Re: Problems using linux-2.5.6-pre1 with a current CVS pull...

[Stephen Lord <lord@xxxxxxx>]

Jordan Breeding wrote:

> Today I decided to upgrade my XFS kernel up to the 2.5.6-pre1 patch when
> it came out.  I created a tree with 2.5.6-pre1 first, then updated my
> xfs tree and then executed the following:
>
> ln -s linux-2.5-xfs/linux linux-2.5.5-xfs
> diff -urN linux-2.5.5-original linux-2.5.5-xfs | patch -p1
> -d/usr/src/linux
>
> I only got around four rejects which were all very trivial excerpts
> which I fixed by hand.  However, when I try and do a `make bzImage` here
> is the error I get:
>
> --begin snip--
> gcc296 -D__KERNEL__ -I/usr/src/linux/include  -Wall -Wstrict-prototypes
> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer
> -pipe -mpreferred-stack-boundary=2 -march=i686  -nostdinc -I
> /usr/lib/gcc-lib/i386-redhat-linux7/2.96/include -I. -funsigned-char
> -DKBUILD_BASENAME=xfs_iget  -c -o xfs_iget.o xfs_iget.c
> gcc296 -D__KERNEL__ -I/usr/src/linux/include  -Wall -Wstrict-prototypes
> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer
> -pipe -mpreferred-stack-boundary=2 -march=i686  -nostdinc -I
> /usr/lib/gcc-lib/i386-redhat-linux7/2.96/include -I. -funsigned-char
> -DKBUILD_BASENAME=xfs_inode  -c -o xfs_inode.o xfs_inode.c
> gcc296 -D__KERNEL__ -I/usr/src/linux/include  -Wall -Wstrict-prototypes
> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer
> -pipe -mpreferred-stack-boundary=2 -march=i686  -nostdinc -I
> /usr/lib/gcc-lib/i386-redhat-linux7/2.96/include -I. -funsigned-char
> -DKBUILD_BASENAME=xfs_inode_item  -c -o xfs_inode_item.o
> xfs_inode_item.c
> gcc296 -D__KERNEL__ -I/usr/src/linux/include  -Wall -Wstrict-prototypes
> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer
> -pipe -mpreferred-stack-boundary=2 -march=i686  -nostdinc -I
> /usr/lib/gcc-lib/i386-redhat-linux7/2.96/include -I. -funsigned-char
> -DKBUILD_BASENAME=xfs_iocore  -c -o xfs_iocore.o xfs_iocore.c
> gcc296 -D__KERNEL__ -I/usr/src/linux/include  -Wall -Wstrict-prototypes
> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer
> -pipe -mpreferred-stack-boundary=2 -march=i686  -nostdinc -I
> /usr/lib/gcc-lib/i386-redhat-linux7/2.96/include -I. -funsigned-char
> -DKBUILD_BASENAME=xfs_itable  -c -o xfs_itable.o xfs_itable.c
> gcc296 -D__KERNEL__ -I/usr/src/linux/include  -Wall -Wstrict-prototypes
> -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer
> -pipe -mpreferred-stack-boundary=2 -march=i686  -nostdinc -I
> /usr/lib/gcc-lib/i386-redhat-linux7/2.96/include -I. -funsigned-char
> -DKBUILD_BASENAME=xfs_dfrag  -c -o xfs_dfrag.o xfs_dfrag.c
> xfs_dfrag.c: In function `xfs_swapext':
> xfs_dfrag.c:226: invalid operands to binary !=
> xfs_dfrag.c:226: invalid operands to binary !=
> make[3]: *** [xfs_dfrag.o] Error 1
> make[3]: Leaving directory `/usr/src/linux/fs/xfs'
> make[2]: *** [first_rule] Error 2
> make[2]: Leaving directory `/usr/src/linux/fs/xfs'
> make[1]: *** [_subdir_xfs] Error 2
> make[1]: Leaving directory `/usr/src/linux/fs'
> make: *** [_dir_fs] Error 2
> --end snip--
>
> And here is line 266 of fs/xfs/xfs_dfrag.c:
>
> --begin snip--
> if (VN_MAPPED(vp)) {
> --end snip--
>
> And VM_MAPPED can be found in fs/xfs/linux/xfs_vnode.h:
>
> --begin snip--
> #define VN_MAPPED(vp)   ((LINVFS_GET_IP(vp)->i_mapping->i_mmap != NULL)
> || \
>                         (LINVFS_GET_IP(vp)->i_mapping->i_mmap_shared !=
> NULL))
> --end snip--
>
> So....what is going on with the current CVS code, or is it just my
> kernel being incorrectly merged with xfs?  Thanks for any help.
>
> Jordan
>
> P.S. - Please Cc: me on responses, thank you.
I have done a merge and built a kernel, not had a chance to boot it yet,
maybe today. This was the one build error, it needs to use list_empty to
check if these lists have entries or not now:

#define       VN_MAPPED(vp)   \
   (!list_empty(&(LINVFS_GET_IP(vp)->i_mapping->i_mmap)) || \
   (!list_empty(&(LINVFS_GET_IP(vp)->i_mapping->i_mmap_shared))))


Steve

p.s. Did the 2.5 bitkeeper tree move - if so were is it now?