Re: [Update][PATCH] PCIe / hotplug: Drop pointless ACPI-based "slot detection" check

[Jarod Wilson <jarod@redhat.com>]

On 6/11/2015 5:16 PM, Rafael J. Wysocki wrote:
> On Thursday, June 11, 2015 04:38:12 PM Jarod Wilson wrote:
>> On 6/11/2015 1:05 PM, Jarod Wilson wrote:
>>> On 5/21/2015 9:21 PM, Rafael J. Wysocki wrote:
>>>> On Thursday, May 21, 2015 11:11:46 AM Bjorn Helgaas wrote:
>>>>> On Tue, May 19, 2015 at 03:27:58PM +0200, Rafael J. Wysocki wrote:
>>>>>> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>>>>>>
>>>>>> Jarod Wilson reports that the expresscard hotplug setup doesn't work
>>>>>> on HP ZBook G2.  The problem turns out to be the ACPI-based "slot
>>>>>> detection" code called from pciehp_probe() which tries to use some
>>>>>> questionable heuristics based on what ACPI objects are present for
>>>>>> the PCIe port device at hand to figure out whether or not to register
>>>>>> a hotplug slot for that port.
>>>>>>
>>>>>> That code is used if there is at least one PCIe port having an ACPI
>>>>>> device configuration object related to hotplug (such as _EJ0 or _RMV)
>>>>>> and the Thunderbolt port on the affected machine has _RMV.  Of course,
>>>>>> Thunderbolt and PCIe native hotplug need not be mutually exclusive
>>>>>> (as they aren't on the machine in question), so that rule is simply
>>>>>> incorrect.
>>>>>>
>>>>>> Moreover, the ACPI-based "slot detection" check does not add any
>>>>>> value if pciehp_probe() is called at all and the service type of the
>>>>>> device object it has been called for is PCIE_PORT_SERVICE_HP, because
>>>>>> PCIe hotplug services are only registered if the _OSC handshake in
>>>>>> acpi_pci_root_add() allows the kernel to control the PCIe native
>>>>>> hotplug feature.  No more checks need to be carried out to decide
>>>>>> whether or not to register a native PCIe hotlug slot in that case.
>>>>>>
>>>>>> For the above reasons, make pciehp_probe() check if it has been
>>>>>> called for the right service type and drop the pointless ACPI-based
>>>>>> "slot detection" check from it.  Also remove the entire code whose
>>>>>> only user is that check (the entire pciehp_acpi.c file goes away
>>>>>> as a result) and drop function headers related to it from the
>>>>>> internal PCIeHP header file.
>>>>>>
>>>>>> Link: http://marc.info/?t=143163219300002&r=1&w=2
>>>>>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=98581
>>>>>> Reported-by: Jarod Wilson <jarod@redhat.com>
>>>>>> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>>>>>
>>>>> This is awesome!  Applied to pci/hotplug for v4.2, with Jarod's
>>>>> reviewed/tested-by.
>>>>
>>>> Thanks!
>>>
>>> Looks like I didn't test enough. I can't explain WHY, but with this
>>> applied, now thunderbolt hot unplug of a network adapter goes haywire,
>>> where prior to the patch, it worked just fine. Still looking into it...
>>
>> Filed bug, dmesg spew can be found in the bug.
>>
>> https://bugzilla.kernel.org/show_bug.cgi?id=99841
>
> If it worked for you previously, can you possibly try to re-create that
> configuration and set of patches applied and retest then?

I tried the current Fedora 4.1-rc7 build first, everything was fine, 
then patched in JUST the one patch, and it went belly up. I'll add some 
extra debugging spew to a build both with and without the one patch, to 
see what differences there are in devices pciehp is claiming and follow 
up in the bug on your other info requests.

-- 
Jarod Wilson
jarod@redhat.com