On Sat, Jun 25, 2011 at 4:20 PM, Randy Dunlap [off-list ref] wrote:

On Sat, 25 Jun 2011 09:35:13 +0300 Artem Bityutskiy wrote:

quoted

On Fri, 2011-06-24 at 13:38 -0700, Randy Dunlap wrote:

quoted

On 06/24/11 13:26, Artem Bityutskiy wrote:

quoted

On Fri, 2011-06-24 at 09:04 -0700, Randy Dunlap wrote:

quoted

On Fri, 24 Jun 2011 15:58:22 +1000 Stephen Rothwell wrote:

quoted

Hi all,

Changes since 20110623:

My fixes tree includes:
        (as an experiment) the early part of mmotm-2011-06-22-13-05
        STAGING: Comedi: Build only on arches providing PAGE_KERNEL_NOCACHE

When CONFIG_MTD=m, CONFIG_MTD_OF_PARTS=y:

drivers/mtd/ofpart.o: In function `init_module':
ofpart.c:(.init.text+0x0): multiple definition of `init_module'
drivers/mtd/mtdcore.o:(.init.text+0x43): first defined here

I guess this is some old issue? I do not think ofpart can be compiled-in
if the mtd core is a module...

[adding driver author, Vitaly]

No, it's a current issue.

I meant that the subject suggests this is about linux-next, but this
problem should be in upstream.

Sorry I misunderstood your comment.

3.0-rc4 does not have this build error when I use the .config file with it...
however, I think that 3.0-rc4 also does not build ofpart.ko as it should.

Looks to me like changes in drivers/mtd/{Kconfig && Makefile} caused this
build error.

The following commit in linux-next appears to be the culprit:

commit 1a6bc893044b961caae1c5a9c6c2065e608da427
Author: Dmitry Eremin-Solenikov [off-list ref]
Date:   Sun May 29 21:32:33 2011 +0400

    mtd: prepare to convert of_mtd_parse_partitions to partition parser

    Prepare to convert of_mtd_parse_partitions() to usual partitions parser:
    1) Register ofpart parser
    2) Internally don't use passed device for error printing
    3) Add device_node to mtd_part_parser_data struct
    4) Move of_mtd_parse_partitions from __devinit to common text section
    5) add ofpart to the default list of partition parsers

    Signed-off-by: Dmitry Eremin-Solenikov [off-list ref]
    Signed-off-by: Artem Bityutskiy [off-list ref]

It adds a module_init() call to ofpart.c which does not work the way
that ofpart.c is linked to mtdcore.o.  Seems like that commit series
still needs work.

g.