Re: [memcg:since-3.5 41/99] (.text+0xa0998): undefined reference to `get_kernel_page'

[Michal Hocko <mhocko@suse.cz>]

On Tue 21-08-12 00:28:13, Wu Fengguang wrote:
> On Mon, Aug 20, 2012 at 05:29:03PM +0200, Michal Hocko wrote:
[...]
> > compiles without any issues, right?
> 
> It have some other issues instead. You didn't receive emails about
> them because my system only complaint about "new" issues :-)

OK
> 
> In file included from /c/wfg/linux/arch/m32r/boot/compressed/misc.c:25:0:
> /c/wfg/linux/arch/m32r/boot/compressed/m32r_sio.c:11:13: warning: conflicting types for built-in function 'putc' [enabled by default]
>   MODPOST 111 modules
> ERROR: "memory_start" [fs/udf/udf.ko] undefined!
> ERROR: "memory_end" [fs/udf/udf.ko] undefined!
> ERROR: "memory_end" [drivers/scsi/sg.ko] undefined!
> ERROR: "memory_start" [drivers/scsi/sg.ko] undefined!
> ERROR: "memory_end" [drivers/i2c/i2c-dev.ko] undefined!
> ERROR: "memory_start" [drivers/i2c/i2c-dev.ko] undefined!
>   GZIP    arch/m32r/boot/compressed/vmlinux.bin.gz
> make[2]: *** [__modpost] Error 1
> make[1]: *** [modules] Error 2
> make[1]: *** Waiting for unfinished jobs....
>   LD      arch/m32r/boot/compressed/piggy.o
> /usr/local/gcc-4.6.3-nolibc/m32r-linux/bin/m32r-linux-ld: target elf32-m32r not found

I assume those are present in the mmotm as well.

> > > Btw, is it meaningful for me to test the bisectability of all the
> > > memcg branches? If not, you may give me a list or pattern for me to
> > > take care.
> > 
> > Well, I do not think that this kind of test makes sense for the tree
> > because fixups are always done as a separate commits to prevent from
> > rebasing. Andrew always folds the fixups into the final patch which is
> > sent to Linus or who ever should push it further.
> 
> Yeah the build script have taken care of these follow-up fixes:
> 
> has_akpm_fix()
> {       
>         [[ $branch = 'next/akpm' || $branch =~ 'memcg/' ]] || return 1
>         
>         [[ $(git_commit_subject $commit) =~ (fix|build|error|warning) ]] && return 0
>         [[ $(git_commit_changelog $commit) =~ (build|compile|gcc|sparse|error|warning) ]] && return 0
>         
>         next_commit=$(echo "$*" | sed -ne "$((iterated_commits+2))p")
>         [[ $(git_commit_subject $next_commit) =~ (fix|build|error|warning) ]] && return 0
>         [[ $(git_commit_changelog $next_commit) =~ (build|compile|gcc|sparse|error|warning) ]] && return 0
>         return 1
> }       

OK, but I still think that the tree is not constructed to be bisectable.
It's main purpose is the provide a common code base for -mm tree which
is easier to work with than linux-next. The compilation coverage is
really great and thanks a lot for doing that but I think that
bisectability testing is not needed.

Thanks
-- 
Michal Hocko
SUSE Labs