RE: [PATCH v3 3/6] PCI: hv: serialize the present/eject work items
> -Original Message- > From: Dexuan Cui > Sent: Tuesday, March 6, 2018 1:22 PM > To: bhelg...@google.com; linux-...@vger.kernel.org; KY Srinivasan >; Stephen Hemminger ; > o...@aepfle.de; a...@canonical.com; jasow...@redhat.com > Cc: linux-ker...@vger.kernel.org; driverdev-devel@linuxdriverproject.org; > Haiyang Zhang ; vkuzn...@redhat.com; > marcelo.ce...@canonical.com; Michael Kelley (EOSG) > ; Dexuan Cui ; Jack > Morgenstein ; sta...@vger.kernel.org > Subject: [PATCH v3 3/6] PCI: hv: serialize the present/eject work items > > When we hot-remove the device, we first receive a PCI_EJECT message and > then receive a PCI_BUS_RELATIONS message with bus_rel->device_count == 0. > > The first message is offloaded to hv_eject_device_work(), and the second is > offloaded to pci_devices_present_work(). Both the paths can be running > list_del(>list_entry), causing general protection fault, because > system_wq can run them concurrently. > > The patch eliminates the race condition. > > Signed-off-by: Dexuan Cui > Tested-by: Adrian Suhov > Tested-by: Chris Valean > Cc: Vitaly Kuznetsov > Cc: Jack Morgenstein > Cc: sta...@vger.kernel.org > Cc: Stephen Hemminger > Cc: K. Y. Srinivasan > --- Acked-by: Haiyang Zhang ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
RE: [PATCH v3 3/6] PCI: hv: serialize the present/eject work items
> -Original Message- > From: Dexuan Cui > Sent: Tuesday, March 6, 2018 10:22 AM > To: bhelg...@google.com; linux-...@vger.kernel.org; KY Srinivasan >; > Stephen Hemminger ; o...@aepfle.de; > a...@canonical.com; > jasow...@redhat.com > Cc: linux-ker...@vger.kernel.org; driverdev-devel@linuxdriverproject.org; > Haiyang Zhang > ; vkuzn...@redhat.com; marcelo.ce...@canonical.com; > Michael > Kelley (EOSG) ; Dexuan Cui > ; Jack > Morgenstein ; sta...@vger.kernel.org > Subject: [PATCH v3 3/6] PCI: hv: serialize the present/eject work items > > When we hot-remove the device, we first receive a PCI_EJECT message and > then receive a PCI_BUS_RELATIONS message with bus_rel->device_count == 0. > > The first message is offloaded to hv_eject_device_work(), and the second > is offloaded to pci_devices_present_work(). Both the paths can be running > list_del(>list_entry), causing general protection fault, because > system_wq can run them concurrently. > > The patch eliminates the race condition. > > Signed-off-by: Dexuan Cui > Tested-by: Adrian Suhov > Tested-by: Chris Valean > Cc: Vitaly Kuznetsov > Cc: Jack Morgenstein > Cc: sta...@vger.kernel.org > Cc: Stephen Hemminger > Cc: K. Y. Srinivasan > --- > drivers/pci/host/pci-hyperv.c | 17 ++--- > 1 file changed, 14 insertions(+), 3 deletions(-) > Reviewed-by: Michael Kelley ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel