Hi there,
I tried to make the MPC8xx IDE Driver loadable as a module (nevermind if
this is useful or not).
It is working - at least in my configuration.
BUT:
Apparently I am too stupid to understand all this Makefile stuff. That's
why I changed the drivers/ide/Makefile to look like this now:
===================================================================
RCS file: /home/cvsroot/linux-2.4.18/drivers/ide/Makefile,v
retrieving revision 1.1.1.1
diff -u -r1.1.1.1 Makefile
--- drivers/ide/Makefile 2002/04/05 11:46:19 1.1.1.1
+++ drivers/ide/Makefile 2002/04/10 12:59:34
@@ -65,7 +65,15 @@
ide-obj-$(CONFIG_BLK_DEV_TRM290) += trm290.o
ide-obj-$(CONFIG_BLK_DEV_UMC8672) += umc8672.o
ide-obj-$(CONFIG_BLK_DEV_VIA82CXXX) += via82cxxx.o
-ide-obj-$(CONFIG_BLK_DEV_MPC8xx_IDE) += ide-m8xx.o
+
+ifeq ($(CONFIG_BLK_DEV_MPC8xx_IDE),y)
+ ide-obj-$(CONFIG_BLK_DEV_MPC8xx_IDE) += ide-m8xx.o
+endif
+
+ifeq ($(CONFIG_BLK_DEV_MPC8xx_IDE),m)
+ obj-$(CONFIG_BLK_DEV_MPC8xx_IDE) += ide-m8xx.o
+endif
+
# The virtualised raid layers MUST come after the ide itself or bad
stuff
# will happen.
Otherwiser eihter there problems during compile or ide-m8xx.o is not
built at all.
I am pretty sure that there is a better i.e. correct way to do this. Any
hints?
Another points is:
When I unload the module and call ide_unregister(0) then the function
m8xx_ide_init_hwif_ports() is called again!
Should I released the region which is ioremaped the first time
m8xx_ide_init_hwif_ports() gets called?
if (!ide_base) {
...
ide_base=(unsigned long)ioremap(ide_phy_base,
ide_phy_end-ide_phy_base);
...
}
Thanks a million,
Steven
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
