From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kiszka Subject: Re: [PATCH] pci-assign: Fix multifunction support Date: Tue, 17 Jan 2012 15:28:36 +0100 Message-ID: <4F158594.2060605@siemens.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: "kvm@vger.kernel.org" , "mtosatti@redhat.com" To: Alex Williamson Return-path: Received: from thoth.sbs.de ([192.35.17.2]:31289 "EHLO thoth.sbs.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753524Ab2AQO2m (ORCPT ); Tue, 17 Jan 2012 09:28:42 -0500 In-Reply-To: Sender: kvm-owner@vger.kernel.org List-ID: On 2012-01-17 14:23, Alex Williamson wrote: > > > ----- Original Message ----- >> On 2012-01-16 18:11, Alex Williamson wrote: >>> The core PCI code sets the multifunction bit in the header before >>> calling the device initfn. For device assignment, we're blasting >>> that value with the actual hardware value, so nobody sees the >>> additional functions if the devices isn't physically multifunction. >>> Switch the HEADER_TYPE to a fully emulated field (all read-only >>> anyway) and add setting and clearing of the multifunction bit to >>> match qemu directive. >>> >>> Signed-off-by: Alex Williamson >>> --- >>> >>> hw/device-assignment.c | 8 +++++++- >>> 1 files changed, 7 insertions(+), 1 deletions(-) >>> >>> diff --git a/hw/device-assignment.c b/hw/device-assignment.c >>> index 2a9e66d..7f4a5ec 100644 >>> --- a/hw/device-assignment.c >>> +++ b/hw/device-assignment.c >>> @@ -540,6 +540,13 @@ again: >>> fprintf(stderr, "%s: read failed, errno = %d\n", __func__, >>> errno); >>> } >>> >>> + /* Restore or clear multifunction, this is always controlled >>> by qemu */ >>> + if (pci_dev->dev.cap_present & QEMU_PCI_CAP_MULTIFUNCTION) { >>> + pci_dev->dev.config[PCI_HEADER_TYPE] |= >>> PCI_HEADER_TYPE_MULTI_FUNCTION; >>> + } else { >>> + pci_dev->dev.config[PCI_HEADER_TYPE] &= >>> ~PCI_HEADER_TYPE_MULTI_FUNCTION; >>> + } >>> + >> >> Why have this in get_*real*_device? Why not fix this up at the caller >> site, i.e. in assigned_initfn? Just for consistency, not a functional >> issue. > > I chose here because we've just overwritten the emulated config space and we then proceed to clean out the BAR registers. As this is close to the point where it gets trashed and we're doing other fixup, it seems appropriate. OK. Jan -- Siemens AG, Corporate Technology, CT T DE IT 1 Corporate Competence Center Embedded Linux