from the quill of "Brian J. Murrell" <[EMAIL PROTECTED]> on
scroll <[EMAIL PROTECTED]>
> from the quill of "Brian J. Murrell" <[EMAIL PROTECTED]> on
> scroll <[EMAIL PROTECTED]>
> > Is it? What are the relevant packages?
>
> Well I think I found a basic problem with lm_sensors. It ain't packaged
> with hackkernel. I guess I will try building the hackkernel package
> here to see what is going on.
In yet another reply to my own message...
I don't really know zilch about this i2c/lm_sensors stuff but I will
take a poke at figuring out what the problem is anyway. :-)
In the patch linux-2.4.0-lm_sensors.patch.bz2, there is a segment:
--- linux-old/drivers/Makefile Tue Oct 17 14:18:23 CEST 2000
+++ linux/drivers/Makefile Tue Oct 17 14:18:23 CEST 2000
@@ -41,5 +41,4 @@
# CONFIG_HAMRADIO can be set without CONFIG_NETDEVICE being set -- ch
subdir-$(CONFIG_HAMRADIO) += net/hamradio
-subdir-$(CONFIG_I2C) += i2c
subdir-$(CONFIG_ACPI) += acpi
Why do we not want to build in the drivers/i2c subdirectory?
Also, it would seem that the lm_sensors patch puts:
subdir-$(CONFIG_SENSORS) += sensors
Right near the bottom of the drivers/Makefile, after the evaluation of
MOD_SUB_DIRS := $(sort $(subdir-m) $(both-m)). Is my suspicion that
"sensors" will not get get put into MOD_SUB_DIRS correct? If so then
it won't get built either.
As a sidebar, because of these problems, hackkernel-2.4.0-0.35mdk builds
modules with unresolved symbols. The depmod in %install discovers and
reports this:
if [ -r System.map ]; then /sbin/depmod -ae -F System.map -b /var/tmp/hackkernel
-2.4.0-build -r 2.4.0-0.35.1mdksmp; fi
depmod: *** Unresolved symbols in /var/tmp/hackkernel-2.4.0-build/lib/modules/2.
4.0-0.35.1mdksmp/kernel/drivers/media/video/bttv.o
depmod: i2c_master_send
depmod: i2c_bit_del_bus
depmod: i2c_bit_add_bus
depmod: i2c_master_recv
[ and so on and so forth ]
Why does the specfile not test for unresolved symbols and bomb out of
the build if it sees it rather than shipping kernel RPMs with broken
modules in it?
I will try fixing these issues, rebuilding a kernel rpm and see what
happens.
b.
--
Brian J. Murrell
