Received-SPF: pass (google.com: domain of linux-kernel+bounces-110060-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
From: David Laight <David.Laight@ACULAB.COM>
To: 'Russell King' <linux@armlinux.org.uk>
CC: Ard Biesheuvel <ardb@kernel.org>, 'Jiangfeng Xiao'
	<xiaojiangfeng@huawei.com>, "arnd@arndb.de" <arnd@arndb.de>,
	"keescook@chromium.org" <keescook@chromium.org>, "haibo.li@mediatek.com"
	<haibo.li@mediatek.com>, "angelogioacchino.delregno@collabora.com"
	<angelogioacchino.delregno@collabora.com>, "amergnat@baylibre.com"
	<amergnat@baylibre.com>, "akpm@linux-foundation.org"
	<akpm@linux-foundation.org>, "dave.hansen@linux.intel.com"
	<dave.hansen@linux.intel.com>, "douzhaolei@huawei.com"
	<douzhaolei@huawei.com>, "gustavoars@kernel.org" <gustavoars@kernel.org>,
	"jpoimboe@kernel.org" <jpoimboe@kernel.org>, "kepler.chenxin@huawei.com"
	<kepler.chenxin@huawei.com>, "kirill.shutemov@linux.intel.com"
	<kirill.shutemov@linux.intel.com>, "linux-hardening@vger.kernel.org"
	<linux-hardening@vger.kernel.org>, "linux-kernel@vger.kernel.org"
	<linux-kernel@vger.kernel.org>, "linux-mm@kvack.org" <linux-mm@kvack.org>,
	"linux-arm-kernel@lists.infradead.org"
	<linux-arm-kernel@lists.infradead.org>, "nixiaoming@huawei.com"
	<nixiaoming@huawei.com>, "peterz@infradead.org" <peterz@infradead.org>,
	"wangbing6@huawei.com" <wangbing6@huawei.com>, "wangfangpeng1@huawei.com"
	<wangfangpeng1@huawei.com>, "jannh@google.com" <jannh@google.com>,
	"willy@infradead.org" <willy@infradead.org>
Subject: RE: [PATCH v2] ARM: unwind: improve unwinders for noreturn case
Thread-Topic: [PATCH v2] ARM: unwind: improve unwinders for noreturn case
Thread-Index: AQHae3ROEuI+AaCprEesIWGaAOB7ebFB9uHAgAAWtoCAAAVW8IAACy4AgAADrmA=
Date: Thu, 21 Mar 2024 12:57:07 +0000
Message-ID: <0fd55e156195440bb1d815dd8300894b@AcuMS.aculab.com>
References: <1709516385-7778-1-git-send-email-xiaojiangfeng@huawei.com>
 <1710906278-23851-1-git-send-email-xiaojiangfeng@huawei.com>
 <ZfqiD8Yw0oOVHW/p@shell.armlinux.org.uk>
 <84a57ca8-8963-ca24-8bd1-ddc5c33bf4da@huawei.com>
 <Zfs7sT6Pxy5yjnPC@shell.armlinux.org.uk>
 <bad25937-fc98-8e11-4279-ab6b3a727c1f@huawei.com>
 <bbcca1e205384cf0b42236e17f3969f7@AcuMS.aculab.com>
 <ZfwYx/8k8FVONg6+@shell.armlinux.org.uk>
 <db930076c837456f999daee5cb76735f@AcuMS.aculab.com>
 <ZfwmomjUwQdCefzh@shell.armlinux.org.uk>
In-Reply-To: <ZfwmomjUwQdCefzh@shell.armlinux.org.uk>
Accept-Language: en-GB, en-US
Precedence: bulk
MIME-Version: 1.0
Content-Language: en-US
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

From: Russell King
> Sent: 21 March 2024 12:23
..
> > That might mean you can get the BL in the middle of a function
> > but where the following instruction is for the 'no stack frame'
> > side of the branch.
> > That is very likely to break any stack offset calculations.
>=20
> No it can't. At any one point in the function, the stack has to be in
> a well defined state, so that access to local variables can work, and
> also the stack can be correctly unwound. If there exists a point in
> the function body which can be reached where the stack could be in two
> different states, then the stack can't be restored to the parent
> context.

Actually you can get there with a function that has a lot of args.
So you can have:
=09if (...) {
=09=09push x
=09=09bl func
=09=09add %sp, #8
=09}
=09code;
which is fine.
But if 'func' is 'noreturn' then the 'add %sp, #8' can be discarded
and then the saved LR is that of 'code' - but the stack offset is wrong.

> > > This is where the problem lies - because the link register value
> > > created by the BL instruction will point to the instruction after the
> > > BL which will _not_ part of the function that invoked the BL. That
> > > will probably cause issues for the ELF unwinder, which means this
> > > issue probably goes beyond _just_ printing the function name.
> >
> > Isn't this already in the unwinder?
> > A BL itself isn't going to fault with PC =3D next-instruction.
>=20
> You are missing the fact that the PC can be the saved LR, and thus
> can very well be the next instruction.

A PC from LR will always be the next instruction.
It is only the PC from a fault frame that is the current one.
The unwinder probably need to be told which one it has.
(Or add 4 the fault frame PC so that the unwinder can subtract
4 from it.)

At least (I don't think) there are any functions where the
called code is responsible for removing arguments.
That is a whole different bag of worms.

=09David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1=
PT, UK
Registration No: 1397386 (Wales)