2008/5/27 Jared D. McNeill <jmcneill%invisible.ca@localhost>:
Chavdar Ivanov wrote:
2008/5/27 Andrew Doran <ad%netbsd.org@localhost>:
On Tue, May 27, 2008 at 01:55:36AM +0200, Adam Hoka wrote:
After compiling a custom kernel I have found the following in my dmesg:
WARNING: module error: builtin module `drm' failed to init
WARNING: module error: builtin module `putter' failed to init
FWIW, the module is not initializing because its char/block device has
already been hooked in by the glue generated by config(1). Two ways to
achieve the same thing and one wins. It's harmless. Some #ifdefs would
sort
the warning, but we need to do some serious work on the kernel config
scheme
to handle both modules and static builds well.
I am seeing the same. If I take out the corresponding lines in my
GENERIC.local, do I have to do something to enable/use the kernel
modules? I haven't seen a panic - I am using drm only at this time.
1. cd sys/modules, make and make install
2. set MKMODULAR=yes in /etc/mk.conf
3. for each of sbin/modload sbin/modunload and sbin/modstat:
USETOOLS=no make clean
USETOOLS=no make
sudo USETOOLS=no make install
4. modload <driver>, currently i915drm and radeondrm are available on i386.
i915drm is available on amd64 but there are stability problems running it
from a module.
Cheers,
Jared
Thanks, all worked as expected. One less thing to modify in the kernel config...
I can't load radeondrm on my office Kayak - it's a 9800SE, which
apparently is RV350 (AFAIU drm supports RV200 and RV300, am I right?).
I am getting:
WARNING: module error: unable to affix module.
But then when I had radeondrm in the kernel config file, everything
showed OK in the dmesg, but drm never actually worked - glxgears would
just freeze the X output, just the mouse cursor remained moving.