From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:45746) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aqhBo-0005NO-Ej for qemu-devel@nongnu.org; Thu, 14 Apr 2016 09:20:09 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1aqhBk-0002yR-LT for qemu-devel@nongnu.org; Thu, 14 Apr 2016 09:20:08 -0400 Received: from mail-vk0-x22e.google.com ([2607:f8b0:400c:c05::22e]:34908) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aqhBk-0002xv-H5 for qemu-devel@nongnu.org; Thu, 14 Apr 2016 09:20:04 -0400 Received: by mail-vk0-x22e.google.com with SMTP id t129so108799899vkg.2 for ; Thu, 14 Apr 2016 06:20:04 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <570F93E9.1050204@gmail.com> References: <570F93E9.1050204@gmail.com> From: Peter Maydell Date: Thu, 14 Apr 2016 14:19:44 +0100 Message-ID: Content-Type: text/plain; charset=UTF-8 Subject: Re: [Qemu-devel] IRQ number used by virtio-net-device in vexpress-a15 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Inti Gonzalez-Herrera Cc: QEMU Developers On 14 April 2016 at 13:58, Inti Gonzalez-Herrera wrote: > I have some problems using virtio-net-device. In particular, th system is > not receiving an interrupt when it transmits a package. However, I can see > how the buffer descriptors are moved to the used ring. I was thinking that > maybe the problem is related to the fact that I don't know what is the exact > IRQ number. > > In other words, let's say that I have a function enable_irq(int, handler) > that enables a given interrupt in the GIC. What is the number n that I > should used as parameter in this platform? > > Right now, I'm using the value (40 + k), where k is the index of the > specific transporter where the network device is attached (from 0 to 3). It > is not working. > I get that information from file hw/arm/vexpress.c, functions > vexpress_common_init and vexpress_modify_dtb (this one is not executed) Problems with interrupt numbers are almost always related to confusion about whether the interrupt number starts with the 0 as the first external interrupt, or 0 as the first internal interrupt (where 0..31 are SGIs and PPIs, and the first external (SPI) interrupt is 32). The '40 + k' in the vexpress.c source code is an external interrupt number (in which numbering scheme the PL011 UARTs are interrupts 5 to 8, and the sp804 timer is interrupt 48, if that helps you cross-check against some other devices in your OS which do have working interrupts). thanks -- PMM