From mboxrd@z Thu Jan 1 00:00:00 1970 From: Josef Vogt Subject: Bug in PCI Passthrough Date: Wed, 20 Apr 2011 11:15:23 +0200 Message-ID: <4DAEA42B.3010600@li-life.li> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: quoted-printable Return-path: List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org Hello all I tried to enable PCI Passthrough in my Debian Squeeze installation.=20 Unfortunately, it didnt work. The following bug showed up (excerpt from=20 dmesg): [ 10.148612] uhci_hcd 0000:00:14.2: PCI INT D -> Link[LNKD] -> GSI 11 (level, low) -> IRQ 11 [ 4.678352] Failed to setup GSI :11, err_code:-22 [ 4.678480] dmfe 0000:00:0e.0: PCI INT A -> Link[LNKB] -> GSI 11 (level, low) -> IRQ 11 [ 6.868419] Failed to setup GSI :10, err_code:-22 Strange thing is, it worked perfectly with Debian Lenny Kernel 2.6.26 [1] I first thought, it was a Debian bug[2], but then I tried the vanilla=20 kernel (xen/stable-2.6.32.x from kernel.org) the bug was exactly the=20 same. See also [1] for the the exact error messages. I also tried to get help from xen users list [3]. Strange side note: It seems to work with opensuse 11.2 (kernel 3.6.31)=20 and also Opensuse 11.3 (kernel 3.6.34) =E2=80=93 I installed them both bu= t didnt=20 get the above error message. The interrupts were assigned correctly. So=20 i started looking at the code, but the used code in Opensuse looks quite=20 different to the xen code in kernel.org. (I didn't find xen.c, where the=20 error message =E2=80=9EFailed to setup GSI=E2=80=9C is located). Is it po= ssible that=20 Opensuse uses Xen source code from an older branch? I also compiled xen/next-2.6.39 (up to this commit:=20 6f3ba85fbef08ce0d833a1943827454b3e5532b7) but then, unfortunately the=20 dom0 crashes. All the above together (except the dom0 crash in 2.6.39, of course)=20 let's me think, this could a bug in xen kernel source code. Greetings! Josef [1]=20 http://lists.xensource.com/archives/html/xen-users/2011-01/msg00411.html [2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D610832 [3]=20 http://lists.xensource.com/archives/html/xen-users/2011-01/msg00388.html