From mboxrd@z Thu Jan 1 00:00:00 1970 From: bugzilla-daemon@freedesktop.org Subject: [Bug 105113] [hawaii, radeonsi, clover] Running Piglit cl/program/execute/{, tail-}calls{, -struct, -workitem-id}.cl cause GPU VM error and ring stalled GPU lockup Date: Fri, 16 Nov 2018 16:13:57 +0000 Message-ID: References: Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1798189606==" Return-path: Received: from culpepper.freedesktop.org (culpepper.freedesktop.org [IPv6:2610:10:20:722:a800:ff:fe98:4b55]) by gabe.freedesktop.org (Postfix) with ESMTP id 8A26E6E71A for ; Fri, 16 Nov 2018 16:13:57 +0000 (UTC) In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" To: dri-devel@lists.freedesktop.org List-Id: dri-devel@lists.freedesktop.org --===============1798189606== Content-Type: multipart/alternative; boundary="15423848371.534D.29858" Content-Transfer-Encoding: 7bit --15423848371.534D.29858 Date: Fri, 16 Nov 2018 16:13:57 +0000 MIME-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Bugzilla-URL: http://bugs.freedesktop.org/ Auto-Submitted: auto-generated https://bugs.freedesktop.org/show_bug.cgi?id=3D105113 --- Comment #6 from Maciej S. Szmigiero --- There are really two issues at play here: 1) If the LLVM-generated code cannot be run properly then it should be simp= ly rejected by whatever is actually in charge of submitting it to the GPU (I g= uess this would be Mesa?). This way an application will know it cannot use OpenCL for computation, at least not with this compute kernel. Instead, it currently looks like many of these test run but give incorrect results, which is obviously rather bad. 2) Some (previous) Mesa + LLVM versions generate a command stream that cras= hes the GPU and, as far as I can remember, sometimes even lockup the whole mach= ine. It should not be possible to crash the GPU, regardless how incorrect a comm= and stream that userspace sends to it is - because otherwise it is possible for an unprivileged user with GPU access to DoS the machine. --=20 You are receiving this mail because: You are the assignee for the bug.= --15423848371.534D.29858 Date: Fri, 16 Nov 2018 16:13:57 +0000 MIME-Version: 1.0 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Bugzilla-URL: http://bugs.freedesktop.org/ Auto-Submitted: auto-generated

Commen= t # 6 on bug 10511= 3 from Maciej S. Szmigiero 
There are really two issues at play here:
1) If the LLVM-generated code cannot be run properly then it should be simp=
ly
rejected by whatever is actually in charge of submitting it to the GPU (I g=
uess
this would be Mesa?).
This way an application will know it cannot use OpenCL for computation, at
least
not with this compute kernel.

Instead, it currently looks like many of these test run but give incorrect
results, which is obviously rather bad.

2) Some (previous) Mesa + LLVM versions generate a command stream that cras=
hes
the GPU and, as far as I can remember, sometimes even lockup the whole mach=
ine.

It should not be possible to crash the GPU, regardless how incorrect a comm=
and
stream that userspace sends to it is - because otherwise it is possible for
an unprivileged user with GPU access to DoS the machine.

You are receiving this mail because:
  • You are the assignee for the bug.
= --15423848371.534D.29858-- --===============1798189606== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: base64 Content-Disposition: inline X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KZHJpLWRldmVs IG1haWxpbmcgbGlzdApkcmktZGV2ZWxAbGlzdHMuZnJlZWRlc2t0b3Aub3JnCmh0dHBzOi8vbGlz dHMuZnJlZWRlc2t0b3Aub3JnL21haWxtYW4vbGlzdGluZm8vZHJpLWRldmVsCg== --===============1798189606==--