2021-03-09 21:44:48

by Kees Cook

Subject: [PATCH v5 3/7] init_on_alloc: Unpessimize default-on builds

Right now, the state of CONFIG_INIT_ON_ALLOC_DEFAULT_ON (and
...ON_FREE...) did not change the assembly ordering of the static branch
tests. Use the new jump_label macro to check CONFIG settings to default
to the "expected" state, unpessimizes the resulting assembly code.

Reviewed-by: Alexander Potapenko <[email protected]>
Link: https://lore.kernel.org/lkml/CAG_fn=X0DVwqLaHJTO6Jw7TGcMSm77GKHinrd0m_6y0SzWOrFA@mail.gmail.com/
Signed-off-by: Kees Cook <[email protected]>
---
include/linux/mm.h | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index bf341a9bfe46..2ccd856ac0d1 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2874,7 +2874,8 @@ static inline void kernel_unpoison_pages(struct page *page, int numpages) { }
DECLARE_STATIC_KEY_MAYBE(CONFIG_INIT_ON_ALLOC_DEFAULT_ON, init_on_alloc);
static inline bool want_init_on_alloc(gfp_t flags)
{
- if (static_branch_unlikely(&init_on_alloc))
+ if (static_branch_maybe(CONFIG_INIT_ON_ALLOC_DEFAULT_ON,
+ &init_on_alloc))
return true;
return flags & __GFP_ZERO;
}
@@ -2882,7 +2883,8 @@ static inline bool want_init_on_alloc(gfp_t flags)
DECLARE_STATIC_KEY_MAYBE(CONFIG_INIT_ON_FREE_DEFAULT_ON, init_on_free);
static inline bool want_init_on_free(void)
{
- return static_branch_unlikely(&init_on_free);
+ return static_branch_maybe(CONFIG_INIT_ON_FREE_DEFAULT_ON,
+ &init_on_free);
}

extern bool _debug_pagealloc_enabled_early;
--
2.25.1

2021-03-10 12:54:11

by Andrey Konovalov

Subject: Re: [PATCH v5 3/7] init_on_alloc: Unpessimize default-on builds

On Tue, Mar 9, 2021 at 10:43 PM Kees Cook <[email protected]> wrote:
>
> Right now, the state of CONFIG_INIT_ON_ALLOC_DEFAULT_ON (and
> ...ON_FREE...) did not change the assembly ordering of the static branch
> tests. Use the new jump_label macro to check CONFIG settings to default
> to the "expected" state, unpessimizes the resulting assembly code.
>
> Reviewed-by: Alexander Potapenko <[email protected]>
> Link: https://lore.kernel.org/lkml/CAG_fn=X0DVwqLaHJTO6Jw7TGcMSm77GKHinrd0m_6y0SzWOrFA@mail.gmail.com/
> Signed-off-by: Kees Cook <[email protected]>
> ---
> include/linux/mm.h | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index bf341a9bfe46..2ccd856ac0d1 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -2874,7 +2874,8 @@ static inline void kernel_unpoison_pages(struct page *page, int numpages) { }
> DECLARE_STATIC_KEY_MAYBE(CONFIG_INIT_ON_ALLOC_DEFAULT_ON, init_on_alloc);
> static inline bool want_init_on_alloc(gfp_t flags)
> {
> - if (static_branch_unlikely(&init_on_alloc))
> + if (static_branch_maybe(CONFIG_INIT_ON_ALLOC_DEFAULT_ON,
> + &init_on_alloc))
> return true;
> return flags & __GFP_ZERO;
> }
> @@ -2882,7 +2883,8 @@ static inline bool want_init_on_alloc(gfp_t flags)
> DECLARE_STATIC_KEY_MAYBE(CONFIG_INIT_ON_FREE_DEFAULT_ON, init_on_free);
> static inline bool want_init_on_free(void)
> {
> - return static_branch_unlikely(&init_on_free);
> + return static_branch_maybe(CONFIG_INIT_ON_FREE_DEFAULT_ON,
> + &init_on_free);
> }
>
> extern bool _debug_pagealloc_enabled_early;

Should we also update slab_want_init_on_alloc() and slab_want_init_on_free()?

2021-03-10 21:05:27

by Kees Cook

Subject: Re: [PATCH v5 3/7] init_on_alloc: Unpessimize default-on builds

On Wed, Mar 10, 2021 at 01:52:04PM +0100, Andrey Konovalov wrote:
> On Tue, Mar 9, 2021 at 10:43 PM Kees Cook <[email protected]> wrote:
> >
> > Right now, the state of CONFIG_INIT_ON_ALLOC_DEFAULT_ON (and
> > ...ON_FREE...) did not change the assembly ordering of the static branch
> > tests. Use the new jump_label macro to check CONFIG settings to default
> > to the "expected" state, unpessimizes the resulting assembly code.
> >
> > Reviewed-by: Alexander Potapenko <[email protected]>
> > Link: https://lore.kernel.org/lkml/CAG_fn=X0DVwqLaHJTO6Jw7TGcMSm77GKHinrd0m_6y0SzWOrFA@mail.gmail.com/
> > Signed-off-by: Kees Cook <[email protected]>
> > ---
> > include/linux/mm.h | 6 ++++--
> > 1 file changed, 4 insertions(+), 2 deletions(-)
> >
> > diff --git a/include/linux/mm.h b/include/linux/mm.h
> > index bf341a9bfe46..2ccd856ac0d1 100644
> > --- a/include/linux/mm.h
> > +++ b/include/linux/mm.h
> > @@ -2874,7 +2874,8 @@ static inline void kernel_unpoison_pages(struct page *page, int numpages) { }
> > DECLARE_STATIC_KEY_MAYBE(CONFIG_INIT_ON_ALLOC_DEFAULT_ON, init_on_alloc);
> > static inline bool want_init_on_alloc(gfp_t flags)
> > {
> > - if (static_branch_unlikely(&init_on_alloc))
> > + if (static_branch_maybe(CONFIG_INIT_ON_ALLOC_DEFAULT_ON,
> > + &init_on_alloc))
> > return true;
> > return flags & __GFP_ZERO;
> > }
> > @@ -2882,7 +2883,8 @@ static inline bool want_init_on_alloc(gfp_t flags)
> > DECLARE_STATIC_KEY_MAYBE(CONFIG_INIT_ON_FREE_DEFAULT_ON, init_on_free);
> > static inline bool want_init_on_free(void)
> > {
> > - return static_branch_unlikely(&init_on_free);
> > + return static_branch_maybe(CONFIG_INIT_ON_FREE_DEFAULT_ON,
> > + &init_on_free);
> > }
> >
> > extern bool _debug_pagealloc_enabled_early;
>
> Should we also update slab_want_init_on_alloc() and slab_want_init_on_free()?

Whoops! Thank you; I will update and resend. :)

--
Kees Cook