Messages in this thread | | | From | Rusty Russell <> | Subject | Re: [PATCH] modules as shared objects | Date | Mon, 25 Nov 2002 12:10:50 +1100 |
| |
In message <20021123140128.A699@twiddle.net> you write: > * Freeing the .init.* sections. Also requires the link script. > I believe the proper solution is to use a different program > header with a different load tag: > > #define PT_INIT_LOAD 0x60000001 > > In this way we don't make a mistake and discard a segment that > we're not supposed to.
Were you thinking of dropping the dynamic at the same time (ie. putting the dynamic and init sections in the tail using the linker script)? Hmm, I guess not since you
I think an "mod_arch_free_partial(struct module *mod, void *start, void *end)" function which archs could use to do whatever fancy magic they wanted (eg. trim exception and really drop the pages, or your vmalloc_drop_pages() function, or nothing at all). x86_64, sparc, sparc64, ppc and arm do/will have their own allocators already.
> +module_core_size(const ElfW(Ehdr) *hdr, struct module *module,
I *really* hate this ElfW(type) notation. I know it's used in binutils etc, but it's a visual bitch to parse. Please please please at least use the #defines or typedefs within the generic kernel/module.c so it's readable.
Other than some gratuitous reformatting 8), looks really good. Implementing PPC now.
Thanks, Rusty. -- Anyone who quotes me in their sig is an idiot. -- Rusty Russell. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |