Re: [PATCH 1/2] module: allow arch overrides for .init section names

[Jessica Yu]

+++ Vincent Whitchurch [06/05/20 17:25 +0200]:
> ARM stores unwind information for .init.text in sections named
> .ARM.extab.init.text and .ARM.exidx.init.text.  Since those aren't
> currently recognized as init sections, they're allocated along with the
> core section, and relocation fails if the core and the init section are
> allocated from different regions and can't reach other.
>
>  final section addresses:
>        ...
>        0x7f800000 .init.text
>        ..
>        0xcbb54078 .ARM.exidx.init.text
>        ..
>
> section 16 reloc 0 sym '': relocation 42 out of range (0xcbb54078 ->
> 0x7f800000)
>
> Allow architectures to override the section name so that ARM can fix
> this.
>
> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@xxxxxxxx>
> ---
> include/linux/moduleloader.h | 2 ++
> kernel/module.c              | 9 +++++++--
> 2 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/moduleloader.h b/include/linux/moduleloader.h
> index ca92aea8a6bd..07d079085f11 100644
> --- a/include/linux/moduleloader.h
> +++ b/include/linux/moduleloader.h
> @@ -29,6 +29,8 @@ void *module_alloc(unsigned long size);
> /* Free memory returned from module_alloc. */
> void module_memfree(void *module_region);
>
> +bool module_init_section(const char *name);
> +

Hi, I just have two small nits. Could you please add a comment on top of
module_init_section to document it? (similar to module_exit_section())

> /* Determines if the section name is an exit section (that is only used during
>  * module unloading)
>  */
> diff --git a/kernel/module.c b/kernel/module.c
> index 33569a01d6e1..a5f2b4e1ef53 100644
> --- a/kernel/module.c
> +++ b/kernel/module.c
> @@ -2370,6 +2370,11 @@ static long get_offset(struct module *mod, unsigned int *size,
> 	return ret;
> }
>
> +bool __weak module_init_section(const char *name)
> +{
> +	return strstarts(name, ".init");
> +}
> +

Could you move this next to module_exit_section()? Just trying to keep
similar functions together. Otherwise, the patch looks fine to me.

Thanks!

Jessica