Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20210921101014.1938382-1-elver@google.com> <20210921101014.1938382-4-elver@google.com>
 <CACT4Y+Z6Rss3+oiN5bcKHYeQgG=nZ9VDqwrhOS4VUZ=_a5NoBw@mail.gmail.com>
In-Reply-To: <CACT4Y+Z6Rss3+oiN5bcKHYeQgG=nZ9VDqwrhOS4VUZ=_a5NoBw@mail.gmail.com>
From:   Marco Elver <elver@google.com>
Date:   Tue, 21 Sep 2021 13:35:42 +0200
Message-ID: <CANpmjNPuW47hwmLm=RXr6sXSzvAmz0_vo3m9UGgUbT_CQ=oSgg@mail.gmail.com>
Subject: Re: [PATCH v2 4/5] kfence: limit currently covered allocations when
 pool nearly full
To:     Dmitry Vyukov <dvyukov@google.com>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Alexander Potapenko <glider@google.com>,
        Jann Horn <jannh@google.com>,
        Aleksandr Nogikh <nogikh@google.com>,
        Taras Madan <tarasmadan@google.com>,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        kasan-dev@googlegroups.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, 21 Sept 2021 at 13:05, Dmitry Vyukov <dvyukov@google.com> wrote:
[...]
> > +/*
> > + * Adds (or subtracts) count @val for allocation stack trace hash
> > + * @alloc_stack_hash from Counting Bloom filter.
> > + */
> > +static void alloc_covered_add(u32 alloc_stack_hash, int val)
> > +{
> > +       int i;
> > +
> > +       if (!alloc_stack_hash)
> > +               return;
>
> Nice!
> I like the hash seed, non-evicting cache and that threshold become a
> command line parameter.
>
> This check is the only place I don't understand. What's special about
> alloc_stack_hash == 0? I see that even double-free's won't call this
> with 0.

Indeed, it's no longer required (I think it was in a previous version
I played with). At this point, it should only be true if jhash()
returns 0 for something, and in that case might not even want the
check.

I can also remove the unnecessary "meta->alloc_stack_hash = 0;" in
kfence_guarded_free().

Unless I hear otherwise, I'll remove the unneeded code (and send a v3
of this series in a few days).