers: 1
Date: Wed, 16 Dec 2009 14:30:44 -0800
From: Andrew Sharp <andy.sharp@lsi.com>
To: "Fisher, Bill" <Bill.Fisher@lsi.com>
Cc: "Kozlovsky, Maxim" <Maxim.Kozlovsky@lsi.com>, "Scheer, Larry"
 <Larry.Scheer@lsi.com>, "Fong, Rendell" <Rendell.Fong@lsi.com>, "Stark,
 Brian" <Brian.Stark@lsi.com>
Subject: Re: Protposed tuxrx branch changes going forward
Message-ID: <20091216143044.2b86065e@ripper.onstor.net>
In-Reply-To: <4B295BF5.6090202@lsi.com>
References: <4B295865.4020108@lsi.com>
	<861DA0537719934884B3D30A2666FECC010BD20CF9@cosmail02.lsi.com>
	<4B295BF5.6090202@lsi.com>
Organization: LSI
X-Mailer: Sylpheed-Claws 2.6.0 (GTK+ 2.8.20; x86_64-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

The ordering is quite important, just like it is for building a house.
You don't want to put the roof on before the walls go in.  We spent
some time carefully thinking the order through for the reasons that
Bill has stated.

On Wed, 16 Dec 2009 15:15:17 -0700 William Fisher <bill.fisher@lsi.com>
wrote:

> Kozlovsky, Maxim wrote:
> > I don't see the need to establish any particular order to check in
> > the changes. If the changes are not affecting each other, there is
> > no reason for ordering. If they are, that's what p4 integrate is
> > for. Whoever has a change that is ready to be checked in should be
> > able to check in that change without waiting for somebody to do
> > something else. Excessive serialization only slows things down, as
> > everybody knows.
> > 
> > 
> 
> 	Yes, and makes more work for al, the others that follow.
> 
> 	The fact we have been using at least three versions
> 	of the header files is not exactly a clean scheme
> 	to continue.
> 
> 	If you want to check in all of your changes
> 	immediately, then when will they be converted
> 	to the common header file set?
> 
> 	This pushes the work onto the next person,
> 	since the first to check-in wins.
> 
> 	It has taken almost 6 weeks to get JonG's changes
> 	into the tree, lots of them affect the FP code
> 	that needs to be moved forrward. I would
> 	think you would want these changes (point 1)
> 	immediately.
> 
> 	Right now the user level code is not correctly
> 	compiled.
> 
> 	The TXRX module is using a different set of
> 	headers, and the Virtual Server needs
> 	some of the old header files from the
> 	BSD tree.
> 
> -----Original Message-----
> > From: William Fisher [mailto:bill.fisher@lsi.com] 
> > Sent: Wednesday, December 16, 2009 2:00 PM
> > To: Sharp, Andy; Scheer, Larry; Fong, Rendell; Kozlovsky, Maxim;
> > Stark, Brian Cc: Fisher, Bill
> > Subject: Protposed tuxrx branch changes going forward
> > 
> > Folks:
> > 
> > Larry, Andy and I talked about the order of changes
> > we need to make to get everything compiled with
> > the correct set of headers, linked and modules
> > loaded in the kernel and user code.
> > 
> > We are proposing the following order:
> > 
> > 1) Larry Merge of the dev tree containing
> >     JonG's changes has been integrated into
> >     the txrx branch as of yesterday noon'ish.
> >     [Larry] => DONE
> > 
> > 2) There are a number of changes to some
> >     of the common header files, in the
> >     nfx-tree/Includes directories that
> >     need to be merged with those in the
> >     linux/include/onstor/*/*.h directories.
> > 
> >     Larry and I will be doing these changes
> >     into order to resolve the differences
> >     and add those changess to the include/linux/onstor
> >     directories and files.
> >     [Larry and Fisher] => STARTED
> > 
> > 3) The TXRX module itself will be converted to
> >     use the include/linux/onstor header files
> >     for kernel modules. This covers about 8
> >     source files. After this step the
> >     nfs-tree/code/{tpl,netee2,ui} directories
> >     and headers will be deleted since they
> >     will be obsolete and any changes moved
> >     forward.
> > 
> >     I will be doing these changes and verifying
> >     that the kernel networking changes still
> >     compile and work.
> >     [Fisher]
> > 
> > 4) The integration of Max's changes for the
> >     subset of the FP code that he has ported.
> >     They can now convert to use the
> >     common set of kernel header files for
> >     modules, this is a small subset, as
> >     well as those specific to the FP code.
> >     He has a number waiting for go and
> >     has been waiting for the header files
> >     to "stabilize".
> >     [Max]
> > 
> > 5) Fix the header files so that the user level
> >     code compiles and lines with the one's used
> >     for the kernel modules.
> >     Currently this is not been
> >     properly addressed.
> >     [Andy and Larry]
> > 
> > 6) Resolve the issues keeping the TXRX
> >     module from loading as it right now.
> >     This can be done sooner since it
> >     is pretty much independent of the above
> >     items.
> >     [Andy]
> > 
> > 7) Integrate the Virtual Server changes
> >     into the tree using the new protocol
> >     registration procedures Rendell has
> >     developed.
> >     [Rendell]
> > 
> > 8) Resolve the req_tbl.[hc] include issues
> >     so that the NFS/CIFS/NETBIOS protocol
> >     state machine support code can be
> >     cleanly separated from the Virtual Server
> >     code Rendell has been developing.
> >     [ALL]
> > 
> > Let me know if I have missed something or
> > you guys think a different order is better.
> > 
> > Later,
> > 
> > -- Bill
> 
