linux-embedded.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
To: Nicolas Ferre <nicolas.ferre@atmel.com>
Cc: Aras Vaichas <arasv@magellan-technology.com>,
	linux-embedded@vger.kernel.org
Subject: Re: AT91SAM9G20 design and boot times
Date: Fri, 12 Jun 2009 16:06:26 +0200	[thread overview]
Message-ID: <20090612140626.GA10372@game.jcrosoft.org> (raw)
In-Reply-To: <4A30FD17.7060508@atmel.com>

On 14:48 Thu 11 Jun     , Nicolas Ferre wrote:
> Hi,
> 
> Aras Vaichas :
> > Hi,
> > 
> > we're designing our next generation RFID reader and I'm planning on
> > using the AT91SAM9G20 in the next design. We currently use the
> > AT91RM9200.
> 
> Good choice ;-)
> 
> > Can I get some feedback from people with regards to general design
> > configurations  that affect the boot time with respect to this CPU (or
> > the AT91SAM9260)
> 
> I presume that you already know what are the configurations we are using
> at Linux4sam.org on our at91sam9g20ek board. Anyway, I answer your
> questions as it may also interest other readers.
> 
> > Some questions are:
> > 
> > * what is your boot time? What is the time from power-on to kernel
> > running, how long does the kernel take to run until init starts, and
> > how long does init take?
> 
> We have not tried to be clever on speeding up boot process but I guess
> that we can save time:
> - removing unneeded bootloader steps (remove u-boot for instance)
> - removing unused drivers
> - lower timeouts (Ethernet phy detection for example)
> - tailor rootfs exactly to your needs
> 
> That said, here are numbers from our not optimized demo
> (AT91Bootstrap+u-boot+linux+buildroot):
> - ~6s to Linux (can certainly be optimized)
> - ~+2s to init
> - ~+5s to login (here also)
I've done some test on 9263
with the following config
cpu+nor (u-boot+linux+busybox)
~4/5s to the prompt

if the boottime is critical I'll recommend you NOR + XIP

> 
> > * what is your system hardware configuration? CPU+NAND, CPU+NOR+NAND,
> > CPU+DATAFLASH+NAND?
> 
> CPU + NAND can lower your BOM
> 
> > * do you boot everything from the NAND or do you use a combination of
> > Dataflash/NOR and NAND?
I've test it on 9263
> 
> Both.
> 
> > * do you use U-Boot, or a minimal custom bootloader that copies a
> > kernel image from NAND to SDRAM and then executes it?
> 
> If boot time is a hot topic, choose AT91Bootstrap that directly loads Linux.
I disagree on this because you will loose a lot's of possibity as bootcount
DFU, store your kernel in your rootfs etc...

> 
> > * do you mount a small partition of the NAND to begin with and then
> > mount the rest later?
> 
I'll consider to switch to UBIFS
and use lzma compression (for the uImage) if no XIP

Best Regards,
J.

  parent reply	other threads:[~2009-06-12 14:06 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <ed62800906102037q71f2a6bdq835793a1ac36c41b@mail.gmail.com>
2009-06-11  3:43 ` AT91SAM9G20 design and boot times Aras Vaichas
2009-06-11 12:48   ` Nicolas Ferre
2009-06-11 13:58     ` Robin Getz
2009-06-12 14:06     ` Jean-Christophe PLAGNIOL-VILLARD [this message]
2009-06-15  6:32       ` Marc Pignat
2009-06-12 18:37   ` AT91SAM9G20 design and power (was and boot times) Christian Gagneraud

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20090612140626.GA10372@game.jcrosoft.org \
    --to=plagnioj@jcrosoft.com \
    --cc=arasv@magellan-technology.com \
    --cc=linux-embedded@vger.kernel.org \
    --cc=nicolas.ferre@atmel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).