[PATCH v3 04/13] mm/execmem, arch: convert remaining overrides of module_alloc to execmem

Mike Rapoport rppt at kernel.org
Wed Oct 4 22:28:24 PDT 2023


On Wed, Oct 04, 2023 at 12:29:36AM +0000, Edgecombe, Rick P wrote:
> On Mon, 2023-09-18 at 10:29 +0300, Mike Rapoport wrote:
> > diff --git a/arch/x86/kernel/module.c b/arch/x86/kernel/module.c
> > index 5f71a0cf4399..9d37375e2f05 100644
> > --- a/arch/x86/kernel/module.c
> > +++ b/arch/x86/kernel/module.c
> >
> > -void *module_alloc(unsigned long size)
> > +struct execmem_params __init *execmem_arch_params(void)
> >  {
> > -       gfp_t gfp_mask = GFP_KERNEL;
> > -       void *p;
> > -
> > -       if (PAGE_ALIGN(size) > MODULES_LEN)
> > -               return NULL;
> > +       unsigned long module_load_offset = 0;
> > +       unsigned long start;
> >  
> > -       p = __vmalloc_node_range(size, MODULE_ALIGN,
> > -                                MODULES_VADDR +
> > get_module_load_offset(),
> > -                                MODULES_END, gfp_mask, PAGE_KERNEL,
> > -                                VM_FLUSH_RESET_PERMS |
> > VM_DEFER_KMEMLEAK,
> > -                                NUMA_NO_NODE,
> > __builtin_return_address(0));
> > +       if (IS_ENABLED(CONFIG_RANDOMIZE_BASE) && kaslr_enabled())
> > +               module_load_offset =
> > +                       get_random_u32_inclusive(1, 1024) *
> > PAGE_SIZE;
> 
> Minor:
> I think you can skip the IS_ENABLED(CONFIG_RANDOMIZE_BASE) part because
> CONFIG_RANDOMIZE_MEMORY depends on CONFIG_RANDOMIZE_BASE (which is
> checked in kaslr_enabled()).

Thanks, I'll look into it.

-- 
Sincerely yours,
Mike.



More information about the linux-riscv mailing list