Re: LKPK (Live Kernel Patching Kit)

Adam Sulmicki (adam@cfar.UMD.EDU)
Sun, 26 Jul 1998 18:52:49 -0400

Rogier Wolff writes:

->It is not that bad. "insmod" must be doing the same thing. You can
->load a module into the running kernel, and the external variables will
->be resolved in the same way.
->When I wrote the first "modules" implementation, I had a script modify
->"" into an assembly program that simply had all the variable
->declarations, together with "org" statements to put them at the right
->spot. If you compile that, you have an "object" to link against which
->doesn't hold any code, but will cause all variables to be resolved
->to the right address.
->I -=THINK=- that insmod does the resolving/linking by itself. I didn't
->want to write a new linker....

I think I kind of get idea how it should work, but actual
implementation of something like this is currently beyond my
skills :(

I think it is insmod which,rightly, complains that it can't
find the variable which is not being expoted.

This or other way, can you take a look at my code,
that is at the sigio patch. Mabye you have some suggestions how
to fix references to wake_up_process and init_taks_union, so that
they point to right address and insmod does not complain.

Currnetly those two are declared as local vars, what is obviously
wrong. If I declare them extern, then it will be fine from module
point of view, but insmod still will be unable to find them.
All the things I'm talking about are in sigio/patch.h

The package is at:


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at