[wplug] geforce4 and redhat

[Vanco, Donald]

Here's what's worked on some of the RH betas that had issues like this:

1) grab http://www.minion.de/files/NVIDIA_kernel-1.0-4191-2.5-tl.diff
2) on line 151 of the patch there's 
+#  define REMAP_PAGE_RANGE(a...)        remap_page_range(## a)
which you need to change to 
+#  define REMAP_PAGE_RANGE(a...)        remap_page_range(vma,## a)
(add ifdefs detecting RH8 2.4 kernels if you want to be fancy)
3) Apply modified patch
4) Compile as usual
5) Edit /etc/X11/prefdm and /usr/X11R6/bin/startx (depending on what you
use) to have a export LD_ASSUME_KERNEL=2.4.19 right in the beginning.


YMMV
Don

Tom McCullough wrote:
> Vanco, Donald wrote:
> 
>> Tom McCullough wrote:
>> 
>> 
>>> bgtrio at yahoo.com wrote:
>>> 
>>> 
>>> 
>>>> Well, the "no such user - using root" thing is pretty normal, but I
>>>> typically only see this message one time if memory serves. I don't
>>>> think I can help you without seeing some more of the output from
>>>> the rpmbuild command though.. 
>>>> 
>>>> You are logged in as root, right?
>>>> 
>>>> B
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>> the error scrolled off the screen too fast,
>>> 
>>> 
>> 
>> 	Under Linux, you can "shift-page_up" to look at what has blown by.
>> Moving to another VC will, of course, blow out this buffer.
>> 
>> Don
>> 
>> _______________________________________________
>> wplug mailing list
>> wplug at wplug.org
>> http://www.wplug.org/mailman/listinfo/wplug
>> 
>> 
>> 
> That's a great little feature I was not aware of.  Thanks.
> The errors are the same that I saw with the tarball.  They seem to
> center around a function in the nv.c file called
> "remap_page_range_R839d959e"  and they cause the nv.o file to not be
> generated. examples:
> 
> warning: passing arg1 of remap........59e makes pointer from integer
> without a cast too few arguments to function remap..............59e
> Incompatible type for argument 4 of remap.............59e
> 
> then the last few are different:
> 
> In function 'nv_get_phys_address'
> warning: implicit declaration of function 'pte_offset'
> invalid type argument of 'unary *'
> make: *** [nv.o] Error 1
> that seems to be the fatal error info.  I am a little confused about
> the compile making nv.o and the last line that says cannot stat
>  nvidia.o? Neither file is present, though.
> curioser and curioser
> tom......................................
> I think I have been inside long enough today.  I'll check back later
> tonight 
> 
> _______________________________________________
> wplug mailing list
> wplug at wplug.org
> http://www.wplug.org/mailman/listinfo/wplug