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>
MAID:2
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/LSI/INBOX	0	4AEF5EA2.5050005@lsi.com
X-Sylpheed-End-Special-Headers: 1
Date: Mon, 2 Nov 2009 15:27:44 -0800
From: Andrew Sharp <andy.sharp@lsi.com>
To: William Fisher <bill.fisher@lsi.com>
Subject: Re: kernel build - please review 33334
Message-ID: <20091102152744.6ab082b1@ripper.onstor.net>
In-Reply-To: <4AEF5EA2.5050005@lsi.com>
References: <20091102140522.5f6eabd2@ripper.onstor.net>
	<4AEF5EA2.5050005@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

OK,

I removed acpu.c from the changelist, and it still builds, so we're
good on that.

Please send me some specifics about what you want me to do with what
comments, and then I think it can go in if you concur.

Cheers,

a


On Mon, 2 Nov 2009 15:35:14 -0700 William Fisher
<bill.fisher@lsi.com> wrote:

> Andrew Sharp wrote:
> > Hi Bill,
> > 
> > I have a fix for some of the issues with the kernel build in this
> > changelist, so I'd like to get a review from you.
> > 
> > I believe you have two issues you would like to bring up/have me
> > address, etc.
> > 
> > One is ... I don't remember, something about comments.  Please tell
> > me which comments, where, and what you don't like and what it
> > should look like after I fix it.  Be brief.  Be technical.
> > 
> 
> 	I think the changes to move the mgmtbus initialization
> 	code to some architectural specific directory are fine.
> 
> 	I think we should rename all CONFIG_ONSTOR_XXX symbols to
> 	CONFIG_LSI_XXX once and for all and not have some
> 	combination of them. I don't care if they are TXRX
> 	versus TUXSTOR but we need to do it only once.
> 
> 	I will send you the exact details.
> 	
> > The other is the acpu.  Please tell me what you want done on that.
> 
> 	For the acpu code, leave what is there today for the time
> 	being. That will allow the rcon shell to work with the
> 	net/tpl, net/neteee2 and net/neteee-ui code that I will
> 	be sending you.
> 
> 	The next step will be to add support for both ACPU
> 	and NCPU threads. The NCPU "thread" will simply
> 	read the ATON IPC queue to handle the TPL messages
> 	sent to the "logical NCPU". This will allow the thread
> 	in the kernel to do the socket operation, and when that
> 	gets done, to do the call-back operation. This is
> 	critical for both the FP and ACPU code to work
> 	using the TPL interfaces. The TPL implementation
> 	is now strict Linux socket operations so at least
> 	the BSD stack is gone.
> 
> 	Hence, we need two threads on the TXRX side, an ACPU
> 	thread for CIFS/NFS/NETBIOS and a NCPU thread for the
> 	IPC messaging from the ACPU. The ACPU code is
> 	full of calls to IPC[ATON] queue operations.
> 	The FP code is full of eee_send_message() calls to
> 	send messages to the NCPU for networking operations.
> 	
> 	Hence there are TWO methods that must be supported
> 	here, the EEE protocol and the IPC scheme.
> 
> 	This design
> 	allows us to move the FP code over wholesale without
> 	breaking it's TPL view of the world.
> 
> 	The fact that none of this code can "block" is the
> 	limitation with the NFS/CIFS/NETBIOS design approach.
> 
> 	The NDMP is also completely in bed with TPL and
> 	all those call-backs. This is the path to get
> 	the stuff running ASAP.
> > 
> > There remains one more thing: neteee/neteee.c builds with 2
> > undefined symbols: eee_rcv_pkt and eee_txrxIsPktToNcpu_kernel.  I'd
> > like to fix that in this changelist -- if possible -- but I'd like
> > a recommendation from you on how to handle it.  Can I comment the
> > function calls out for now?  Is there something better I should do?
> 
> 	I will resend you the message I sent that resolves these
> 	symbols.
> > 
> > If I can get this changelist squared away today or tomorrow we can
> > claim at least for a few days that the kernel builds, and then
> > maybe get Max out of our hair.
> > 
> As soon after I send you the one we discussed this morning.
> 
> -- Bill
> 
> 
> 
