Drivers: hv: vmbus: do not mark HV_PCIE as perf_device
authorDexuan Cui <decui@microsoft.com>
Tue, 27 Mar 2018 22:01:02 +0000 (15:01 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 20 Apr 2018 06:20:41 +0000 (08:20 +0200)
commit 238064f13d057390a8c5e1a6a80f4f0a0ec46499 upstream.

The pci-hyperv driver's channel callback hv_pci_onchannelcallback() is not
really a hot path, so we don't need to mark it as a perf_device, meaning
with this patch all HV_PCIE channels' target_cpu will be CPU0.

Signed-off-by: Dexuan Cui <decui@microsoft.com>
Cc: stable@vger.kernel.org
Cc: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/hv/channel_mgmt.c

index d8bc4b91019225f5c83d79e68a67f858cfa9b56e..9360cdce740e8591f007dcf21033dfa1d25f3146 100644 (file)
@@ -70,7 +70,7 @@ static const struct vmbus_device vmbus_devs[] = {
        /* PCIE */
        { .dev_type = HV_PCIE,
          HV_PCIE_GUID,
-         .perf_device = true,
+         .perf_device = false,
        },
 
        /* Synthetic Frame Buffer */