The problem was found when I tried to inject PCI config error by
PHB3 PAPR error injection registers into Broadcom Austin 4-ports
NIC adapter. The frozen PE was reported successfully and EEH core
started to recover it. However, I run into fenced PHB when dumping
PCI config space as EEH logs. I was told that PCI config requests
should not be progagated to the adapter until PE reset is done
successfully. Otherise, we would run out of PHB internal credits
and trigger PCT (PCIE Completion Timeout), which leads to the
fenced PHB.
The patch introduces another PE flag EEH_PE_CFG_RESTRICTED, which
is set during PE initialization time if the PE includes the specific
PCI devices that need block PCI config access until PE reset is done.
When the PE becomes frozen for the first time, EEH_PE_CFG_BLOCKED is
set if the PE has flag EEH_PE_CFG_RESTRICTED. Then the PCI config
access to the PE will be dropped by platform PCI accessors until
PE reset is done successfully. The mechanism is shared by PowerNV
platform owned PE or userland owned ones. It's not used on pSeries
platform yet.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Function eeh_pe_state_mark() could possibly have combination of
multiple EEH PE state as its argument. The patch fixes the condition
used to check if EEH_PE_ISOLATED is included.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The problem was reported by Carol: In the scenario of passing mlx4
adapter to guest, EEH error could be recovered successfully. When
returning the device back to host, the driver (mlx4_core.ko)
couldn't be loaded successfully because of error number -5 (-EIO)
returned from mlx4_get_ownership(), which hits offlined PCI device.
The root cause is that we missed to put the affected devices into
normal state on clearing PE isolated state right after PE reset.
The patch fixes above issue by putting the affected devices to
normal state when clearing PE isolated state in eeh_pe_state_clear().
Cc: stable@vger.kernel.org
Reported-by: Carol L. Soto <clsoto@us.ibm.com>
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The patch allows PE (struct eeh_pe) instance to have auxillary data,
whose size is configurable on basis of platform. For PowerNV, the
auxillary data will be used to cache PHB diag-data for that PE
(frozen PE or fenced PHB). In turn, we can retrieve the diag-data
at any later points.
It's useful for the case of VFIO PCI devices where the error log
should be cached, and then be retrieved by the guest at later point.
Also, it can avoid PHB diag-data overwritting if another frozen PE
reported and the previous diag-data isn't fetched by guest.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
pr_warn() is equal to pr_warning(), but the former is a bit more
formal according to commit fc62f2f ("kernel.h: add pr_warn for
symmetry to dev_warn, netdev_warn").
The patch replaces pr_warning() with pr_warn().
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
pci_get_slot() is called with hold of PCI bus semaphore and it's not
safe to be called in interrupt context. However, we possibly checks
EEH error and calls the function in interrupt context. To avoid using
pci_get_slot(), we turn into device tree for fetching location code.
Otherwise, we might run into WARN_ON() as following messages indicate:
WARNING: at drivers/pci/search.c:223
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.16.0-rc3+ #72
task: c000000001367af0 ti: c000000001444000 task.ti: c000000001444000
NIP: c000000000497b70 LR: c000000000037530 CTR: 000000003003d114
REGS: c000000001446fa0 TRAP: 0700 Not tainted (3.16.0-rc3+)
MSR: 9000000000029032 <SF,HV,EE,ME,IR,DR,RI> CR: 48002422 XER: 20000000
CFAR: c00000000003752c SOFTE: 0
:
NIP [c000000000497b70] .pci_get_slot+0x40/0x110
LR [c000000000037530] .eeh_pe_loc_get+0x150/0x190
Call Trace:
.of_get_property+0x30/0x60 (unreliable)
.eeh_pe_loc_get+0x150/0x190
.eeh_dev_check_failure+0x1b4/0x550
.eeh_check_failure+0x90/0xf0
.lpfc_sli_check_eratt+0x504/0x7c0 [lpfc]
.lpfc_poll_eratt+0x64/0x100 [lpfc]
.call_timer_fn+0x64/0x190
.run_timer_softirq+0x2cc/0x3e0
Cc: stable@vger.kernel.org
Signed-off-by: Mike Qiu <qiudayu@linux.vnet.ibm.com>
Acked-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The sysfs entries are lost because of commit 2213fb1 ("powerpc/eeh:
Skip eeh sysfs when eeh is disabled"). That commit added condition
to create sysfs entries with EEH_ENABLED, which isn't populated
when trying to create sysfs entries on PowerNV platform during system
boot time. The patch fixes the issue by:
* Reoder EEH initialization functions so that they're same on
PowerNV/pSeries.
* Cache PE's primary bus by PowerNV platform instead of EEH core
to avoid kernel crash caused by the function reorder. Another
benefit with this is to avoid one eeh_probe_mode_dev() in EEH
core.
Signed-off-by: Mike Qiu <qiudayu@linux.vnet.ibm.com>
Acked-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
As Ben suggested, it's meaningful to dump PE's location code
for site engineers when hitting EEH errors. The patch introduces
function eeh_pe_loc_get() to retireve the location code from
dev-tree so that we can output it when hitting EEH errors.
If primary PE bus is root bus, the PHB's dev-node would be tried
prior to root port's dev-node. Otherwise, the upstream bridge's
dev-node of the primary PE bus will be check for the location code
directly.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The issue was detected in a bit complicated test case where
we have multiple hierarchical PEs shown as following figure:
+-----------------+
| PE#3 p2p#0 |
| p2p#1 |
+-----------------+
|
+-----------------+
| PE#4 pdev#0 |
| pdev#1 |
+-----------------+
PE#4 (have 2 PCI devices) is the child of PE#3, which has 2 p2p
bridges. We accidentally had less-known scenario: PE#4 was removed
permanently from the system because of permanent failure (e.g.
exceeding the max allowd failure times in last hour), then we detects
EEH errors on PE#3 and tried to recover it. However, eeh_dev instances
for pdev#0/1 were not detached from PE#4, which was still connected to
PE#3. All of that was because of the fact that we rely on count-based
pcibios_release_device(), which isn't reliable enough. When doing
recovery for PE#3, we still apply hotplug on PE#4 and pdev#0/1, which
are not valid any more. Eventually, we run into kernel crash.
The patch fixes above issue from two aspects. For unplug, we simply
skip those permanently removed PE, whose state is (EEH_PE_STATE_ISOLATED
&& !EEH_PE_STATE_RECOVERING) and its frozen count should be greater
than EEH_MAX_ALLOWED_FREEZES. For plug, we marked all permanently
removed EEH devices with EEH_DEV_REMOVED and return 0xFF's on read
its PCI config so that PCI core will omit them.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
After reset on the specific PE or PHB, we never configure AER
correctly on PowerNV platform. We needn't care it on pSeries
platform. The patch introduces additional EEH operation eeh_ops::
restore_config() so that we have chance to configure AER correctly
for PowerNV platform.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
None of these files are actually using any __init type directives
and hence don't need to include <linux/init.h>. Most are just a
left over from __devinit and __cpuinit removal, or simply due to
code getting copied from one driver to the next.
The one instance where we add an include for init.h covers off
a case where that file was implicitly getting it from another
header which itself didn't need it.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
While restoring BARs for one specific PCI device, the pci_dev
instance should have been released. So it's not reliable to use
the pci_dev instance on restoring BARs. However, we still need
some information (e.g. PCIe capability position, header type) from
the pci_dev instance. So we have to store those information to
EEH device in advance.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
When EEH error happens to one specific PE, some devices with drivers
supporting EEH won't except hotplug on the device. However, there
might have other deivces without driver, or with driver without EEH
support. For the case, we need do partial hotplug in order to make
sure that the PE becomes absolutely quite during reset. Otherise,
the PE reset might fail and leads to failure of error recovery.
The current code doesn't handle that 'mixed' case properly, it either
uses the error callbacks to the drivers, or tries hotplug, but doesn't
handle a PE (EEH domain) composed of a combination of the two.
The patch intends to support so-called "partial" hotplug for EEH:
Before we do reset, we stop and remove those PCI devices without
EEH sensitive driver. The corresponding EEH devices are not detached
from its PE, but with special flag. After the reset is done, those
EEH devices with the special flag will be scanned one by one.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Currently, we're trasversing the EEH devices list using list_for_each_entry().
That's not safe enough because the EEH devices might be removed from
its parent PE while doing iteration. The patch replaces that with
list_for_each_entry_safe().
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
When we do normal hotplug, the PE (shadow EEH structure) shouldn't be
kept around.
However, we need to keep it if the hotplug an artifial one caused by
EEH errors recovery.
Since we remove EEH device through the PCI hook pcibios_release_device(),
the flag "purge_pe" passed to various functions is meaningless. So the patch
removes the meaningless flag and introduce new flag "EEH_PE_KEEP"
to save the PE while doing hotplug during EEH error recovery.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
After reset (e.g. complete reset) in order to bring the fenced PHB
back, the PCIe link might not be ready yet. The patch intends to
make sure the PCIe link is ready before accessing its subordinate
PCI devices. The patch also fixes that wrong values restored to
PCI_COMMAND register for PCI bridges.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Originally, eeh_mutex was introduced to protect the PE hierarchy
tree and the attached EEH devices because EEH core was possiblly
running with multiple threads to access the PE hierarchy tree.
However, we now have only one kthread in EEH core. So we needn't
the eeh_mutex and just remove it.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We're not expecting that one specific PE got frozen for over 5
times in last hour. Otherwise, the PE will be removed from the
system upon newly coming EEH errors. The patch introduces time
stamp to trace the first error on specific PE in last hour and
function to update that accordingly. Besides, the time stamp
is recovered during PE hotplug path as we did for frozen count.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
There're several types of PEs can be supported for now: PHB, Bus
and Device dependent PE. For PCI bus dependent PE, tracing the
corresponding PCI bus from PE (struct eeh_pe) would make the code
more efficient. The patch also enables the retrieval of PCI bus based
on the PCI bus dependent PE.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
While processing EEH event interrupt from P7IOC, we need function
to retrieve the PE according to the indicated EEH device. The patch
makes function eeh_pe_get() public so that other source files can call
it for that purpose. Also, the patch fixes referring to wrong BDF
(Bus/Device/Function) address while searching PE in function
__eeh_pe_get().
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
One of the possible cases indicated by P7IOC interrupt is fenced
PHB. For that case, we need fetch the PE corresponding to the PHB
and disable the PHB and all subordinate PCI buses/devices, recover
from the fenced state and eventually enable the whole PHB. We need
one function to fetch the PHB PE outside eeh_pe.c and the patch is
going to make eeh_phb_pe_get() public for that purpose.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The patch moves the common part of EEH core into arch/powerpc/kernel
directory so that we needn't PPC_PSERIES while compiling POWERNV
platform:
* Move the EEH common part into arch/powerpc/kernel
* Move the functions for PCI hotplug from pSeries platform to
arch/powerpc/kernel/pci-hotplug.c
* Move CONFIG_EEH from arch/powerpc/platforms/pseries/Kconfig to
arch/powerpc/platforms/Kconfig
* Adjust makefile accordingly
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Gavin Shan