Message ID | 20240514135303.176134-3-krishnak@linux.ibm.com |
---|---|
State | New |
Headers | show |
Series | PCI hotplug driver fixes | expand |
On Tue, May 14, 2024 at 11:54 PM Krishna Kumar <krishnak@linux.ibm.com> wrote: > > There is an issue with the hotplug operation when it's done on the > bridge/switch slot. The bridge-port and devices behind the bridge, which > become offline by hot-unplug operation, don't get hot-plugged/enabled by > doing hot-plug operation on that slot. Only the first port of the bridge > gets enabled and the remaining port/devices remain unplugged. The hot > plug/unplug operation is done by the hotplug driver > (drivers/pci/hotplug/pnv_php.c). > > Root Cause Analysis: This behavior is due to missing code for the DPC > switch/bridge. I don't see anything touching DPC in this series? > *snip* > > Command for reproducing the issue : > > For hot unplug/disable - echo 0 > /sys/bus/pci/slots/C5/power > For hot plug/enable - echo 1 > /sys/bus/pci/slots/C5/power > > where C5 is slot associated with bridge. > > Scenario/Tests: > Output of lspci -nn before test is given below. This snippet contains > devices used for testing on Powernv machine. > > 0004:02:00.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] > 0004:02:01.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] > 0004:02:02.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] > 0004:02:03.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] > 0004:08:00.0 Serial Attached SCSI controller [0107]: > Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 [1000:00c9] (rev 01) > 0004:09:00.0 Serial Attached SCSI controller [0107]: > Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 [1000:00c9] (rev 01) > > Output of lspci -tv before test is as follows: > > # lspci -tv > +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]--+-00.0-[03-07]-- > | | +-01.0-[08]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 > | | +-02.0-[09]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 > | | \-03.0-[0a-0e]-- > | \-00.1 PMC-Sierra Inc. Device 4052 > > C5(bridge) and C6(End Point) slot address are as below: > # cat /sys/bus/pci/slots/C5/address > 0004:02:00 > # cat /sys/bus/pci/slots/C6/address > 0004:09:00 Uh, if I'm reading this right it looks like your "slot" C5 is actually the PCIe switch's internal bus which is definitely not hot pluggable. I find it helps to look at the PCI topology in terms of where the physical PCIe links are. Here we've got: - A link between the PHB (0004:00:00.0) and the switch upstream port (0004:01:00.0) - A link from switch downstream port 0 (0004:02:00.0) to nothing - A link from switch downstream port 1 (0004:02:01.0) to a SAS card - A link from switch downstream port 2 (0004:02:02.0) to a SAS card - A link from switch downstream port 2 (0004:02:03.0) to nothing Note that there's no PCIe link between the switch upstream port (0004:01:00.0) and the downstream ports on bus 0004:02. The connection between those is invisible to us because it's custom bus logic internal to the PCIe switch ASIC. What I think has happened here is that system firmware has supplied bad PCIe slot information to OPAL which has resulted in pnv_php advertising a slot in the wrong place. Assuming this following the usual IBM convention I'd expect the bridge device for C5 to be the PHB's root port and the bus should be 0004:01. It might be worth adding some logic to pnv_php to verify the PCI bridge upstream of the slot actually has the PCIe slot capability to guard against this problem. > Hot-unplug operation on slot associated with bridge: > # echo 0 > /sys/bus/pci/slots/C5/power > # lspci -tv > +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]-- > | \-00.1 PMC-Sierra Inc. Device 4052 Yep, "powering off" C5 doesn't remove the upstream port device. This would create problems if you physically removed the card from C5 since the kernel would assume the switch device is still present. > *snip* > diff --git a/arch/powerpc/kernel/pci_dn.c b/arch/powerpc/kernel/pci_dn.c > index 38561d6a2079..bea612759832 100644 > --- a/arch/powerpc/kernel/pci_dn.c > +++ b/arch/powerpc/kernel/pci_dn.c > @@ -493,4 +493,36 @@ static void pci_dev_pdn_setup(struct pci_dev *pdev) > pdn = pci_get_pdn(pdev); > pdev->dev.archdata.pci_data = pdn; > } > + > +void pci_traverse_sibling_nodes_and_scan_slot(struct device_node *start, struct pci_bus *bus) > +{ > + struct device_node *dn; > + int slotno; > + > + u32 class = 0; > + > + if (!of_property_read_u32(start->child, "class-code", &class)) { > + /* Call of pci_scan_slot for non-bridge/EP case */ > + if (!((class >> 8) == PCI_CLASS_BRIDGE_PCI)) { > + slotno = PCI_SLOT(PCI_DN(start->child)->devfn); > + pci_scan_slot(bus, PCI_DEVFN(slotno, 0)); > + return; > + } > + } > + > + /* Iterate all siblings */ > + for_each_child_of_node(start, dn) { > + class = 0; > + > + if (!of_property_read_u32(start->child, "class-code", &class)) { > + /* Call of pci_scan_slot on each sibling-nodes/bridge-ports */ > + if ((class >> 8) == PCI_CLASS_BRIDGE_PCI) { > + slotno = PCI_SLOT(PCI_DN(dn)->devfn); > + pci_scan_slot(bus, PCI_DEVFN(slotno, 0)); > + } > + } > + } If you're going to iterate over all the DT nodes why not just scan all of them rather than special casing bridges? IIRC current logic is the way it is because PowerVM only puts single devices under a PHB and in the PowerNV (pnv_php) case the PCIe spec guarantees that only device 0 will be present on the end of a link. If you want to handle the more generic case then feel free, but do it properly.
On Fri, May 17, 2024 at 9:15 PM krishna kumar <krishnak@linux.ibm.com> wrote: > > > Uh, if I'm reading this right it looks like your "slot" C5 is actually > > the PCIe switch's internal bus which is definitely not hot pluggable. > > It's a hotplug slot. Please see the snippet below: > > :~$ sudo lspci -vvv -s 0004:02:00.0 | grep --color HotPlug > SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- > :~$ > > :~$ sudo lspci -vvv -s 0004:02:01.0 | grep --color HotPlug > SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- > :~$ > > :~$ sudo lspci -vvv -s 0004:02:02.0 | grep --color HotPlug > SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- > :~$ > > :~$ sudo lspci -vvv -s 0004:02:03.0 | grep --color HotPlug > SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- > :~$ All this is showing is that the switch downstream ports on bus 0004:02 have a slot capability. I already know that (see what I said previously about physical links). The fact the downstream ports have a slot capability also has absolutely nothing to do with anything I was saying. Look at the lspci output for 0004:01:00.0 which is the switch's upstream port. The upstream port device will not have a slot capability because it's a bridge into the virtual PCI bus that is internal to the switch. > It seems like your explanation about the missing 0004:01:00.0 may be > correct and could be due to a firmware bug. However, the scope of this > patch does not relate to this issue. Additionally, if it starts with > 0004:01:00.0 to 0004:01:03.0, the behavior of hot-unplug and hot-plug > operations will remain inconsistent. This patch aims to address the > inconsistent behavior of hot-unplug and hot-plug. > > *snip* > > > It might be worth adding some logic to pnv_php to verify the PCI > > bridge upstream of the slot actually has the PCIe slot capability to > > guard against this problem. > > We can have a look at this problem in another patch. The point of this series is to fix the behaviour of pnv_php, is it not? Powering off a PCI(e) slot is supposed to render it safe to remove the card in that slot. Currently if you "power off" C5, the kernel is still going to have active references to the switch's upstream port device (at 0004:01:00.0) and the switch management function (at 0004:01:00.1). If the kernel has active references to PCI devices physically located in the slot we supposedly powered off, then the hotplug driver isn't doing its job. The asymmetry between hot add and removal that you're trying to fix here is a side effect of the fact that pnv_php is advertising the wrong thing as a slot. I think you should stop pnv_php from advertising something as a slot when it's not actually a slot because that's the root of all your problems. > We wanted to handle the more generic case and did not want to be confined to > only one device assumption. We want to fix the current inconsistent behavior > more generically. Right, as I said above I don't think handing the more generic case is actually required if pnv_php is doing its job properly. It doesn't hurt though. > Regarding the fix, the fix is obvious: really? > We have to traverse > and find the bridge ports from DT and invoke pci_scan_slot() on them. This will > discover and create the entry for bridge ports (0004:02:00.0 to 0004:02:00.3 on > the given bus- 0004:02). There is already an existing function, pci_scan_bridge() > which is doing invocation of pci_scan_slot () for the devices behind the bridge, > in this case for SAS device. So eventually, we are doing a scan of all the entities > behind the slot. I already read your patch so I'm not sure why you feel the need to re-describe it in tedious detail. > Would you like me to combine the non-bridge and bridge cases into one? I can attempt > to do this. Hopefully, if we incorporate the iterate sibling logic case correctly, > we may not need to maintain these two separate cases for bridge and non-bridge. I > will attempt this, and if it works, I will include it in the next patch. Thanks. Yes, do that. Also, do not post HTML emails to linux development lists. It breaks plain text inline quoting which makes your messages annoying to reply to. Some linux development lists will also silently drop HTML emails. Please talk to the other LTC engineers about how to set up your mail client to send plain text emails to avoid these problems in the future. Oliver
Hi Oliver, Thanks for your suggestions. Pls find my response: On 5/20/24 20:29, Oliver O'Halloran wrote: > On Fri, May 17, 2024 at 9:15 PM krishna kumar <krishnak@linux.ibm.com> wrote: >>> Uh, if I'm reading this right it looks like your "slot" C5 is actually >>> the PCIe switch's internal bus which is definitely not hot pluggable. >> It's a hotplug slot. Please see the snippet below: >> >> :~$ sudo lspci -vvv -s 0004:02:00.0 | grep --color HotPlug >> SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- >> :~$ >> >> :~$ sudo lspci -vvv -s 0004:02:01.0 | grep --color HotPlug >> SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- >> :~$ >> >> :~$ sudo lspci -vvv -s 0004:02:02.0 | grep --color HotPlug >> SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- >> :~$ >> >> :~$ sudo lspci -vvv -s 0004:02:03.0 | grep --color HotPlug >> SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- >> :~$ > All this is showing is that the switch downstream ports on bus 0004:02 > have a slot capability. I already know that (see what I said > previously about physical links). The fact the downstream ports have a > slot capability also has absolutely nothing to do with anything I was > saying. Look at the lspci output for 0004:01:00.0 which is the > switch's upstream port. The upstream port device will not have a slot > capability because it's a bridge into the virtual PCI bus that is > internal to the switch. Let me try to understand your suggestion and what needs to be done now: lspci -nn snippet in current scenario: 0004:01:00.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:01:00.1 Memory controller [0580]: PMC-Sierra Inc. Device [11f8:4052] 0004:02:00.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:02:01.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:02:02.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:02:03.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] lspci -tv snippet in current scenario: +-[0001:00]---00.0-[01-0a]--+-00.0-[02-0a]--+-00.0-[03-07]-- | | +-01.0-[08]--+-00.0 Broadcom Inc. and subsidiaries NetXtreme BCM5719 Gigabit Ethernet PCIe | | | +-00.1 Broadcom Inc. and subsidiaries NetXtreme BCM5719 Gigabit Ethernet PCIe | | | +-00.2 Broadcom Inc. and subsidiaries NetXtreme BCM5719 Gigabit Ethernet PCIe | | | \-00.3 Broadcom Inc. and subsidiaries NetXtreme BCM5719 Gigabit Ethernet PCIe | | +-02.0-[09]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | \-03.0-[0a]----00.0 IBM PCI-E IPR SAS Adapter (ASIC) | \-00.1 PMC-Sierra Inc. Device 4052 C5 bus address: [root@ltczzess2-lp1 ~]# cat /sys/bus/pci/slots/C5/address 0004:02:00 [root@ltczzess2-lp1 ~]# 0004:01:00.0 doesn't have hotplug capability but 0004:02:00.0 does have this capability. Below snippet tells about this: [root@ltczzess2-lp1 ~]# sudo lspci -vvv -s 0004:01:00.0 | grep --color HotPlug [root@ltczzess2-lp1 ~]# [root@ltczzess2-lp1 ~]# sudo lspci -vvv -s 0004:02:00.0 | grep --color HotPlug SltCap: AttnBtn- PwrCtrl+ MRL- AttnInd- PwrInd- HotPlug+ Surprise- [root@ltczzess2-lp1 ~]# In Function - pnv_php_register_one() is responsible for slot creation from hotplug capable device node: Below is the current code that does check the device node for hot plug capability and takes the decision /* Check if it's hotpluggable slot */ ret = of_property_read_u32(dn, "ibm,slot-pluggable", &prop32); if (ret || !prop32){ return -ENXIO; } Its obvious that 0004:01:00.0 does not get above criteria fulfilled but 0004:02:00.0 does, so is the current behavior (Upstream port is not became C5 slot but downstream port became C5 slot). I am summarizing your suggested changes. Please let me know if I've got it right: 1. Do you want me to modify the code so that the C5 device-bdf and bus-address become 0004:01:00/0004:01 instead of 0004:02:00/0004:01? 2. When performing a hot-unplug operation on C5, should all devices from 0004:01 be removed? And should all devices from 0004:02 also be removed? I think the answer is yes, but please confirm. 3. When performing a hot-plug operation on C5, should all the devices removed earlier from 0004:01 and 0004:02 be re-attached? 4. Will there be any PCIe topology changes in this workflow? Once you confirm the above requirements, we can discuss how to proceed further. I have some follow up questions from your last mail and I am putting these questions in below paragraphs as inline statements. It will confirm me if we should do above things or not. >> It seems like your explanation about the missing 0004:01:00.0 may be >> correct and could be due to a firmware bug. However, the scope of this >> patch does not relate to this issue. Additionally, if it starts with >> 0004:01:00.0 to 0004:01:03.0, the behavior of hot-unplug and hot-plug >> operations will remain inconsistent. This patch aims to address the >> inconsistent behavior of hot-unplug and hot-plug. >> >> *snip* >> >>> It might be worth adding some logic to pnv_php to verify the PCI >>> bridge upstream of the slot actually has the PCIe slot capability to >>> guard against this problem. >> We can have a look at this problem in another patch. > The point of this series is to fix the behaviour of pnv_php, is it > not? Yes and we will do necessary things. > Powering off a PCI(e) slot is supposed to render it safe to > remove the card in that slot. Do you mean physical-removal of the device after power-off ? > Currently if you "power off" C5, the > kernel is still going to have active references to the switch's > upstream port device (at 0004:01:00.0) and the switch management > function (at 0004:01:00.1). Yes, since we are only operating on the downstream port of C5, upstream ports' reference to the kernel will remain the same. > If the kernel has active references to PCI > devices physically located in the slot we supposedly powered off, then > the hotplug driver isn't doing its job. We have only powered off the downstream ports, not the upstream port. The upstream port will remain powered on. Do you mean to say that it will cause a problem if we physically remove the device while the upstream port is powered on and the downstream port is powered off? Will it cause a kernel crash? Is this the reason for designating the upstream port as a C5 slot and performing a hot-plug operation on it? Is it correct to select a device port that is not hot-pluggable, designate it as a C5 slot, and perform a hot-plug operation on it? > The asymmetry between hot add > and removal that you're trying to fix here is a side effect of the > fact that pnv_php is advertising the wrong thing as a slot. Pnv-php is displaying the information, what it receives from the device node property. We will attempt to modify the code path that is responsible for this. I am not sure yet what additional code is needed for this, but I will figure it out. Is it okay to change this code? > I think > you should stop pnv_php from advertising something as a slot when it's > not actually a slot because that's the root of all your problems. Okay, I am aligned but need some more clarification. Currently, we are observing this behavior with the PMC-Sierra bridge. Will this behavior occur with all bridges? In other words, will the upstream port capability not be hot-pluggable for all bridges and switches, and therefore not be considered for slot selection? In a previous email, you mentioned that this problem is due to a firmware bug, causing the driver to behave incorrectly and advertise the wrong port as a slot. Assuming the firmware bug is not present, what will be the behavior? Will there be any expected PCI-topology changes in the above "lspci -tv" command? Also, if the firmware bug is not present, do we still need to make changes to the driver code? Best Regards, Krishna >> We wanted to handle the more generic case and did not want to be confined to >> only one device assumption. We want to fix the current inconsistent behavior >> more generically. > Right, as I said above I don't think handing the more generic case is > actually required if pnv_php is doing its job properly. It doesn't > hurt though. > >> Regarding the fix, the fix is obvious: > really? > >> We have to traverse >> and find the bridge ports from DT and invoke pci_scan_slot() on them. This will >> discover and create the entry for bridge ports (0004:02:00.0 to 0004:02:00.3 on >> the given bus- 0004:02). There is already an existing function, pci_scan_bridge() >> which is doing invocation of pci_scan_slot () for the devices behind the bridge, >> in this case for SAS device. So eventually, we are doing a scan of all the entities >> behind the slot. > I already read your patch so I'm not sure why you feel the need to > re-describe it in tedious detail. > >> Would you like me to combine the non-bridge and bridge cases into one? I can attempt >> to do this. Hopefully, if we incorporate the iterate sibling logic case correctly, >> we may not need to maintain these two separate cases for bridge and non-bridge. I >> will attempt this, and if it works, I will include it in the next patch. Thanks. > Yes, do that. > > Also, do not post HTML emails to linux development lists. It breaks > plain text inline quoting which makes your messages annoying to reply > to. Some linux development lists will also silently drop HTML emails. > Please talk to the other LTC engineers about how to set up your mail > client to send plain text emails to avoid these problems in the > future. > > Oliver
diff --git a/arch/powerpc/include/asm/ppc-pci.h b/arch/powerpc/include/asm/ppc-pci.h index a8b7e8682f5b..83db8d0798ac 100644 --- a/arch/powerpc/include/asm/ppc-pci.h +++ b/arch/powerpc/include/asm/ppc-pci.h @@ -28,6 +28,10 @@ struct pci_dn; void *pci_traverse_device_nodes(struct device_node *start, void *(*fn)(struct device_node *, void *), void *data); + +void pci_traverse_sibling_nodes_and_scan_slot(struct device_node *start, + struct pci_bus *bus); + extern void pci_devs_phb_init_dynamic(struct pci_controller *phb); #if defined(CONFIG_IOMMU_API) && (defined(CONFIG_PPC_PSERIES) || \ diff --git a/arch/powerpc/kernel/pci-hotplug.c b/arch/powerpc/kernel/pci-hotplug.c index 0fe251c6ac2c..639a3d592fe2 100644 --- a/arch/powerpc/kernel/pci-hotplug.c +++ b/arch/powerpc/kernel/pci-hotplug.c @@ -106,7 +106,7 @@ EXPORT_SYMBOL_GPL(pci_hp_remove_devices); */ void pci_hp_add_devices(struct pci_bus *bus) { - int slotno, mode, max; + int mode, max; struct pci_dev *dev; struct pci_controller *phb; struct device_node *dn = pci_bus_to_OF_node(bus); @@ -129,8 +129,7 @@ void pci_hp_add_devices(struct pci_bus *bus) * order for fully rescan all the way down to pick them up. * They can have been removed during partial hotplug. */ - slotno = PCI_SLOT(PCI_DN(dn->child)->devfn); - pci_scan_slot(bus, PCI_DEVFN(slotno, 0)); + pci_traverse_sibling_nodes_and_scan_slot(dn, bus); max = bus->busn_res.start; /* * Scan bridges that are already configured. We don't touch diff --git a/arch/powerpc/kernel/pci_dn.c b/arch/powerpc/kernel/pci_dn.c index 38561d6a2079..bea612759832 100644 --- a/arch/powerpc/kernel/pci_dn.c +++ b/arch/powerpc/kernel/pci_dn.c @@ -493,4 +493,36 @@ static void pci_dev_pdn_setup(struct pci_dev *pdev) pdn = pci_get_pdn(pdev); pdev->dev.archdata.pci_data = pdn; } + +void pci_traverse_sibling_nodes_and_scan_slot(struct device_node *start, struct pci_bus *bus) +{ + struct device_node *dn; + int slotno; + + u32 class = 0; + + if (!of_property_read_u32(start->child, "class-code", &class)) { + /* Call of pci_scan_slot for non-bridge/EP case */ + if (!((class >> 8) == PCI_CLASS_BRIDGE_PCI)) { + slotno = PCI_SLOT(PCI_DN(start->child)->devfn); + pci_scan_slot(bus, PCI_DEVFN(slotno, 0)); + return; + } + } + + /* Iterate all siblings */ + for_each_child_of_node(start, dn) { + class = 0; + + if (!of_property_read_u32(start->child, "class-code", &class)) { + /* Call of pci_scan_slot on each sibling-nodes/bridge-ports */ + if ((class >> 8) == PCI_CLASS_BRIDGE_PCI) { + slotno = PCI_SLOT(PCI_DN(dn)->devfn); + pci_scan_slot(bus, PCI_DEVFN(slotno, 0)); + } + } + } + +} + DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, pci_dev_pdn_setup);
There is an issue with the hotplug operation when it's done on the bridge/switch slot. The bridge-port and devices behind the bridge, which become offline by hot-unplug operation, don't get hot-plugged/enabled by doing hot-plug operation on that slot. Only the first port of the bridge gets enabled and the remaining port/devices remain unplugged. The hot plug/unplug operation is done by the hotplug driver (drivers/pci/hotplug/pnv_php.c). Root Cause Analysis: This behavior is due to missing code for the DPC switch/bridge. The existing driver depends on pci_hp_add_devices() function for device enablement. This function calls pci_scan_slot() on only one device-node/port of the bridge, not on all the siblings' device-node/port. The missing code needs to be added which will find all the sibling device-nodes/bridge-ports and will run explicit pci_scan_slot() on those. A new function has been added for this purpose which gets invoked from pci_hp_add_devices(). This new function pci_traverse_sibling_nodes_and_scan_slot() gets all the sibling bridge-ports by traversal and explicitly invokes pci_scan_slot on them. Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: Christophe Leroy <christophe.leroy@csgroup.eu> Cc: "Aneesh Kumar K.V" <aneesh.kumar@kernel.org> Cc: Bjorn Helgaas <bhelgaas@google.com> Cc: Gaurav Batra <gbatra@linux.ibm.com> Cc: Nathan Lynch <nathanl@linux.ibm.com> Cc: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Krishna Kumar <krishnak@linux.ibm.com> --- Command for reproducing the issue : For hot unplug/disable - echo 0 > /sys/bus/pci/slots/C5/power For hot plug/enable - echo 1 > /sys/bus/pci/slots/C5/power where C5 is slot associated with bridge. Scenario/Tests: Output of lspci -nn before test is given below. This snippet contains devices used for testing on Powernv machine. 0004:02:00.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:02:01.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:02:02.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:02:03.0 PCI bridge [0604]: PMC-Sierra Inc. Device [11f8:4052] 0004:08:00.0 Serial Attached SCSI controller [0107]: Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 [1000:00c9] (rev 01) 0004:09:00.0 Serial Attached SCSI controller [0107]: Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 [1000:00c9] (rev 01) Output of lspci -tv before test is as follows: # lspci -tv +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]--+-00.0-[03-07]-- | | +-01.0-[08]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | +-02.0-[09]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | \-03.0-[0a-0e]-- | \-00.1 PMC-Sierra Inc. Device 4052 C5(bridge) and C6(End Point) slot address are as below: # cat /sys/bus/pci/slots/C5/address 0004:02:00 # cat /sys/bus/pci/slots/C6/address 0004:09:00 Hot-unplug operation on slot associated with bridge: # echo 0 > /sys/bus/pci/slots/C5/power # lspci -tv +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]-- | \-00.1 PMC-Sierra Inc. Device 4052 From the above lspci -tv output, it can be observed that hot unplug operation has removed all the PMC-Sierra bridge ports like: 00.0-[03-07], 01.0-[08], 02.0-[09], 03.0-[0a-0e] and the SAS devices behind the bridge-port. Without the fix, when the hot plug operation is done on the same slot, it adds only the first bridge port and doesn't restore all the bridge-ports and devices that it unplugged earlier. Below snippet shows this. Hot-plug operation on the bridge slot without the fix: # echo 1 > /sys/bus/pci/slots/C5/power # lspci -tv +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]--+-00.0-[03-07]-- | \-00.1 PMC-Sierra Inc. Device 4052 After the fix, it restores all the devices in the same manner how it unplugged them earlier during the hot unplug operation. The below snippet shows the same. Hot-plug operation on bridge slot with the fix: # echo 1 > /sys/bus/pci/slots/C5/power # lspci -tv +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]--+-00.0-[03-07]-- | | +-01.0-[08]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | +-02.0-[09]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | \-03.0-[0a-0e]-- | \-00.1 PMC-Sierra Inc. Device 4052 Removal of End point device behind bridge are also intact and behaving correctly. Hot-unplug operation on Endpoint device C6: # echo 0 > /sys/bus/pci/slots/C6/power # lspci -tv +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]--+-00.0-[03-07]-- | | +-01.0-[08]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | +-02.0-[09]-- | | \-03.0-[0a-0e]-- | \-00.1 PMC-Sierra Inc. Device 4052 Hot-plug operation on Endpoint device C6: # echo 1 > /sys/bus/pci/slots/C6/power # lspci -tv +-[0004:00]---00.0-[01-0e]--+-00.0-[02-0e]--+-00.0-[03-07]-- | | +-01.0-[08]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | +-02.0-[09]----00.0 Broadcom / LSI SAS3216 PCI-Express Fusion-MPT SAS-3 | | \-03.0-[0a-0e]-- | \-00.1 PMC-Sierra Inc. Device 4052 arch/powerpc/include/asm/ppc-pci.h | 4 ++++ arch/powerpc/kernel/pci-hotplug.c | 5 ++--- arch/powerpc/kernel/pci_dn.c | 32 ++++++++++++++++++++++++++++++ 3 files changed, 38 insertions(+), 3 deletions(-)