Does LVM compile on Sparc64?

Hi Everybody

I got the source of Linux 2.2.20 and applied the patches of LVM 1.0.1 (including Stephen Tweedie's rawio patch and linux-2.2.18-sparc64.diff).

The patches do apply well, but it seems like there are some missing definitions in include/linux/lvm.h:

duke:/usr/src/linux# make vmlinux
[snipped]
sparc64-linux-gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -fno-strict-aliasing -m64 -pipe -mno-fpu -mcpu=ultrasparc -mcmodel=medlow -ffixed-g4 -fcall-used-g5 -fcall-used-g7 -Wno-sign-compare   -c -o ioctl32.o ioctl32.c
ioctl32.c: In function `do_lvm_ioctl':
ioctl32.c:2570: warning: assignment makes pointer from integer without a cast
ioctl32.c:2604: structure has no member named `inode'
ioctl32.c:2645: warning: assignment from incompatible pointer type
ioctl32.c:2646: structure has no member named `inode'
ioctl32.c:2653: structure has no member named `inode'
ioctl32.c:2666: structure has no member named `inode'
ioctl32.c:2545: warning: `v' might be used uninitialized in this function
make[1]: *** [ioctl32.o] Error 1
make[1]: Leaving directory `/usr/src/linux-2.2.20-RAID/arch/sparc64/kernel'
make: *** [_dir_arch/sparc64/kernel] Error 2
duke:/usr/src/linux#

ioctl32.c:2604 looks like:

v->pv[i]->pe = NULL; v->pv[i]->inode = NULL;

v makes use of vg_t and v->pv makes use of pv_t (both defined in lvm.h)

lvm.h on sparc64 is completly different than on Intel. I'm just wondering now, if all the stuff gets fixed, if I apply the struct member 'inode' to pv_t, like it is done on Intel.

Or are there any other patches, to run LVM 1.0.1 on sparc64?

Thanks & Cheers
Thomas

  .-.   Thomas Bader · [email protected] · http://www.t-bader.ch/  .-.
  oo|                                                                     oo|
 /`'\     Einen Unix-Shellaccount gibt es unter http://www.trash.net/    /`'\
(\_;/)       PGP Key-ID: 0x3A4B7F5D (RSA)  0x7584F5D8 (DSA/EG)          (\_;/)

See http://lists.sistina.com/pipermail/linux-lvm/2001-December/009950.html

for a patch that should work.

I'll make a proper 2.2 one next week.

patrick

Hi

Oh nice. I didn't found this patch yet (searched with google).

I applied it and it compiled without error. But now, my machine does not boot anymore - it hangs directly after it has detected my raid arrays:

-->-->-->--
md: updating md0 RAID superblock on device
sdb1 [events: 00000022](write) sdb1's sb offset: 3145664
md: serializing resync, md0 has overlapping physical units with md2!
sda1 [events: 00000022](write) sda1's sb offset: 3145664 .
... autorun DONE.
--<--<--<--

After this, it won't continue booting up. I booted a backup copy of my kernel image. I saw, that after the above step ("... autorun DONE.") the kernel tries to mount my root partition read-only.

My / is on a RAID1 array (/dev/md0) with ext2. Does the LVM code touch anything of these two?

Any other hints?

Thanks & Cheers
Thomas

  .-.   Thomas Bader · [email protected] · http://www.t-bader.ch/  .-.
  oo|                                                                     oo|
 /`'\     Einen Unix-Shellaccount gibt es unter http://www.trash.net/    /`'\
(\_;/)       PGP Key-ID: 0x3A4B7F5D (RSA)  0x7584F5D8 (DSA/EG)          (\_;/)

I honestly don't know what that should be, sorry. I know for sure that LVM has worked for me on SPARC64 with those patches applied so all I can suggest is to check the rest of your kernel config and maybe do a clean rebuild of the kernel.

patrick