Re: [PATCH 10/12] arm64: kasan: simplify and inline MTE functions

From: Andrey Konovalov <andreyknvl@google.com>
To: Catalin Marinas <catalin.marinas@arm.com>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>,
	Dmitry Vyukov <dvyukov@google.com>,
	Alexander Potapenko <glider@google.com>,
	Marco Elver <elver@google.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Will Deacon <will.deacon@arm.com>,
	Andrey Ryabinin <aryabinin@virtuozzo.com>,
	Peter Collingbourne <pcc@google.com>,
	Evgenii Stepanov <eugenis@google.com>,
	Branislav Rankov <Branislav.Rankov@arm.com>,
	Kevin Brodsky <kevin.brodsky@arm.com>,
	kasan-dev <kasan-dev@googlegroups.com>,
	Linux ARM <linux-arm-kernel@lists.infradead.org>,
	Linux Memory Management List <linux-mm@kvack.org>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 10/12] arm64: kasan: simplify and inline MTE functions
Date: Tue, 2 Feb 2021 19:04:38 +0100	[thread overview]
Message-ID: <CAAeHK+y31RvnR2UPtekuscAd=Ogk5zouW_kzxPm7-mVotpqQOA@mail.gmail.com> (raw)
In-Reply-To: <20210202154200.GC26895@gaia>

On Tue, Feb 2, 2021 at 4:42 PM Catalin Marinas <catalin.marinas@arm.com> wrote:
>
> On Mon, Feb 01, 2021 at 08:43:34PM +0100, Andrey Konovalov wrote:
> > +/*
> > + * Assign allocation tags for a region of memory based on the pointer tag.
> > + * Note: The address must be non-NULL and MTE_GRANULE_SIZE aligned and
> > + * size must be non-zero and MTE_GRANULE_SIZE aligned.
> > + */
>
> OK, so we rely on the caller to sanity-check the range. Fine by me but I
> can see (un)poison_range() only doing this for the size. Do we guarantee
> that the start address is aligned?

See the previous patch in the series. kasan_poison() checks and warns
on both unaligned addr and size. kasan_unpoison() checks addr and
rounds up size.

> > +static __always_inline void mte_set_mem_tag_range(void *addr, size_t size, u8 tag)
> > +{
> > +     u64 curr, end;
> > +
> > +     if (!size)
> > +             return;
> > +
> > +     curr = (u64)__tag_set(addr, tag);
> > +     end = curr + size;
> > +
> > +     do {
> > +             /*
> > +              * 'asm volatile' is required to prevent the compiler to move
> > +              * the statement outside of the loop.
> > +              */
> > +             asm volatile(__MTE_PREAMBLE "stg %0, [%0]"
> > +                          :
> > +                          : "r" (curr)
> > +                          : "memory");
> > +
> > +             curr += MTE_GRANULE_SIZE;
> > +     } while (curr != end);
> > +}
> >
> >  void mte_enable_kernel_sync(void);
> >  void mte_enable_kernel_async(void);
> > @@ -47,10 +95,12 @@ static inline u8 mte_get_mem_tag(void *addr)
> >  {
> >       return 0xFF;
> >  }
> > +
> >  static inline u8 mte_get_random_tag(void)
> >  {
> >       return 0xFF;
> >  }
> > +
> >  static inline void *mte_set_mem_tag_range(void *addr, size_t size, u8 tag)
>
> This function used to return a pointer and that's what the dummy static
> inline does here. However, the new mte_set_mem_tag_range() doesn't
> return anything. We should have consistency between the two (the new
> static void definition is fine by me).

Right, forgot to update the empty function definition. Will do in v2.

>
> Otherwise the patch looks fine.
>
> Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>

Thanks!