yeahuh / rpms / qemu-kvm

Forked from rpms/qemu-kvm 2 years ago
Clone
Blob Blame History Raw
From aad727a5ecde1ad4935eb8427604d4df5a1f1f35 Mon Sep 17 00:00:00 2001
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Date: Fri, 5 May 2017 19:06:14 +0200
Subject: [PATCH 2/7] usb-xhci: Fix PCI capability order

RH-Author: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-id: <20170505190614.15987-2-dgilbert@redhat.com>
Patchwork-id: 75038
O-Subject: [RHEL-7.4 qemu-kvm-rhev PATCH 1/1] usb-xhci: Fix PCI capability order
Bugzilla: 1447874
RH-Acked-by: Laszlo Ersek <lersek@redhat.com>
RH-Acked-by: Michael S. Tsirkin <mst@redhat.com>
RH-Acked-by: Gerd Hoffmann <kraxel@redhat.com>
RH-Acked-by: Juan Quintela <quintela@redhat.com>

From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>

bz: https://bugzilla.redhat.com/show_bug.cgi?id=1447874
brew: https://brewweb.engineering.redhat.com/brew/taskinfo?taskID=13141090
upstream: No, see below

Upstream commit 1108b2f8a9 in 2.7.0 changed the order
of the PCI capability chain in the XHCI pci device in the case
where the device has the PCIe endpoint capability (i.e. only
older machine types, pc-i440fx-2.0 upstream, pc-i440fx-rhel7.0.0
apparently for us).

Changing the order breaks migration compatibility; fixing this
upstream would mean breaking the same case going from 2.7.0->current
that currently works 2.7.0->2.9.0 - so upstream it's a choice
of two breakages.

Since we never released 2.7.0/2.8.0 we can fix this downstream.

This reverts the order so that we create the capabilities in the
order:
   PCIe
   MSI
   MSI-X

The symptom is:
qemu-kvm: get_pci_config_device: Bad config data: i=0x71 read: a0 device: 0 cmask: ff wmask: 0 w1cmask:0
qemu-kvm: Failed to load PCIDevice:config
qemu-kvm: Failed to load xhci:parent_obj
qemu-kvm: error while loading state for instance 0x0 of device '0000:00:0d.0/xhci'
qemu-kvm: load of migration failed: Invalid argument

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
---
 hw/usb/hcd-xhci.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/hw/usb/hcd-xhci.c b/hw/usb/hcd-xhci.c
index c38594c..1555aa4 100644
--- a/hw/usb/hcd-xhci.c
+++ b/hw/usb/hcd-xhci.c
@@ -3575,6 +3575,12 @@ static void usb_xhci_realize(struct PCIDevice *dev, Error **errp)
         xhci->max_pstreams_mask = 0;
     }
 
+    if (pci_bus_is_express(dev->bus) ||
+        xhci_get_flag(xhci, XHCI_FLAG_FORCE_PCIE_ENDCAP)) {
+        ret = pcie_endpoint_cap_init(dev, 0xa0);
+        assert(ret >= 0);
+    }
+
     if (xhci->msi != ON_OFF_AUTO_OFF) {
         ret = msi_init(dev, 0x70, xhci->numintrs, true, false, &err);
         /* Any error other than -ENOTSUP(board's MSI support is broken)
@@ -3623,12 +3629,6 @@ static void usb_xhci_realize(struct PCIDevice *dev, Error **errp)
                      PCI_BASE_ADDRESS_SPACE_MEMORY|PCI_BASE_ADDRESS_MEM_TYPE_64,
                      &xhci->mem);
 
-    if (pci_bus_is_express(dev->bus) ||
-        xhci_get_flag(xhci, XHCI_FLAG_FORCE_PCIE_ENDCAP)) {
-        ret = pcie_endpoint_cap_init(dev, 0xa0);
-        assert(ret >= 0);
-    }
-
     if (xhci->msix != ON_OFF_AUTO_OFF) {
         /* TODO check for errors, and should fail when msix=on */
         msix_init(dev, xhci->numintrs,
-- 
1.8.3.1