Re: [PATCH v8 30/43] arm64: kasan: Allow enabling in-kernel MTE

[Andrey Konovalov]

On Thu, Nov 5, 2020 at 12:13 PM Vincenzo Frascino
<vincenzo.frascino@xxxxxxx> wrote:
>
> Hi Andrey,
>
> On 11/4/20 11:18 PM, Andrey Konovalov wrote:
> > Hardware tag-based KASAN relies on Memory Tagging Extension (MTE)
> > feature and requires it to be enabled. MTE supports
> >
> > This patch adds a new mte_init_tags() helper, that enables MTE in
> > Synchronous mode in EL1 and is intended to be called from KASAN runtime
> > during initialization.
> >
> > The Tag Checking operation causes a synchronous data abort as
> > a consequence of a tag check fault when MTE is configured in
> > synchronous mode.
> >
> > As part of this change enable match-all tag for EL1 to allow the
> > kernel to access user pages without faulting. This is required because
> > the kernel does not have knowledge of the tags set by the user in a
> > page.
> >
> > Note: For MTE, the TCF bit field in SCTLR_EL1 affects only EL1 in a
> > similar way as TCF0 affects EL0.
> >
> > MTE that is built on top of the Top Byte Ignore (TBI) feature hence we
> > enable it as part of this patch as well.
> >
>
> seems that in this patch you dropped me as author. Would you mind to clarify the
> reason?

Sorry, a mistake while squashing/rebasing, will fix in the next version.

>
> > Signed-off-by: Vincenzo Frascino <vincenzo.frascino@xxxxxxx>
> > Co-developed-by: Andrey Konovalov <andreyknvl@xxxxxxxxxx>
> > Signed-off-by: Andrey Konovalov <andreyknvl@xxxxxxxxxx>
> > ---
> > Change-Id: I4d67497268bb7f0c2fc5dcacefa1e273df4af71d
> > ---
> >  arch/arm64/include/asm/mte-kasan.h |  6 ++++++
> >  arch/arm64/kernel/mte.c            |  7 +++++++
> >  arch/arm64/mm/proc.S               | 23 ++++++++++++++++++++---
> >  3 files changed, 33 insertions(+), 3 deletions(-)
> >
> > diff --git a/arch/arm64/include/asm/mte-kasan.h b/arch/arm64/include/asm/mte-kasan.h
> > index 3a70fb1807fd..ae75feaea2d4 100644
> > --- a/arch/arm64/include/asm/mte-kasan.h
> > +++ b/arch/arm64/include/asm/mte-kasan.h
> > @@ -29,6 +29,8 @@ u8 mte_get_mem_tag(void *addr);
> >  u8 mte_get_random_tag(void);
> >  void *mte_set_mem_tag_range(void *addr, size_t size, u8 tag);
> >
> > +void __init mte_init_tags(u64 max_tag);
> > +
> >  #else /* CONFIG_ARM64_MTE */
> >
> >  static inline u8 mte_get_ptr_tag(void *ptr)
> > @@ -49,6 +51,10 @@ static inline void *mte_set_mem_tag_range(void *addr, size_t size, u8 tag)
> >       return addr;
> >  }
> >
> > +static inline void mte_init_tags(u64 max_tag)
> > +{
> > +}
> > +
> >  #endif /* CONFIG_ARM64_MTE */
> >
> >  #endif /* __ASSEMBLY__ */
> > diff --git a/arch/arm64/kernel/mte.c b/arch/arm64/kernel/mte.c
> > index 06ba6c923ab7..fcfbefcc3174 100644
> > --- a/arch/arm64/kernel/mte.c
> > +++ b/arch/arm64/kernel/mte.c
> > @@ -121,6 +121,13 @@ void *mte_set_mem_tag_range(void *addr, size_t size, u8 tag)
> >       return ptr;
> >  }
> >
> > +void __init mte_init_tags(u64 max_tag)
> > +{
> > +     /* Enable MTE Sync Mode for EL1. */
> > +     sysreg_clear_set(sctlr_el1, SCTLR_ELx_TCF_MASK, SCTLR_ELx_TCF_SYNC);
> > +     isb();
>
> I am fine with the approach of letting cpu_enable_mte() call directly
> kasan_init_tags(), but how does it work of the other 2 implementation of KASAN?
> Is it still called in arch_setup()?

Yes, the other 2 modes are initialized in setup_arch().

> I would prefer to keep the code that initializes the sync mode in
> cpu_enable_mte() (calling kasan_init_tags() before then that)

This won't work, we'll later need to make the decision about whether
to turn on MTE at all in KASAN runtime based on KASAN boot flags.

> or in a separate
> function since setting the mode has nothing to do with initializing the tags.

This will work. Any preference on the name of this function?

Alternatively we can rename mte_init_tags() to something else and let
it handle both RRND and sync/async.

Thanks!