The following patch seems to fix building IDE as a module
in 2.5.41. I am running this code on the machine that I'm using to
compose this email.
A couple of notes:
1. For the time being, I have reassembled some of the
drivers/ide object files into ide-mod.o again, because there are some
circular references (not necessarily a bad thing) that modprobe cannot
otherwise handle. For now, this includes putting ide-probe in
ide-mod.
2. I have changed cmd640.o and legacy.o from dep_bool to
dep_tristate and made them also depend on $CONFIG_BLK_DEV_IDE, so
that they will only be offered as modules if ide-mod.o is a
module (since, like any IDE driver, they require some symbols in
ide-mod.o).
These changes are probably not perfect, but I think they
should be an improvement with no real disadvantages in comparison to
what they replace, so I would encourage you to integrate the changes
if you see no problem. Please let me know what you want to do, or if
there is something more you'd like me to do regarding this patch.
--
Adam J. Richter __ ______________ 575 Oroville Road
[email protected] \ / Milpitas, California 95035
+1 408 309-6081 | g g d r a s i l United States of America
"Free Software For The Rest Of Us."
On Thu, 2002-10-10 at 14:44, Adam J. Richter wrote:
> 1. For the time being, I have reassembled some of the
> drivers/ide object files into ide-mod.o again, because there are some
> circular references (not necessarily a bad thing) that modprobe cannot
> otherwise handle. For now, this includes putting ide-probe in
> ide-mod.
I don't think that is actually avoidable. We end up with a lump that I'd
probably call "ide-core" which is fine.
> 2. I have changed cmd640.o and legacy.o from dep_bool to
> dep_tristate and made them also depend on $CONFIG_BLK_DEV_IDE, so
> that they will only be offered as modules if ide-mod.o is a
> module (since, like any IDE driver, they require some symbols in
> ide-mod.o).
Sounds right
> These changes are probably not perfect, but I think they
> should be an improvement with no real disadvantages in comparison to
> what they replace, so I would encourage you to integrate the changes
> if you see no problem. Please let me know what you want to do, or if
> there is something more you'd like me to do regarding this patch.
I'll go over it again, but I have no major problem with applying them
and working from there to clean up the corners. Firstly however do one
little thing - dont put extern blah() in the code, add them to the right
header files.
On Thu, Oct 10, 2002 at 03:18:36PM +0100, Alan Cox wrote:
> I'll go over it again, but I have no major problem with applying them
> and working from there to clean up the corners. Firstly however do one
> little thing - dont put extern blah() in the code, add them to the right
> header files.
Sure. Here is the updated patch. I've verified that it
compiles without complaint.
--
Adam J. Richter __ ______________ 575 Oroville Road
[email protected] \ / Milpitas, California 95035
+1 408 309-6081 | g g d r a s i l United States of America
"Free Software For The Rest Of Us."