hi i use suse 10.1 64bits i'm not able to build: forcedeth.c my Makefile obj-m := forcedeth.o when i do make -C /usr/src/linux-2.6.16.21-0.13 SUBDIRS=$PWD modules make: Entering directory `/usr/src/linux-2.6.16.21-0.13' Makefile:450: .config: Aucun fichier ou répertoire de ce type WARNING: Symbol version dump /usr/src/linux-2.6.16.21-0.13/Module.symvers is missing; modules will have no dependencies and modversions. CC [M] /home/collinm/Download/nvidia/forcedeth.o /home/collinm/Download/nvidia/forcedeth.c: In function ‘nv_set_sg’: /home/collinm/Download/nvidia/forcedeth.c:3886: warning: control reaches end of non-void function /bin/sh: scripts/basic/fixdep: No such file or directory make[1]: *** [/home/collinm/Download/nvidia/forcedeth.o] Error 1 make: *** [_module_/home/collinm/Download/nvidia] Error2 make: Leaving directory `/usr/src/linux-2.6.16.21-0.13' nvidia staff said me: Your kernel source seems to be either broken or incomplete i installed nvidia driver and webcam driver withtout problem (need to build and have kernel source) uname -a return: linux64 2.6.16.21-0.13-default any idea? thanks
Le mardi 29 août 2006 14:02, Anders Johansson a écrit :
On Tuesday 29 August 2006 19:55, Marc Collin wrote:
/bin/sh: scripts/basic/fixdep: No such file or directory
Did you run
cd /usr/src/linux-2.6.16.21-0.13 make scripts make prepare
before trying to compile?
no that work now... thanks for theses commands ok 2 files is generated /home/collinm/Download/nvidia/forcedeth.mod.o /home/collinm/Download/nvidia/forcedeth.ko where i need to put theses file?
On Tuesday 29 August 2006 20:09, Marc Collin wrote:
ok 2 files is generated
/home/collinm/Download/nvidia/forcedeth.mod.o /home/collinm/Download/nvidia/forcedeth.ko
where i need to put theses file?
The forcedeth.ko file needs to be put somewhere under /lib/modules/2.6.16.21-0.13/, probably under the "extra" directory (but it doesn't really matter). Then run "depmod -a", and then you can use it The other file is just a temporary file used during the compile, you can ignore it or delete it
On Tuesday 29 August 2006 20:09, Marc Collin wrote:
ok 2 files is generated
/home/collinm/Download/nvidia/forcedeth.mod.o /home/collinm/Download/nvidia/forcedeth.ko
where i need to put theses file?
Sorry, my other mail was stupid. Since the forcedeth driver already exists in the shipped kernel, you need to either delete it first before you move this one into the kernel directory, or overwrite the old one with the new one. By the way, where did you get this version from?
Le mardi 29 août 2006 14:25, Anders Johansson a écrit :
On Tuesday 29 August 2006 20:09, Marc Collin wrote:
ok 2 files is generated
/home/collinm/Download/nvidia/forcedeth.mod.o /home/collinm/Download/nvidia/forcedeth.ko
where i need to put theses file?
Sorry, my other mail was stupid.
Since the forcedeth driver already exists in the shipped kernel, you need to either delete it first before you move this one into the kernel directory, or overwrite the old one with the new one.
By the way, where did you get this version from?
nvidia web site nvidia go out this file there is a couple of day..... i hope suse will add it to their next kernel..... it's suppose to correct problem with a dual boot (win/linux) if you boot on windows, reboot to linux, the network card will not work because ms lock the card....... i hope this news release will correct the problem my brother use both system not a real problem for me, i only use suse :)
Le mardi 29 août 2006 14:25, Anders Johansson a écrit :
On Tuesday 29 August 2006 20:09, Marc Collin wrote:
ok 2 files is generated
/home/collinm/Download/nvidia/forcedeth.mod.o /home/collinm/Download/nvidia/forcedeth.ko
where i need to put theses file?
Sorry, my other mail was stupid.
Since the forcedeth driver already exists in the shipped kernel, you need to either delete it first before you move this one into the kernel directory, or overwrite the old one with the new one.
surely a rmmod is nedded?
By the way, where did you get this version from?
On Tuesday 29 August 2006 23:53, Marc Collin wrote:
Le mardi 29 août 2006 14:25, Anders Johansson a écrit :
On Tuesday 29 August 2006 20:09, Marc Collin wrote:
ok 2 files is generated
/home/collinm/Download/nvidia/forcedeth.mod.o /home/collinm/Download/nvidia/forcedeth.ko
where i need to put theses file?
Sorry, my other mail was stupid.
Since the forcedeth driver already exists in the shipped kernel, you need to either delete it first before you move this one into the kernel directory, or overwrite the old one with the new one.
surely a rmmod is nedded?
To actually use the driver, assuming you have the old one loaded, yes, you need to unload it first, and to do that you probably have to stop the network first too, with "rcnetwork stop", and then "rmmod forcedeth", before you can use the new one
Le mardi 29 août 2006 18:05, Anders Johansson a écrit :
On Tuesday 29 August 2006 23:53, Marc Collin wrote:
Le mardi 29 août 2006 14:25, Anders Johansson a écrit :
On Tuesday 29 August 2006 20:09, Marc Collin wrote:
ok 2 files is generated
/home/collinm/Download/nvidia/forcedeth.mod.o /home/collinm/Download/nvidia/forcedeth.ko
where i need to put theses file?
Sorry, my other mail was stupid.
Since the forcedeth driver already exists in the shipped kernel, you need to either delete it first before you move this one into the kernel directory, or overwrite the old one with the new one.
surely a rmmod is nedded?
To actually use the driver, assuming you have the old one loaded, yes, you need to unload it first, and to do that you probably have to stop the network first too, with "rcnetwork stop", and then "rmmod forcedeth", before you can use the new one
my brother use the same system than me (suse 10.1 bits) i have an athlon 64bits, him have a athlon 64bits dual core (it use smp kernel) can i take the forcedeth.ko i build and copy it to his system? thanks
On Tuesday 29 August 2006 19:55, Marc Collin wrote:
/home/collinm/Download/nvidia/forcedeth.c: In function ‘nv_set_sg’: /home/collinm/Download/nvidia/forcedeth.c:3886: warning: control reaches end of non-void function
This warning comes from incorrectly-written code. If you let this warning go the driver will have "undefined behaviour". You can edit the code and maybe find the proper return value for nv_set_sg(), and add the proper 'return' statement there. The problem is that the function is declared as returning a value but does not actually do so.
/bin/sh: scripts/basic/fixdep: No such file or directory make[1]: *** [/home/collinm/Download/nvidia/forcedeth.o] Error 1 make: *** [_module_/home/collinm/Download/nvidia] Error2
This is caused by the Makefile looking for a script in a relative path, but you are not building from the top of the kernel source tree. Try building from there, instead of using the -C option to 'make'. Maybe it'll fix the problem, maybe no.
nvidia staff said me:
Your kernel source seems to be either broken or incomplete
Maybe, but the first warning is from broken nvidia code. Some compilers won't complain about this error, but most will (or should). -- ----- stephan@s11n.net http://s11n.net "...pleasure is a grace and is not obedient to the commands of the will." -- Alan W. Watts
On Tuesday 29 August 2006 20:08, stephan beal wrote:
On Tuesday 29 August 2006 19:55, Marc Collin wrote:
/home/collinm/Download/nvidia/forcedeth.c: In function ‘nv_set_sg’: /home/collinm/Download/nvidia/forcedeth.c:3886: warning: control reaches end of non-void function
This warning comes from incorrectly-written code. If you let this warning go the driver will have "undefined behaviour". You can edit the code and maybe find the proper return value for nv_set_sg(), and add the proper 'return' statement there. The problem is that the function is declared as returning a value but does not actually do so.
I haven't seen the actual code, but it doesn't have to be incorrect. gcc doesn't do much in the way of logic analysis, and for example int foo(){ int i; i = 1; if(i) return 0; } will also generate the error. gcc just doesn't see that even if all possible code paths don't return something, all logically accessible paths do Of course this puts more responsibility on the programmer to make sure that he hasn't missed a code path. But it doesn't have to be an actual error
participants (3)
-
Anders Johansson -
Marc Collin -
stephan beal