[PATCH] UBI: take 2

[PATCH] UBI: take 2

[Artem Bityutskiy]

This is take 2 of the previous mail with David's comments in mind.

Hello Andrew,

we have announced UBI several months ago in the MTD mailing list. It was
successfully used in our setup and we've got positive feedback.

In short, it is kind of LVM layer but for flash (MTD) devices which
hides flash devices complexities like bad eraseblocks (on NANDs) and
wear. The documentation is available at the MTD web site:
http://www.linux-mtd.infradead.org/doc/ubi.html
http://www.linux-mtd.infradead.org/faq/ubi.html

The source code is available at the UBI GIT tree:
git://git.infradead.org/ubi-2.6.git

The UBI GIT tree is based upon the mtd-2.6.git
(git://git.infradead.org/mtd-2.6.git)

There is also an 'mtd' branch available in the UBI GIT which contains
the current mtd-2.6.git. So that you can use git-diff mtd and have UBI
patches.

There is also web interface to the git trees available:
UBI: http://git.infradead.org/?p=ubi-2.6.git;a=summary
MTD: http://git.infradead.org/?p=mtd-2.6.git;a=summary

Plain patches may be found at
http://linux-mtd.infradead.org/~dedekind/ubi/

Please, include the patches to your tree.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: [PATCH] UBI: take 2

[Andrew Morton]

On Wed, 29 Nov 2006 20:17:26 +0200
Artem Bityutskiy <dedekind@infradead.org> wrote:

> This is take 2 of the previous mail with David's comments in mind.
> 
> Hello Andrew,
> 
> we have announced UBI several months ago in the MTD mailing list. It was
> successfully used in our setup and we've got positive feedback.
> 
> In short, it is kind of LVM layer but for flash (MTD) devices which
> hides flash devices complexities like bad eraseblocks (on NANDs) and
> wear. The documentation is available at the MTD web site:
> http://www.linux-mtd.infradead.org/doc/ubi.html
> http://www.linux-mtd.infradead.org/faq/ubi.html
> 
> The source code is available at the UBI GIT tree:
> git://git.infradead.org/ubi-2.6.git

Got that, thanks.  It needs a bit of help:

--- a/drivers/mtd/ubi/cdev.c~git-ubi-fix
+++ a/drivers/mtd/ubi/cdev.c
@@ -1185,7 +1185,7 @@ static ssize_t vol_cdev_direct_write(str
 			 len, vol_id, lnum, off);
 
 		err = ubi_eba_write_leb(ubi, vol_id, lnum, tbuf, off, len,
-					UBI_DATA_UNKNOWN, &written, 0, NULL);
+					UBI_DATA_UNKNOWN, &written, NULL);
 		if (unlikely(err)) {
 			count -= written;
 			*offp += written;
_

Re: [PATCH] UBI: take 2

[Josh Boyer]

On Thu, 2006-11-30 at 17:06 -0800, Andrew Morton wrote:
> >
> > The source code is available at the UBI GIT tree:
> > git://git.infradead.org/ubi-2.6.git
> 
> Got that, thanks.  It needs a bit of help:
> 
> --- a/drivers/mtd/ubi/cdev.c~git-ubi-fix
> +++ a/drivers/mtd/ubi/cdev.c
> @@ -1185,7 +1185,7 @@ static ssize_t vol_cdev_direct_write(str
>  			 len, vol_id, lnum, off);
> 
>  		err = ubi_eba_write_leb(ubi, vol_id, lnum, tbuf, off, len,
> -					UBI_DATA_UNKNOWN, &written, 0, NULL);
> +					UBI_DATA_UNKNOWN, &written, NULL);
>  		if (unlikely(err)) {
>  			count -= written;
>  			*offp += written;
> _

Nice catch.  Odd that gcc doesn't throw a warning on my system with it
being like that.

  CC      drivers/mtd/ubi/badeb.o
  CC      drivers/mtd/ubi/upd.o
  CC      drivers/mtd/ubi/sysfs.o
  CC      drivers/mtd/ubi/cdev.o
  CC      drivers/mtd/ubi/uif.o
  CC      drivers/mtd/ubi/vtbl.o
  CC      drivers/mtd/ubi/volmgmt.o
  CC      drivers/mtd/ubi/eba.o

Anyway, we'll get the patch into the tree ASAP.

josh

Re: [PATCH] UBI: take 2

[Artem Bityutskiy]

Thank you Andrew,

On Thu, 2006-11-30 at 22:24 -0600, Josh Boyer wrote:
> On Thu, 2006-11-30 at 17:06 -0800, Andrew Morton wrote:
> >  		err = ubi_eba_write_leb(ubi, vol_id, lnum, tbuf, off, len,
> > -					UBI_DATA_UNKNOWN, &written, 0, NULL);
> > +					UBI_DATA_UNKNOWN, &written, NULL);
> Nice catch.  Odd that gcc doesn't throw a warning on my system with it
> being like that.

Because this piece of code is for debugging purposes only and it is
usually disabled. You probably haven't ever enabled it.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: [PATCH] UBI: take 2

[Artem Bityutskiy]

Hello Andrew,

On Thu, 2006-11-30 at 17:06 -0800, Andrew Morton wrote:
> --- a/drivers/mtd/ubi/cdev.c~git-ubi-fix
> +++ a/drivers/mtd/ubi/cdev.c
> @@ -1185,7 +1185,7 @@ static ssize_t vol_cdev_direct_write(str
>  			 len, vol_id, lnum, off);
>  
>  		err = ubi_eba_write_leb(ubi, vol_id, lnum, tbuf, off, len,
> -					UBI_DATA_UNKNOWN, &written, 0, NULL);
> +					UBI_DATA_UNKNOWN, &written, NULL);

I have re-based the ubi-2.6.git tree, my apologies for this. I hope I
have not caused you many troubles by this.

I also I have applied your fix.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)