Skip to contentSkip to conversation

PCI,pciehp: Don't set slot off when found device already exists.

PATCH Applied Yijing Wang 2 participants 4 messages Jan 13, 2014 – Feb 12, 2014 PCI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 PCI,pciehp: Don't set slot off when found device already exists.

If we found device already exists during hot add device, we should leave it, not to set slot off.

---
 drivers/pci/hotplug/pciehp_ctrl.c |    3 ++-
 drivers/pci/hotplug/pciehp_pci.c  |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
drivers/pci/hotplug/pciehp_ctrl.c +2 -1
index 38f0186..f27b0ff 100644
@@ -224,7 +224,8 @@ static int board_added(struct slot *p_slot)
 	if (retval) {
 		ctrl_err(ctrl, "Cannot add device at %04x:%02x:00\n",
 			 pci_domain_nr(parent), parent->number);
-		goto err_exit;
+		if (retval != -EEXIST)
+			goto err_exit;
 	}
 
 	if (PWR_LED(ctrl))
drivers/pci/hotplug/pciehp_pci.c +1 -1
index aac7a40..9b87b49 100644
@@ -48,7 +48,7 @@ int pciehp_configure_device(struct slot *p_slot)
 			 "at %04x:%02x:00, cannot hot-add\n", pci_name(dev),
 			 pci_domain_nr(parent), parent->number);
 		pci_dev_put(dev);
-		return -EINVAL;
+		return -EEXIST;
 	}
 
 	num = pci_scan_slot(parent, PCI_DEVFN(0, 0));
Signature · 1 line
1.7.1
maintainer
#1 lore raw
Quoted 2 lines · Yijing Wang

If we found device already exists during hot add device, we should leave it, not to set slot off.

Does this fix a defect? What's the scenario where this fixes a problem in the current code? I think this change makes sense, but I'd like to have a better description of the problem that we're fixing.

I see the following path that looks like it might be a problem with the current code if slot power is off when we resume after a suspend:

pciehp_resume
pciehp_enable_slot
board_added
pciehp_configure_device

Is this the scenario you're addressing?

Quoted 41 lines
---
 drivers/pci/hotplug/pciehp_ctrl.c |    3 ++-
 drivers/pci/hotplug/pciehp_pci.c  |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
drivers/pci/hotplug/pciehp_ctrl.c +2 -1
index 38f0186..f27b0ff 100644
@@ -224,7 +224,8 @@ static int board_added(struct slot *p_slot)
 	if (retval) {
 		ctrl_err(ctrl, "Cannot add device at %04x:%02x:00\n",
 			 pci_domain_nr(parent), parent->number);
-		goto err_exit;
+		if (retval != -EEXIST)
+			goto err_exit;
 	}
 
 	if (PWR_LED(ctrl))
drivers/pci/hotplug/pciehp_pci.c +1 -1
index aac7a40..9b87b49 100644
@@ -48,7 +48,7 @@ int pciehp_configure_device(struct slot *p_slot)
 			 "at %04x:%02x:00, cannot hot-add\n", pci_name(dev),
 			 pci_domain_nr(parent), parent->number);
 		pci_dev_put(dev);
-		return -EINVAL;
+		return -EEXIST;
 	}
 
 	num = pci_scan_slot(parent, PCI_DEVFN(0, 0));
Signature · 7 lines
1.7.1


--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw
Quoted 11 lines · Bjorn Helgaas
Quoted 2 lines · Yijing Wang

If we found device already exists during hot add device, we should leave it, not to set slot off.

Does this fix a defect? What's the scenario where this fixes a problem in the current code? I think this change makes sense, but I'd like to have a better description of the problem that we're fixing.

I see the following path that looks like it might be a problem with the current code if slot power is off when we resume after a suspend:

Quoted 6 lines

pciehp_resume
pciehp_enable_slot
board_added
pciehp_configure_device

Is this the scenario you're addressing?

Yes, it is.

I found this case by reviewing the code. But I have no platform supports pciehp resume. So I can not test it in real.

Thanks!
Yijing.

Quoted 44 lines
Quoted 42 lines
---
 drivers/pci/hotplug/pciehp_ctrl.c |    3 ++-
 drivers/pci/hotplug/pciehp_pci.c  |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
drivers/pci/hotplug/pciehp_ctrl.c +2 -1
index 38f0186..f27b0ff 100644
@@ -224,7 +224,8 @@ static int board_added(struct slot *p_slot)
 	if (retval) {
 		ctrl_err(ctrl, "Cannot add device at %04x:%02x:00\n",
 			 pci_domain_nr(parent), parent->number);
-		goto err_exit;
+		if (retval != -EEXIST)
+			goto err_exit;
 	}
 
 	if (PWR_LED(ctrl))
drivers/pci/hotplug/pciehp_pci.c +1 -1
index aac7a40..9b87b49 100644
@@ -48,7 +48,7 @@ int pciehp_configure_device(struct slot *p_slot)
 			 "at %04x:%02x:00, cannot hot-add\n", pci_name(dev),
 			 pci_domain_nr(parent), parent->number);
 		pci_dev_put(dev);
-		return -EINVAL;
+		return -EEXIST;
 	}
 
 	num = pci_scan_slot(parent, PCI_DEVFN(0, 0));
Signature · 7 lines
1.7.1


--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Signature · 2 lines
Thanks!
Yijing
maintainer Applied
#3 lore raw
In reply to Yijing Wang
Quoted 4 lines · Yijing Wang

If we found device already exists during hot add device, we should leave it, not to set slot off.

Applied to pci/pciehp for v3.15, thanks!

Bjorn

Quoted 40 lines
---
 drivers/pci/hotplug/pciehp_ctrl.c |    3 ++-
 drivers/pci/hotplug/pciehp_pci.c  |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
drivers/pci/hotplug/pciehp_ctrl.c +2 -1
index 38f0186..f27b0ff 100644
@@ -224,7 +224,8 @@ static int board_added(struct slot *p_slot)
 	if (retval) {
 		ctrl_err(ctrl, "Cannot add device at %04x:%02x:00\n",
 			 pci_domain_nr(parent), parent->number);
-		goto err_exit;
+		if (retval != -EEXIST)
+			goto err_exit;
 	}
 
 	if (PWR_LED(ctrl))
drivers/pci/hotplug/pciehp_pci.c +1 -1
index aac7a40..9b87b49 100644
@@ -48,7 +48,7 @@ int pciehp_configure_device(struct slot *p_slot)
 			 "at %04x:%02x:00, cannot hot-add\n", pci_name(dev),
 			 pci_domain_nr(parent), parent->number);
 		pci_dev_put(dev);
-		return -EINVAL;
+		return -EEXIST;
 	}
 
 	num = pci_scan_slot(parent, PCI_DEVFN(0, 0));
Signature · 7 lines
1.7.1


--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html