pci: question about the life cycle of pci_dev and its parent pci_bus

pci: question about the life cycle of pci_dev and its parent pci_bus

[Liu ping fan]

Hi,

I have a question about the life cycle of pci_dev and its parent pci_bus.
In pci_destroy_dev(), we remove the pci_dev from the bus_list, so
pci_dev is isolated from its parent pci_bus, and their life cycle are
independent too. But it seems that in pci_dev, we can still access the
pci_bus from the pointer pci_dev->bus.  To resolve such issue, I think
during the pci_dev removal,  there should be process to 1.prevent the
reader incoming 2. ensure the current reader has exit.  But can not
find them.

Could anyone tell me?

Thanx, pingfan

Re: pci: question about the life cycle of pci_dev and its parent pci_bus

[Jiang Liu]

Hi Pingfan,

	We have realized the same when implementing lock mechanisms for PCI hotplug
operations, and we are planning to holding a reference to parent bus for each pci dev.
	Regards!
	Gerry

On 08/23/2012 01:24 PM, Liu ping fan wrote:
> Hi,
> 
> I have a question about the life cycle of pci_dev and its parent pci_bus.
> In pci_destroy_dev(), we remove the pci_dev from the bus_list, so
> pci_dev is isolated from its parent pci_bus, and their life cycle are
> independent too. But it seems that in pci_dev, we can still access the
> pci_bus from the pointer pci_dev->bus.  To resolve such issue, I think
> during the pci_dev removal,  there should be process to 1.prevent the
> reader incoming 2. ensure the current reader has exit.  But can not
> find them.
> 
> Could anyone tell me?
> 
> Thanx, pingfan
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: pci: question about the life cycle of pci_dev and its parent pci_bus

[Liu ping fan]

On Fri, Aug 24, 2012 at 10:30 PM, Jiang Liu <liuj97@gmail.com> wrote:
> Hi Pingfan,
>
>         We have realized the same when implementing lock mechanisms for PCI hotplug
> operations, and we are planning to holding a reference to parent bus for each pci dev.

Thanks! You save me from searching such codes.
Regards,
pingfan

>         Regards!
>         Gerry
>
> On 08/23/2012 01:24 PM, Liu ping fan wrote:
>> Hi,
>>
>> I have a question about the life cycle of pci_dev and its parent pci_bus.
>> In pci_destroy_dev(), we remove the pci_dev from the bus_list, so
>> pci_dev is isolated from its parent pci_bus, and their life cycle are
>> independent too. But it seems that in pci_dev, we can still access the
>> pci_bus from the pointer pci_dev->bus.  To resolve such issue, I think
>> during the pci_dev removal,  there should be process to 1.prevent the
>> reader incoming 2. ensure the current reader has exit.  But can not
>> find them.
>>
>> Could anyone tell me?
>>
>> Thanx, pingfan
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
>>
>