From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id 235703858404; Wed, 31 Aug 2022 16:35:30 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 235703858404
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1661963730;
	bh=j+HASOLjJblL0NKtkhSaZ38w9rzX/umA9O7b+uwX4EY=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=SI0pyVzPoMKgW823UrXHbJhNuD/hJ4SPp66gCXXcmsu5DHsDdw0QvscFLliOYxzet
	 cvdr8dvYsCA1uu/7D0viPDOVclpf82pVyVsoMrOQeguk200HAnXwz7tUwh90vyE5Np
	 urzxt0hIjXmOZvUH1nbDhBJCHim0VEF7Uo/h1pvY=
From: "rearnsha at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/106763] Armv8.2 vmov.f16 instruction sometimes causes
 SIGILL
Date: Wed, 31 Aug 2022 16:35:29 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 12.2.1
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rearnsha at gcc dot gnu.org
X-Bugzilla-Status: RESOLVED
X-Bugzilla-Resolution: INVALID
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-106763-4-3rPPIKIveP@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-106763-4@http.gcc.gnu.org/bugzilla/>
References: <bug-106763-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D106763

--- Comment #8 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
I spoke to our kernel experts about this and they think my hypothesis is qu=
ite
likely to be correct.  They also noted that kernel version 4.9.118 is about=
 200
releases out of date on the 4.9 LTS series.

But more importantly, they say that trying to run a 32-bit kernel on 64-bit=
 Arm
v8 hardware is not recommended and that you should seriously consider switc=
hing
to a 64-bit kernel (which can still run a 32-bit userland if you need it to=
).

I do note that the code here:
=20
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arc=
h/arm/kernel/entry-armv.S#n628=20
does not have an entry for CP#9, which would be needed for FP16 support as =
all
FP16 instructions are in that part of the encoding space.  I have no idea if
changing that would be enough.=