Re: powerpc fixes

[Robert Millan]

On Thu, Jan 31, 2008 at 12:26:46AM +0100, Jordi Mallach wrote:
> The following patches fix the powerpc build for the latest Debian
> package. Sorry for not diffing this against HEAD.
> 
> However, the resulting binaries don't work as expected:
> 
> 18111:address@hidden:/boot/grub$ sudo grub-install "(hd0)"
> grub-probe: error: /dev/stdin:1: Bad device name
> grub-probe: error: /dev/stdin:1: Bad device name
> /boot/grub must be a mount point.
> 
> grub-mkdevicemap generated this when I called it by hand:
> (/address@hidden/address@hidden/address@hidden)       /dev/hda

Are you using the latest ofpathname?

> diff -Nuar grub2-1.95+20080128/conf/powerpc-ieee1275.mk 
> buildd/grub2-1.95+20080128/conf/powerpc-ieee1275.mk
> --- grub2-1.95+20080128/conf/powerpc-ieee1275.mk      2008-01-27 
> 22:44:02.000000000 +0000
> +++ buildd/grub2-1.95+20080128/conf/powerpc-ieee1275.mk       2008-01-30 
> 22:45:59.000000000 +0000

Please skip *.mk files in patches ;-)

> @@ -69,7 +70,7 @@
>       fs/affs.c fs/cpio.c fs/ext2.c fs/fat.c fs/hfs.c         \
>       fs/hfsplus.c fs/iso9660.c fs/jfs.c fs/minix.c           \
>       fs/ntfs.c fs/ntfscomp.c fs/reiserfs.c fs/sfs.c          \
> -     fs/ufs.c fs/xfs.c                                       \
> +     fs/ufs.c fs/xfs.c fs/fshelp.c                           \
>       \
>       io/gzio.c                                                       \
>       kern/device.c kern/disk.c kern/dl.c kern/elf.c kern/env.c       \

This won't apply cleanly to HEAD.

> diff -Nuar grub2-1.95+20080128/include/grub/powerpc/libgcc.h 
> buildd/grub2-1.95+20080128/include/grub/powerpc/libgcc.h
> --- grub2-1.95+20080128/include/grub/powerpc/libgcc.h 2008-01-10 
> 07:27:08.000000000 +0000
> +++ buildd/grub2-1.95+20080128/include/grub/powerpc/libgcc.h  2008-01-30 
> 22:39:43.000000000 +0000
> @@ -25,21 +25,14 @@
>  void EXPORT_FUNC (__fixunsdfsi) (void);
>  void EXPORT_FUNC (__floatsidf) (void);
>  void EXPORT_FUNC (__floatsisf) (void);
> +void EXPORT_FUNC (__floatundisf) (void);
>  void EXPORT_FUNC (__lshrdi3) (void);
> -void EXPORT_FUNC (__make_dp) (void);
> -void EXPORT_FUNC (__make_fp) (void);
>  void EXPORT_FUNC (__muldf3) (void);
>  void EXPORT_FUNC (__mulsf3) (void);
> -void EXPORT_FUNC (__pack_d) (void);
> -void EXPORT_FUNC (__pack_f) (void);
>  void EXPORT_FUNC (__subdf3) (void);
>  void EXPORT_FUNC (__subsf3) (void);
> -void EXPORT_FUNC (__thenan_df) (void);
> -void EXPORT_FUNC (__thenan_sf) (void);
>  void EXPORT_FUNC (__trampoline_setup) (void);
>  void EXPORT_FUNC (__truncdfsf2) (void);
>  void EXPORT_FUNC (__ucmpdi2) (void);
> -void EXPORT_FUNC (__unpack_d) (void);
> -void EXPORT_FUNC (__unpack_f) (void);
>  void EXPORT_FUNC (__floatdisf) (void);
>  void EXPORT_FUNC (__cmpdi2) (void);

This part (as discussed in the other thread) is wrong.  We need to extract
symbol list dynamicaly (if we don't want to break arbitrary versions of gcc).

-- 
Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call… if you are unable to speak?
(as seen on /.)