On Sat, 14 Sep 2013, Bjorn Helgaas wrote:
> This gets to the reason I'm interested -- we have PCI device removal
> issues, and I have been thinking of them largely as reference counting
> problems. I'd like to make the assertion that if you have a pci_bus
> or a pci_dev pointer, e.g., acquired by
On Fri, Sep 13, 2013 at 8:11 PM, Alan Stern wrote:
> On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
>
>> Thanks. Maybe this is more relevant than I thought. I'd sure like to
>> copy your strategy rather than reinvent something.
>
> Well, I don't know if this will really end up being all that relevant
On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
> Thanks. Maybe this is more relevant than I thought. I'd sure like to
> copy your strategy rather than reinvent something.
Well, I don't know if this will really end up being all that relevant
for PCI, but you're welcome to steal any ideas you like. :
On Fri, Sep 13, 2013 at 2:14 PM, Alan Stern wrote:
> On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
>
>> On Fri, Sep 13, 2013 at 12:25 PM, Alan Stern
>> wrote:
>> > On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
>> >
>> >> usb_bus_list_lock protects the usb_bus_list, and we don't touch that list
>> >> in
On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
> On Fri, Sep 13, 2013 at 12:25 PM, Alan Stern
> wrote:
> > On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
> >
> >> usb_bus_list_lock protects the usb_bus_list, and we don't touch that list
> >> in usb_disconnect(), so there's no reason to hold the lock here.
On Fri, Sep 13, 2013 at 12:25 PM, Alan Stern wrote:
> On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
>
>> usb_bus_list_lock protects the usb_bus_list, and we don't touch that list
>> in usb_disconnect(), so there's no reason to hold the lock here.
>
> The code says:
>
> /* used when updating list of hc
usb_bus_list_lock protects the usb_bus_list, and we don't touch that list
in usb_disconnect(), so there's no reason to hold the lock here.
Signed-off-by: Bjorn Helgaas
---
drivers/usb/core/hcd.c |4
1 file changed, 4 deletions(-)
diff --git a/drivers/usb/core/hcd.c b/drivers/usb/core/h
On Fri, 13 Sep 2013, Bjorn Helgaas wrote:
> usb_bus_list_lock protects the usb_bus_list, and we don't touch that list
> in usb_disconnect(), so there's no reason to hold the lock here.
The code says:
/* used when updating list of hcds */
DEFINE_MUTEX(usb_bus_list_lock);/* exported only f
