* pci_bus_lock question
@ 2004-07-21 22:10 John Rose
2004-07-21 22:21 ` John Rose
0 siblings, 1 reply; 5+ messages in thread
From: John Rose @ 2004-07-21 22:10 UTC (permalink / raw)
To: Greg KH; +Cc: Mike Wortman, lkml
Is the intended purpose of pci_bus_lock to synchronize access to _just_
the global list of pci devices, or also to the pci_root_buses list?
If it is intended to protect the latter, I see many unfortunate places
where it's not being used :)
Thanks-
John
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: pci_bus_lock question
2004-07-21 22:10 pci_bus_lock question John Rose
@ 2004-07-21 22:21 ` John Rose
2004-07-22 7:08 ` Greg KH
0 siblings, 1 reply; 5+ messages in thread
From: John Rose @ 2004-07-21 22:21 UTC (permalink / raw)
To: Greg KH; +Cc: Mike Wortman, lkml
But then, most of these violations are in __init functions. I think I
just answered my own question :)
Thanks-
John
On Wed, 2004-07-21 at 17:10, John Rose wrote:
> Is the intended purpose of pci_bus_lock to synchronize access to _just_
> the global list of pci devices, or also to the pci_root_buses list?
>
> If it is intended to protect the latter, I see many unfortunate places
> where it's not being used :)
>
> Thanks-
> John
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: pci_bus_lock question
2004-07-21 22:21 ` John Rose
@ 2004-07-22 7:08 ` Greg KH
2004-07-22 15:11 ` John Rose
0 siblings, 1 reply; 5+ messages in thread
From: Greg KH @ 2004-07-22 7:08 UTC (permalink / raw)
To: John Rose; +Cc: Greg KH, Mike Wortman, lkml
On Wed, Jul 21, 2004 at 05:21:08PM -0500, John Rose wrote:
> But then, most of these violations are in __init functions. I think I
> just answered my own question :)
Yes, we don't protect the lists in those __init functions, as it isn't
needed at that point in time.
thanks,
greg k-h
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: pci_bus_lock question
2004-07-22 7:08 ` Greg KH
@ 2004-07-22 15:11 ` John Rose
2004-07-24 2:05 ` Greg KH
0 siblings, 1 reply; 5+ messages in thread
From: John Rose @ 2004-07-22 15:11 UTC (permalink / raw)
To: Greg KH; +Cc: Mike Wortman, lkml
I need to remove a bus from the pci_root_buses() list, and I need to do
so from a module. Would it be preferable to export the pci_bus_lock
symbol or create wrappers in the PCI core that safely add/remove buses
to/from this list?
I'm guessing the latter :)
Thanks-
John
On Thu, 2004-07-22 at 02:08, Greg KH wrote:
> On Wed, Jul 21, 2004 at 05:21:08PM -0500, John Rose wrote:
> > But then, most of these violations are in __init functions. I think I
> > just answered my own question :)
>
> Yes, we don't protect the lists in those __init functions, as it isn't
> needed at that point in time.
>
> thanks,
>
> greg k-h
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: pci_bus_lock question
2004-07-22 15:11 ` John Rose
@ 2004-07-24 2:05 ` Greg KH
0 siblings, 0 replies; 5+ messages in thread
From: Greg KH @ 2004-07-24 2:05 UTC (permalink / raw)
To: John Rose; +Cc: Mike Wortman, lkml
On Thu, Jul 22, 2004 at 10:11:21AM -0500, John Rose wrote:
> I need to remove a bus from the pci_root_buses() list, and I need to do
> so from a module. Would it be preferable to export the pci_bus_lock
> symbol or create wrappers in the PCI core that safely add/remove buses
> to/from this list?
>
> I'm guessing the latter :)
Good guess :)
greg k-h
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2004-07-24 3:36 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-07-21 22:10 pci_bus_lock question John Rose
2004-07-21 22:21 ` John Rose
2004-07-22 7:08 ` Greg KH
2004-07-22 15:11 ` John Rose
2004-07-24 2:05 ` Greg KH
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox