Re: [PATCH v2 3/9] PCI: preserve dev->subordinate until pci_stop_dev() has been called

[Jiang Liu <liuj97@gmail.com>]

On 09/15/2012 01:09 PM, Yinghai Lu wrote:
> On Fri, Sep 14, 2012 at 8:05 PM, Jiang Liu <liuj97@gmail.com> wrote:
>> From: Yijing Wang <wangyijing@huawei.com>
>>
>> Changeset 2ed168eeb3edec029aa0eca5cb981d6376f931f9 "PCI: Fold stop and
>> remove helpers into their callers" has changed the behavior when
>> removing a PCI device.
>>
>> Previously, for a PCI bridge device with secondary bus, dev->subordinate
>> is valid when calling PCI bus notification callbacks for
>> BUS_NOTIFY_ADD_DEVICE/BUS_NOTIFY_DEL_DEVICE events. Now dev->subordinate
>> has been reset to NULL when calling callbacks for BUS_NOTIFY_DEL_DEVICE
>> events, which may break some PCI bus notification callbacks.
>>
>> So revert to the original behavior to keep dev->subordinate valid for
>> both BUS_NOTIFY_ADD_DEVICE events and BUS_NOTIFY_DEL_DEVICE events.
>>
>> Signed-off-by: Yijing Wang <wangyijing@huawei.com>
>> Signed-off-by: Jiang Liu <jiang.liu@huawei.com>
>> ---
>>  drivers/pci/remove.c |    6 +++---
>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/pci/remove.c b/drivers/pci/remove.c
>> index 86a4636..6244956 100644
>> --- a/drivers/pci/remove.c
>> +++ b/drivers/pci/remove.c
>> @@ -79,16 +79,16 @@ void pci_stop_and_remove_bus_device(struct pci_dev *dev)
>>          * iterator.  Therefore, iterate in reverse so we remove the VFs
>>          * first, then the PF.
>>          */
>> -       if (bus) {
>> +       if (bus)
>>                 list_for_each_entry_safe_reverse(child, tmp,
>>                                                  &bus->devices, bus_list)
>>                         pci_stop_and_remove_bus_device(child);
>> +       pci_stop_dev(dev);
>>
>> +       if (bus) {
>>                 pci_remove_bus(bus);
>>                 dev->subordinate = NULL;
>>         }
>> -
>> -       pci_stop_dev(dev);
>>         pci_destroy_dev(dev);
>>  }
>>  EXPORT_SYMBOL(pci_stop_and_remove_bus_device);
> 
> for support root bus support with acpi_pci_root (ioapic/iommu) will
> need pci_stop_bus_devices()
> 
> so I just put those function back...
> 
> http://git.kernel.org/?p=linux/kernel/git/yinghai/linux-yinghai.git;a=commitdiff;h=247a40206cf44488f21bc6074cf0ba2805d4d840
Hi Yinghai,
	I will just drop this patch when merging with your pci root bus hotplug work.
	--Gerry