From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:37521) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VRUiV-0004pO-5u for qemu-devel@nongnu.org; Wed, 02 Oct 2013 18:16:29 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VRUiN-0005Sd-LJ for qemu-devel@nongnu.org; Wed, 02 Oct 2013 18:16:23 -0400 Received: from mx1.redhat.com ([209.132.183.28]:53974) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VRUiN-0005SO-Dq for qemu-devel@nongnu.org; Wed, 02 Oct 2013 18:16:15 -0400 Message-ID: <1380752181.14443.15.camel@localhost.localdomain> From: Marcel Apfelbaum Date: Thu, 03 Oct 2013 01:16:21 +0300 In-Reply-To: <1380729490.14271.84.camel@ul30vt.home> References: <1380717694-13091-1-git-send-email-marcel.a@redhat.com> <1380717694-13091-6-git-send-email-marcel.a@redhat.com> <1380729490.14271.84.camel@ul30vt.home> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH RFC v2 5/9] hw/vfio: set interrupts using pci irq wrappers List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Alex Williamson Cc: kwolf@redhat.com, peter.maydell@linaro.org, peter.crosthwaite@xilinx.com, anthony@codemonkey.ws, mst@redhat.com, sw@weilnetz.de, jasowang@redhat.com, qemu-devel@nongnu.org, dkoch@verizon.com, keith.busch@intel.com, kraxel@redhat.com, stefanha@redhat.com, dmitry@daynix.com, pbonzini@redhat.com, afaerber@suse.de, ehabkost@redhat.com On Wed, 2013-10-02 at 09:58 -0600, Alex Williamson wrote: > On Wed, 2013-10-02 at 15:41 +0300, Marcel Apfelbaum wrote: > > pci_set_irq and the other pci irq wrappers use > > PCI_INTERRUPT_PIN config register to compute device > > INTx pin to assert/deassert. > > > > Save INTx pin into the config register before calling > > pci_set_irq > > > > Signed-off-by: Marcel Apfelbaum > > --- > > hw/misc/vfio.c | 11 ++++++----- > > 1 file changed, 6 insertions(+), 5 deletions(-) > > Seems ok, but why not take advantage of the pci_irq_raise/lower() > wrappers? BTW, with PCI being active low, should those be > assert/deassert to avoid confusion confusion with the actual signal > level? Thanks, Thanks for the review! I can use pci_irq_raise/lower(), but I wanted to preserve the current form, re-factoring: qemu_set_irq -> pci_set_irq, qemu_irq_lower -> pci_irq_lower ... If you think is worth it, I'll change it. (in all the places) About assert/deassert instead of lower/raise, I am afraid it will confuse users having two different set of naming for interrupts usage. Is easier to understand that pci_irq_lower behaves the same as qemu_pci_lower, then pci_irq_deassert. What do you think? Thanks, Marcel > > Alex > > > > > diff --git a/hw/misc/vfio.c b/hw/misc/vfio.c > > index a1c08fb..3d7297c 100644 > > --- a/hw/misc/vfio.c > > +++ b/hw/misc/vfio.c > > @@ -297,7 +297,7 @@ static void vfio_intx_interrupt(void *opaque) > > 'A' + vdev->intx.pin); > > > > vdev->intx.pending = true; > > - qemu_set_irq(vdev->pdev.irq[vdev->intx.pin], 1); > > + pci_set_irq(&vdev->pdev, 1); > > vfio_mmap_set_enabled(vdev, false); > > if (vdev->intx.mmap_timeout) { > > timer_mod(vdev->intx.mmap_timer, > > @@ -315,7 +315,7 @@ static void vfio_eoi(VFIODevice *vdev) > > vdev->host.bus, vdev->host.slot, vdev->host.function); > > > > vdev->intx.pending = false; > > - qemu_set_irq(vdev->pdev.irq[vdev->intx.pin], 0); > > + pci_set_irq(&vdev->pdev, 0); > > vfio_unmask_intx(vdev); > > } > > > > @@ -341,7 +341,7 @@ static void vfio_enable_intx_kvm(VFIODevice *vdev) > > qemu_set_fd_handler(irqfd.fd, NULL, NULL, vdev); > > vfio_mask_intx(vdev); > > vdev->intx.pending = false; > > - qemu_set_irq(vdev->pdev.irq[vdev->intx.pin], 0); > > + pci_set_irq(&vdev->pdev, 0); > > > > /* Get an eventfd for resample/unmask */ > > if (event_notifier_init(&vdev->intx.unmask, 0)) { > > @@ -417,7 +417,7 @@ static void vfio_disable_intx_kvm(VFIODevice *vdev) > > */ > > vfio_mask_intx(vdev); > > vdev->intx.pending = false; > > - qemu_set_irq(vdev->pdev.irq[vdev->intx.pin], 0); > > + pci_set_irq(&vdev->pdev, 0); > > > > /* Tell KVM to stop listening for an INTx irqfd */ > > if (kvm_vm_ioctl(kvm_state, KVM_IRQFD, &irqfd)) { > > @@ -488,6 +488,7 @@ static int vfio_enable_intx(VFIODevice *vdev) > > vfio_disable_interrupts(vdev); > > > > vdev->intx.pin = pin - 1; /* Pin A (1) -> irq[0] */ > > + pci_config_set_interrupt_pin(vdev->pdev.config, pin); > > > > #ifdef CONFIG_KVM > > /* > > @@ -547,7 +548,7 @@ static void vfio_disable_intx(VFIODevice *vdev) > > vfio_disable_intx_kvm(vdev); > > vfio_disable_irqindex(vdev, VFIO_PCI_INTX_IRQ_INDEX); > > vdev->intx.pending = false; > > - qemu_set_irq(vdev->pdev.irq[vdev->intx.pin], 0); > > + pci_set_irq(&vdev->pdev, 0); > > vfio_mmap_set_enabled(vdev, true); > > > > fd = event_notifier_get_fd(&vdev->intx.interrupt); > > >