"make" error during module install

Hello, all!

I am new here, and looking for help. I built a new AMD Sempron (w/SuSE 10.0 x86_64) machine for my lab, and want to install National Instruments software to manage my PCI DAQ card. The problem? The NI-KAL module that I am trying to compile (upon which the rest of the installation seems to hinge) returns errors when I attempt a "make install" -

error: ‘struct <anonymous>’ has no member named ‘pud’

along with a host of warnings. I am sorta a Linux newbie, but I have performed all the operations in NI-KAL's README files, such as "./configure" (as well as "make modules_prepare" and "make modules") and am baffled. I have been poking around the internets and it seems others have this trouble with 64-bit systems and 32-bit applications. Might this error be specific to SuSE 10.0? I am thinking of going back to 9.2, which has been vetted for non-64-bit NI software. . .

I am thinking of going back to 9.2, which has been vetted for non-64-bit NI software. . .

I think this may be a 64-bit problem judging by a quick Google search on the error you are getting. Have you had this NI daq running on a Linux platform before such as SuSE 9.2?

I have National Instruments software running at work in a MS environment but not had experience using under Linux. (I know that NI do support Linux).

Originally posted by pturner7@Jan 13 2006, 04:00 PM
I think this may be a 64-bit problem judging by a quick Google search on the error you are getting.* Have you had this NI daq running on a Linux platform before such as SuSE 9.2?

I have National Instruments software running at work in a MS environment but not had experience using under Linux.* (I know that NI do support Linux).

[snapback]111481[/snapback]

Yeah, NI only supports 32-bit, SuSE 9.1 and 9.2 systems. :angry: I have looked in the NI "Developer Zone" but didn't find anything useful. I also have had to modify the install script just to get LabVIEW installed (which seems to work fine), and have been trying to work around this new difficulty. I am hoping to find some info on what the error itself means. Perhaps I could modify the NI-KAL source to reflect the system I have?

I am hoping to find some info on what the error itself means.

Seems like it could be driver problem?

Sorry, can't be any more help here.

Being x86-64 contains two sets of libraries lib & lib64, we need to pass compile flags to point the makefiles to the correct library. Two ways I know of are "linux32" & "-m32".

I recommend searching for both before starting example: man linux32; google/linux for "-m32", http://marc.theaimsgroup.com/?l=suse-amd64...mpile+32bit&q=b, etc.

If we think of ./configure as a tool to prepare your system to get ready for the makefile, then think "./configure -m32". If you are in a hurry give it a go; but I recommend doing a little reading 1st.

I see NI also has a forum...

Let us know how it turns out...

Oh yes, there is another way, however it means editing the makefile for every instance of lib reference & pointing it to the correct lib. It is much easier to pass a compile flag.

Have fun.

Snake had a good point. If you still have issues, check the version of GCC you are using. May have to either upgrade or down grade for better compatability.
Cheers,
Bandit

Good point Bandit!

Thanks all!

Will take everything under advisement and try a few things. . .

Oh, and it IS pretty fun.

Hmmmmm. Neither "linux32" nor "./configure -m32" seems to work. I think this might be a memory issue, perhaps? The code that the error message refers me to looks like this:

static inline pte_t *nNIKAL100_getKernelPageTableEntry(nLinux_mm* mm, void *address)
{
pgd_t *pgd;
pmd_t *pmd;
pgd = pgd_offset(mm, (nNIKAL100_tUPtr)address);
pmd = pmd_offset(pgd, (nNIKAL100_tUPtr)address);
nNIKAL100_compileTimeAssert(sizeof(pte_t) <= 16, "pte_t is larger then 128 bits!\n");
#ifdef nNIKAL100_kPTEOffsetKernel
return pte_offset_kernel(pmd, (nNIKAL100_tUPtr)address);
#else
return pte_offset(pmd, (nNIKAL100_tUPtr)address);
#endif
}

For the life of me, I can't tell what the "pud" referred to is. The nice lady at National Instruments says this:

How are you? I am sorry I could not get back to you earlier; I have
been in training all week. I did research your
issue some more, and from what I've seen, even though SUSE 10 is not
officially supported, it seems to be a
Kernel issue. People who have had similar issues were using 2.6.10
and above, and using the 2.6.8 kernel seemed
to work for them. You can give this a shot and see if it resolves the
issue. Feel free to call or reply to this email if you have
any questions. Have a great day!

Perhaps it is the kernal I have?