linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [RFC] PCI/Hotplug: Clear power fault detected when servicing faults
@ 2016-09-20 22:10 Jon Derrick
  2016-09-21 16:39 ` Jon Derrick
  0 siblings, 1 reply; 2+ messages in thread
From: Jon Derrick @ 2016-09-20 22:10 UTC (permalink / raw)
  To: helgaas; +Cc: Jon Derrick, keith.busch, linux-pci

A slot is allowed to implement power fault detection without having a
power controller present. Current code can set the controller's
power_fault_detected when seeing a power fault interrupt, but cannot
clear the flag without a power controller present. This patch clears the
flag when the power fault event is serviced by the interrupt event
handler.

This is mostly based on my interpretation of the spec for Slot Status's
PFD and Slot Cap's PCP. The relevant pieces are at [1] and [2].

[1]: PCIe r3.1a 7.8.11: Slot Status PFD description:
Note that, depending on hardware capability, it is possible that a power
fault can be detected at any time, independent of the Power Controller
Control setting or the occupancy of the slot.

[2]: PCIe r3.1a 6.7.2.4: Power Fault Detected description seems to imply
that the status bit is independent of the Power Controller Present and
Control bits.

Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
---
 drivers/pci/hotplug/pciehp_ctrl.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/pci/hotplug/pciehp_ctrl.c b/drivers/pci/hotplug/pciehp_ctrl.c
index b413c3c..898a82a 100644
--- a/drivers/pci/hotplug/pciehp_ctrl.c
+++ b/drivers/pci/hotplug/pciehp_ctrl.c
@@ -381,6 +381,7 @@ static void interrupt_event_handler(struct work_struct *work)
 		handle_button_press_event(p_slot);
 		break;
 	case INT_POWER_FAULT:
+		ctrl->power_fault_detected = 0;
 		if (!POWER_CTRL(ctrl))
 			break;
 		pciehp_set_attention_status(p_slot, 1);
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [RFC] PCI/Hotplug: Clear power fault detected when servicing faults
  2016-09-20 22:10 [RFC] PCI/Hotplug: Clear power fault detected when servicing faults Jon Derrick
@ 2016-09-21 16:39 ` Jon Derrick
  0 siblings, 0 replies; 2+ messages in thread
From: Jon Derrick @ 2016-09-21 16:39 UTC (permalink / raw)
  To: helgaas; +Cc: keith.busch, linux-pci

I was far too eager with this one. Found a case this morning which resulted in a soft lockup. Please kill this for now.

On Tue, Sep 20, 2016 at 04:10:37PM -0600, Jon Derrick wrote:
> A slot is allowed to implement power fault detection without having a
> power controller present. Current code can set the controller's
> power_fault_detected when seeing a power fault interrupt, but cannot
> clear the flag without a power controller present. This patch clears the
> flag when the power fault event is serviced by the interrupt event
> handler.

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2016-09-21 16:39 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-09-20 22:10 [RFC] PCI/Hotplug: Clear power fault detected when servicing faults Jon Derrick
2016-09-21 16:39 ` Jon Derrick

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).