Re: [PATCH 5/9] kasan/loongarch: call kasan_init_generic in kasan_init

[Sabyrzhan Tasbolatov <snovitoll@gmail.com>]

On Thu, Jun 26, 2025 at 6:29 PM Peter Zijlstra <peterz@infradead.org> wrote:
>
> On Wed, Jun 25, 2025 at 02:52:20PM +0500, Sabyrzhan Tasbolatov wrote:
> > Call kasan_init_generic() which enables the static flag
> > to mark generic KASAN initialized, otherwise it's an inline stub.
> >
> > Replace `kasan_arch_is_ready` with `kasan_enabled`.
> > Delete the flag `kasan_early_stage` in favor of the global static key
> > enabled via kasan_enabled().
> >
> > printk banner is printed earlier right where `kasan_early_stage`
> > was flipped, just to keep the same flow.
> >
> > Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218315
> > Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
> > ---
> >  arch/loongarch/include/asm/kasan.h | 7 -------
> >  arch/loongarch/mm/kasan_init.c     | 7 ++-----
> >  2 files changed, 2 insertions(+), 12 deletions(-)
> >
> > diff --git a/arch/loongarch/include/asm/kasan.h b/arch/loongarch/include/asm/kasan.h
> > index 7f52bd31b9d..b0b74871257 100644
> > --- a/arch/loongarch/include/asm/kasan.h
> > +++ b/arch/loongarch/include/asm/kasan.h
> > @@ -66,7 +66,6 @@
> >  #define XKPRANGE_WC_SHADOW_OFFSET    (KASAN_SHADOW_START + XKPRANGE_WC_KASAN_OFFSET)
> >  #define XKVRANGE_VC_SHADOW_OFFSET    (KASAN_SHADOW_START + XKVRANGE_VC_KASAN_OFFSET)
> >
> > -extern bool kasan_early_stage;
> >  extern unsigned char kasan_early_shadow_page[PAGE_SIZE];
> >
> >  #define kasan_mem_to_shadow kasan_mem_to_shadow
> > @@ -75,12 +74,6 @@ void *kasan_mem_to_shadow(const void *addr);
> >  #define kasan_shadow_to_mem kasan_shadow_to_mem
> >  const void *kasan_shadow_to_mem(const void *shadow_addr);
> >
> > -#define kasan_arch_is_ready kasan_arch_is_ready
> > -static __always_inline bool kasan_arch_is_ready(void)
> > -{
> > -     return !kasan_early_stage;
> > -}
> > -
> >  #define addr_has_metadata addr_has_metadata
> >  static __always_inline bool addr_has_metadata(const void *addr)
> >  {
> > diff --git a/arch/loongarch/mm/kasan_init.c b/arch/loongarch/mm/kasan_init.c
> > index d2681272d8f..cf8315f9119 100644
> > --- a/arch/loongarch/mm/kasan_init.c
> > +++ b/arch/loongarch/mm/kasan_init.c
> > @@ -40,11 +40,9 @@ static pgd_t kasan_pg_dir[PTRS_PER_PGD] __initdata __aligned(PAGE_SIZE);
> >  #define __pte_none(early, pte) (early ? pte_none(pte) : \
> >  ((pte_val(pte) & _PFN_MASK) == (unsigned long)__pa(kasan_early_shadow_page)))
> >
> > -bool kasan_early_stage = true;
> > -
> >  void *kasan_mem_to_shadow(const void *addr)
> >  {
> > -     if (!kasan_arch_is_ready()) {
> > +     if (!kasan_enabled()) {
> >               return (void *)(kasan_early_shadow_page);
> >       } else {
> >               unsigned long maddr = (unsigned long)addr;
> > @@ -298,7 +296,7 @@ void __init kasan_init(void)
> >       kasan_populate_early_shadow(kasan_mem_to_shadow((void *)VMALLOC_START),
> >                                       kasan_mem_to_shadow((void *)KFENCE_AREA_END));
> >
> > -     kasan_early_stage = false;
> > +     kasan_init_generic();
> >
> >       /* Populate the linear mapping */
> >       for_each_mem_range(i, &pa_start, &pa_end) {
> > @@ -329,5 +327,4 @@ void __init kasan_init(void)
> >
> >       /* At this point kasan is fully initialized. Enable error messages */
> >       init_task.kasan_depth = 0;
> > -     pr_info("KernelAddressSanitizer initialized.\n");
> >  }
>
> This one is weird because its the only arch that does things after
> marking early_state false.
>
> Is that really correct, or should kasan_init_generic() be last, like all
> the other architectures?

It really differs from other arch kasan_init(). I can't verify that
kasan_init_generic()
can be placed at the end of kasan_init() because right after
switching the KASAN flag, there's kasan_enabled() check in
kasan_mem_to_shadow().

In arch/loongarch/mm/kasan_init.c:

void *kasan_mem_to_shadow(const void *addr)
{
        if (!kasan_enabled()) {
                return (void *)(kasan_early_shadow_page);
        } else {
...
}

void __init kasan_init(void)
{
...
        kasan_populate_early_shadow(kasan_mem_to_shadow((void *)VMALLOC_START),
        kasan_mem_to_shadow((void *)KFENCE_AREA_END));

        kasan_init_generic();

        /* Populate the linear mapping */
        for_each_mem_range(i, &pa_start, &pa_end) {
....
        kasan_map_populate((unsigned long)kasan_mem_to_shadow(start),
}

>
> Also, please move init_task.kasan_depth = 0 into the generic thing.
> ARM64 might have fooled you with the wrapper function, but they all do
> this right before that pr_info you're taking out.

Please check "[PATCH 1/9] kasan: unify static kasan_flag_enabled across modes",
where I've replied to Christophe:
https://lore.kernel.org/all/CACzwLxj3KWdy-mBu-te1OFf2FZ8eTp5CieYswF5NVY4qPWD93Q@mail.gmail.com/

I can try to put `init_task.kasan_depth = 0;` in kasan_init_generic(),
but in ARM64 kasan_init() we'll still need to have this line for
HW_TAGS, SW_TAGS mode.