Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
MIME-Version: 1.0
References: <cover.1648049113.git.andreyknvl@google.com> <f75c58b17bfaa419f84286cd174e3a08f971b779.1648049113.git.andreyknvl@google.com>
 <YkVzTbafttTHWETU@FVFF77S0Q05N>
In-Reply-To: <YkVzTbafttTHWETU@FVFF77S0Q05N>
From:   Andrey Konovalov <andreyknvl@gmail.com>
Date:   Tue, 5 Apr 2022 17:22:19 +0200
Message-ID: <CA+fCnZekoAMEcS+0905JzP=Gu81R_F_em5Un8JL+FOF2Jj3rqg@mail.gmail.com>
Subject: Re: [PATCH v2 2/4] arm64, scs: save scs_sp values per-cpu when
 switching stacks
To:     Mark Rutland <mark.rutland@arm.com>
Cc:     andrey.konovalov@linux.dev, Marco Elver <elver@google.com>,
        Alexander Potapenko <glider@google.com>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will@kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        Dmitry Vyukov <dvyukov@google.com>,
        Andrey Ryabinin <ryabinin.a.a@gmail.com>,
        kasan-dev <kasan-dev@googlegroups.com>,
        Vincenzo Frascino <vincenzo.frascino@arm.com>,
        Sami Tolvanen <samitolvanen@google.com>,
        Peter Collingbourne <pcc@google.com>,
        Evgenii Stepanov <eugenis@google.com>,
        Florian Mayer <fmayer@google.com>,
        Linux Memory Management List <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Andrey Konovalov <andreyknvl@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Mar 31, 2022 at 11:24 AM Mark Rutland <mark.rutland@arm.com> wrote:
>
> On Wed, Mar 23, 2022 at 04:32:53PM +0100, andrey.konovalov@linux.dev wrote:
> > From: Andrey Konovalov <andreyknvl@google.com>
> >
> > Instead of trying to retrieve the SCS pointers from the stack, change
> > interrupt handlers (for hard IRQ, Normal and Critical SDEI) to save the
> > previous SCS pointer in a per-CPU variable.
>
> I'm *really* not keen on *always* poking this in the entry code for the
> uncommon case of unwind. It complicates the entry code and means we're always
> paying a cost for potentially no benefit. At a high-level, I don't think this
> is the right approach.

This also gives a 5% slowdown, which is not acceptable.

What we can do instead, is to not collect frames from the higher
exception levels at all. This would leave SCS-based stack collection
method impaired, but this is probably fine for KASAN's use case:
currently, stack depot filters out higher-level frames anyway, so
KASAN never saves them. And the lower-level part of the stack trace is
enough to identify the allocation.

Thanks!


> For the regular unwinder, I want to rework things such that we can identify
> exception boundaries and look into the regs (e.g. so that we can recover the
> PC+LR+FP and avoid duplicating part of this in a frame record), and I'd much
> prefer that we did the same here.