AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:<20081003160636.3e696e63@ripper.onstor.net>
CFG:
PT:0
S:andy.sharp@onstor.com
RQ:
SSV:onstor-exch02.onstor.net
NSV:
SSH:
R:<bfisher@onstor.com>
MAID:1
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/andys@onstor.net@onstor-exch02.onstor.net/INBOX	0	BB375AF679D4A34E9CA8DFA650E2B04E0BCFED29@onstor-exch02.onstor.net
X-Sylpheed-End-Special-Headers: 1
Date: Fri, 3 Oct 2008 16:06:45 -0700
From: Andrew Sharp <andy.sharp@onstor.com>
To: Bill Fisher <bfisher@onstor.com>
Subject: Re: Latest prom Load numbers
Message-ID: <20081003160645.074f7645@ripper.onstor.net>
In-Reply-To: <BB375AF679D4A34E9CA8DFA650E2B04E0BCFED29@onstor-exch02.onstor.net>
References: <48E56D6D.5050309@onstor.com>
	<BB375AF679D4A34E9CA8DFA650E2B04E0BCFED29@onstor-exch02.onstor.net>
Organization: Onstor
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=UTF-8
Content-Transfer-Encoding: quoted-printable

Bill,

That means that we're giving up on getting any more performance, and we
just need to wrap up what you've got so far and move on to the next
project.

Speaking of which, how's the git learning progressing?  You can get
git rpm's for your workstation, and/or you can just slogin to ripper
and use git there.

Regardless of which option you follow, you can get the latest
2.6.2[23456] linux-mips source with the following command:

git clone http://ripper/git/linux-mips-2.6 <directory>

man git-clone should give you all the documentation you need for that
command.

The directory should be in your home directory so you can access it
anywhere.  That will create your own repository which you can work in.
The various releases are contained in that repo as branches, so you can
check out whatever branch you want.

Let me know if you have any questions.

Cheers,

a

On Fri, 3 Oct 2008 15:42:03 -0700 "Brian Stark"
<brian.stark@onstor.com> wrote:

> Bill,
>=20
> I tested out Unigen compact flash with your PROM changes, and it
> loaded up just fine.  The load time is slightly shorter than Wintec
> -- 32.6 sec vs 33.5 sec. =20
>=20
> Let's plan for a release on Monday.  Please send Warren the relevant
> changelist so that he can do the code review.  Once this is done, the
> changes can be checked in and Warren will then do the official
> build.  He has a utility that computes the file checksums and then
> creates images with the version and checksum in the filename.  This
> is helpful for Ops to know that they have the correct version when
> ordering programmed parts from the CM.
>=20
> Once built, Ops will then do focused testing and if all is ok,
> they'll then order programmed parts.  We'll also make the image
> available for machines in the lab.
>=20
>=20
> Brian
>=20
>=20
>=20
> -----Original Message-----
> From: Bill Fisher=20
> Sent: Thursday, October 02, 2008 5:55 PM
> To: Brian Stark
> Cc: Bill Fisher; Andy Sharp; Warren Gale
> Subject: Re: Latest prom Load numbers
>=20
> Brian Stark wrote:
>=20
> >I got the following when loading optimized images:
> >
> >diag menu cmd: 55
> >  loop count: 1
> >Read files from both Compact Flash cards (CF)
> >
> >Diag Read files from dev: sda1
> >ext2_load_file: /dev/sda1/boot/vmlinux.bin, location
> >0xffffffff82000000 disk model: CF
> >1GB                                  =E2=98=BA disk geometry: cylinders=
=3D2044
> >heads=3D16 sectors=3D63 ext2fs_file_read: file * 0xffffffff80d1e020,
> >size 3079776(0x2efe60), location 0xffffffff82000000 Loading time
> >dev: sda1: time msecs: 7951
> >
> >ext2_load_file: /dev/sda1/boot/txrx_cg.bin, location 0x42000000
> >disk model: CF 1GB                                  =E2=98=BA
> >disk geometry: cylinders=3D2044 heads=3D16 sectors=3D63
> >ext2fs_file_read: file * 0xffffffff80d1d620, size 7748832(0x763ce0),
> >location 0x42000000 Loading time dev: sda1: time msecs: 14705
> >
> >ext2_load_file: /dev/sda1/boot/fp_cg.bin, location 0x44000000
> >disk model: CF 1GB                                  =E2=98=BA
> >disk geometry: cylinders=3D2044 heads=3D16 sectors=3D63
> >ext2fs_file_read: file * 0xffffffff80d1cc20, size 6105128(0x5d2828),
> >location 0x44000000 Loading time dev: sda1: time msecs: 10895
> >
> >
> >
> >Total time is 33.5 seconds.  The previous PROM was taking 1 minute
> >15 seconds to load these same images. =20
> >
> >
> > =20
> >
> After more looking, the wdc prom  code does a special 16-bit read in
> the "interrupt" handlers when reading
> or writing "data" from/to the control register, hence I previously=20
> mis-spoke.
>=20
> Everything else uses the 8-Bit register operations except this
> special case. Hence it isn't clear if we can
> trivially get another 2X read improvement. I will have to start
> timing various procedures to see where the
> next batch of low hanging fruit can be harvested.
>=20
> When do you want to release these? Tomorrow, after we at least try the
> other CF vendor, right? We will run these through some internal
> QA testing cycle, right, like loading them onto several machines and
> doing "golden" master bits testing, right?
>=20
> Later,
>=20
> -- Bill
>=20
