[parisc-linux] 64bit kernel & nfs module: Global Offset Table overflow (used 1068, allowed 1023)

[parisc-linux] 64bit kernel & nfs module: Global Offset Table overflow (used 1068, allowed 1023)

[Helge Deller]

With a 64bit Kernel (git head):

c3000:/home/cvs# modprobe nfs
FATAL: Error inserting nfs (/lib/modules/2.6.19-rc2-pa0/kernel/fs/nfs/nfs.ko): Invalid argument

dmesg says:
nfs: Global Offset Table overflow (used 1068, allowed 1023)


In the source code arch/parisc/kernel/module.c I found:

/* Maximum number of GOT entries. We use a long displacement ldd from
 * the bottom of the table, which has a maximum signed displacement of
 * 0x3fff; however, since we're only going forward, this becomes
 * 0x1fff, and thus, since each GOT entry is 8 bytes long we can have
 * at most 1023 entries */
#define MAX_GOTS        1023

Since I don't have any real idea about GOT and modules in general:
Isn't it possible to increase this number for 64bit kernels ?

Helge
_______________________________________________
parisc-linux mailing list
parisc-linux@lists.parisc-linux.org
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Re: [parisc-linux] 64bit kernel & nfs module: Global Offset Table overflow (used 1068, allowed 1023)

[Matthew Wilcox]

On Wed, Oct 18, 2006 at 10:03:11PM +0200, Helge Deller wrote:
> With a 64bit Kernel (git head):
> 
> c3000:/home/cvs# modprobe nfs
> FATAL: Error inserting nfs (/lib/modules/2.6.19-rc2-pa0/kernel/fs/nfs/nfs.ko): Invalid argument
> 
> dmesg says:
> nfs: Global Offset Table overflow (used 1068, allowed 1023)
> 
> 
> In the source code arch/parisc/kernel/module.c I found:
> 
> /* Maximum number of GOT entries. We use a long displacement ldd from
>  * the bottom of the table, which has a maximum signed displacement of
>  * 0x3fff; however, since we're only going forward, this becomes
>  * 0x1fff, and thus, since each GOT entry is 8 bytes long we can have
>  * at most 1023 entries */
> #define MAX_GOTS        1023
> 
> Since I don't have any real idea about GOT and modules in general:
> Isn't it possible to increase this number for 64bit kernels ?

We'd need to make the GOT register point 0x2000 above where it currently
does.  Then we could use a signed +0x1fff / -0x2000 offset.

I have no idea how to make this happen.