From: Artem Bityutskiy <dedekind1@gmail.com>
To: twebb <taliaferro62@gmail.com>
Cc: "linux-mtd@lists.infradead.org" <linux-mtd@lists.infradead.org>,
Adrian Hunter <adrian.hunter@nokia.com>
Subject: Re: ubifs wear leveling across entire device
Date: Tue, 08 Dec 2009 08:36:47 +0200 [thread overview]
Message-ID: <1260254207.19669.8.camel@localhost> (raw)
In-Reply-To: <dbdb2ea60912071444y149c536bt3106a5ad1da9b221@mail.gmail.com>
On Mon, 2009-12-07 at 17:44 -0500, twebb wrote:
> >
> > Wear-levelling occurs within a single UBI device which is generally
> > a single MTD partition.
> >
> > You need 1 MTD partition and 1 UBI device with 3 volumes. Make the
> > 3 volumes total less than the whole UBI device and you can add a 4th
> > volume later.
> >
>
> If I make 1 MTD partition and 1 corresponding UBI device, should (can)
> it include areas that will never be accessed via UBI-aware tools? For
> example, is it safe to have on a 4GB physical device a single 4GB MTD
> partition with a single corresponding UBI device, but only 3GB of the
> 4GB device contain N UBI volumes only accessed via appropriate UBI
> user-space (or u-boot) tools while the remaining 1GB of the 4GB is
> accessed with mtd-utils?
This contradicts to what you originally asked for. "Give me
wear-levelling across whole device, but do not touch this 1GiB".
> There are two reasons for this approach:
> 1 - to (hopefully) get the benefits of wear leveling across the entire
> physical device (though I think this may not work in the scenario I've
> described)
If you want untouched 1 GiB, make 3GiB + 1GiB partitions, give 3GiB to
UBI.
> 2 - to allow access to a portion of the flash by non-UBI aware code
>
> I assume the alternative (w/o the benefit of wear leveling across the
> entire device) is to define 2 MTD partitions: 1 for "non-UBI" access
> and 1 for a single UBI device with multiple UBI volumes?
Yes.
Also, you can emulate MTD devices on top of UBI volumes, so you can use
many of the MTD tools in linux.
--
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
prev parent reply other threads:[~2009-12-08 6:37 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-12-03 17:31 ubifs wear leveling across entire device twebb
2009-12-04 11:31 ` Adrian Hunter
2009-12-07 15:41 ` twebb
2009-12-07 20:05 ` Adrian Hunter
2009-12-07 22:44 ` twebb
2009-12-08 6:36 ` Artem Bityutskiy [this message]
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=1260254207.19669.8.camel@localhost \
--to=dedekind1@gmail.com \
--cc=adrian.hunter@nokia.com \
--cc=linux-mtd@lists.infradead.org \
--cc=taliaferro62@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.