[uml-devel] uml-2.6.8-rc2 (2.6.8-rc2 w/ uml-patch-2.6.7-01)

[uml-devel] uml-2.6.8-rc2 (2.6.8-rc2 w/ uml-patch-2.6.7-01)

[gboutwel]

  I'm testing the 2.6.8 waters. Looks like it's not been included

in the rc2 version yet. It's in the mm1 version though... so

there's hope. Here's what I've done:



1) tar jzf linux-2.6.7.tar.bz2



2) mv linux-2.6.7 linux-2.6.8-rc2-1um



3) bzcat patch-2.6.8-rc2.bz2 | patch -p1



4) bzcat uml-patch-2.6.7-01.bz2 | patch -p1 --dry-run | less



* not bad, just shows 1 reject, which is the removal of an

#include from init_setup.c



5) cp ~/2.6.7/config-2.6.7-1um .config



6) make oldconfig ARCH=um



7) make linux ARCH=um



* this fails on arch/um/drivers/stdio_console.c with:

arch/um/drivers/stdio_console.c:207: error: conflicting types

for `console_device'

include/linux/console.h:107: error: previous declaration of `console_device'



Where:



include/linux/console.h:107 extern struct tty_driver *console_device(int

*);

arch/um/drivers/stdio_console.c:207 has struct tty_driver *console_device(struct

console *c, int *index)



so, on a wild 'whim' I changed console_device in stdio_console.c

to uml_console_device.



That compiles, but then it ends with an linker error:



vmlinux(.text+0x2d8b0): In function `alloc_pidmap':

: undefined reference to `find_next_zero_bit'

vmlinux(.text+0x4faef): In function `get_unused_fd':

: undefined reference to `find_next_zero_bit'

vmlinux(.text+0x5fcd8): In function `locate_fd':

: undefined reference to `find_next_zero_bit'

vmlinux(.text+0x7e3f1): In function `scan_bitmap_block':

: undefined reference to `find_next_zero_bit'

vmlinux(.text+0x9f49e): In function `reiserfs_in_journal':

: undefined reference to `find_next_zero_bit'

vmlinux(.text+0xa6bd0): more undefined references to `find_next_zero_bit'

follow

vmlinux(.text+0xb7905): In function `sub_alloc':

: undefined reference to `find_next_bit'

vmlinux(.init.text+0x43d0): In function `__alloc_bootmem_core':

: undefined reference to `find_next_zero_bit'

collect2: ld returned 1 exit status

make: *** [linux] Error 1



Probably something simple, but I wanted to discuss with folks

more in the now on if the above (stdio_console) fix was the right

one, and to find out if anyone else has an idea of how to fixe

the find_next_zero_bit error.



Thanks,

George







----------------------------------------------------

Need to store your files online? 

Upload them for free in Praize Briefcase

http://www.praize.com/cgi-bin/members/mybriefcase.cgi



-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] uml-2.6.8-rc2 (2.6.8-rc2 w/ uml-patch-2.6.7-01)

[David T Hollis]

On Fri, 2004-07-23 at 21:43 +0000, gboutwel wrote:
> vmlinux(.text+0x5fcd8): In function `locate_fd':
> 
> : undefined reference to `find_next_zero_bit'
> 
> vmlinux(.text+0x7e3f1): In function `scan_bitmap_block':
> 
> : undefined reference to `find_next_zero_bit'
> 
> vmlinux(.text+0x9f49e): In function `reiserfs_in_journal':
> 
> : undefined reference to `find_next_zero_bit'
> 
> vmlinux(.text+0xa6bd0): more undefined references to `find_next_zero_bit'
> 
> follow
> 
> vmlinux(.text+0xb7905): In function `sub_alloc':
> 
> : undefined reference to `find_next_bit'
> 
> vmlinux(.init.text+0x43d0): In function `__alloc_bootmem_core':
> 
> : undefined reference to `find_next_zero_bit'
> 
> collect2: ld returned 1 exit status
> 
> make: *** [linux] Error 1
> 
> 
> 
> Probably something simple, but I wanted to discuss with folks
> 
> more in the now on if the above (stdio_console) fix was the right
> 
> one, and to find out if anyone else has an idea of how to fixe
> 
> the find_next_zero_bit error.
> 

I ran into this around a week or so ago as well.  Unfortunately, I don't
have a patch but here is the rough idea of what I did to work around it.
find_next_zero_bit is defined in arch/i386/lib/bitops.c.  If you look in
arch/um/sys-i386, a couple files get symlinked from other parts of the
tree (see the Makefile).  If you add bitops.c to this, you will get past
the find_next_zero_bit problem.  You would add bitops.o to the obj-y,
bitops.c to SYMLINKS, bitops.c-dir := arch/i386/lib.  I think thats
about it.

Good luck
-- 
David T Hollis <dhollis@davehollis.com>



-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] uml-2.6.8-rc2 (2.6.8-rc2 w/ uml-patch-2.6.7-01)

[BlaisorBlade]

Alle 05:38, sabato 24 luglio 2004, David T Hollis ha scritto:
> On Fri, 2004-07-23 at 21:43 +0000, gboutwel wrote:
> >
> > Probably something simple, but I wanted to discuss with folks
> > more in the now on if the above (stdio_console) fix was the right
> > one, and to find out if anyone else has an idea of how to fixe
> > the find_next_zero_bit error.

Yes, I saw that error in -mm and I agree that renaming is right; about the 
other one, the reported fix (below) is correct; also, it would be nice to 
delete include/asm-um/bitops.h (which includes the i386 version of the 
include; since those funcs were moved to out-of-line ones, the trick must be 
done that way).

Note however that we must not hardcode that we steel things from i386 arch, 
but we must use $(SUBARCH) - or it will break for any other archs (the only 
working one, at the moment, is the x86_64 one).


> I ran into this around a week or so ago as well.  Unfortunately, I don't
> have a patch but here is the rough idea of what I did to work around it.
> find_next_zero_bit is defined in arch/i386/lib/bitops.c.  If you look in
> arch/um/sys-i386, a couple files get symlinked from other parts of the
> tree (see the Makefile).  If you add bitops.c to this, you will get past
> the find_next_zero_bit problem.  You would add bitops.o to the obj-y,
> bitops.c to SYMLINKS, bitops.c-dir := arch/i386/lib.  I think thats
> about it.

-- 
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729



-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] uml-2.6.8-rc2 (2.6.8-rc2 w/ uml-patch-2.6.7-01)

[gboutwel]

> > Probably something simple, but I wanted to discuss with folks

> > more in the now on if the above (stdio_console) fix was the
right

> > one, and to find out if anyone else has an idea of how to
fixe

> > the find_next_zero_bit error.



> Yes, I saw that error in -mm and I agree that renaming is right;
about the 

> other one, the reported fix (below) is correct; also, it would
be nice to 

> delete include/asm-um/bitops.h (which includes the i386 version
of the 

> include; since those funcs were moved to out-of-line ones,
the trick must be 

> done that way).



When I delete include/asm-um/bitops.h I can nolonger compile
(main.o complains aobut implicit definition of find_next_bit
& find_next_zero_bit).



What am I supposed to replace include/asm-um/bitops.h with? 
Am I somehow supposed to symlink it to an SUBARCH?



> Note however that we must not hardcode that we steel things
from i386 arch, 

> but we must use $(SUBARCH) - or it will break for any other
archs (the only 

> working one, at the moment, is the x86_64 one).



I've tweaked my 2.6.8-rc2-1um to use $(SUBARCH), as well..  If
I can figure out how to handle bitops.h similarly, I'll post
an patch of my changes vs an vanilla 2.6.8-rc2 tree.



George



----------------------------------------

Single? Find Christian Singles like you!

http://www.praize.com/singles/





-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel