Hi Bjorn,
On the system that this bug happens, an MCA event is generated while
kernel crashed:
Transaction Address: memory write to address 0x00000ae041428 (LMMIO
- SBL Blade 1 SFW DDR Memory)
I guess the there is some module trying to visit the address
0x00000ae041428 right after this line is run:
pci_write_config_word(dev, PCI_COMMAND,
orig_cmd & ~(PCI_COMMAND_MEMORY | PCI_COMMAND_IO));
The output of lspci -vvv is followed.
40:01.0 PCI bridge: Intel Corporation 5520/5500/X58 I/O Hub PCI Express
Root Port 1 (rev 22) (prog-if 00 [Normal decode])
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr+ Stepping- SERR+ FastB2B- DisINTx+
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort-
<TAbort- <MAbort- >SERR- <PERR- INTx-
Latency: 0, Cache Line Size: 64 bytes
Bus: primary=40, secondary=41, subordinate=41, sec-latency=0
I/O behind bridge: 0000f000-00000fff
Memory behind bridge: ae000000-af8fffff
Prefetchable memory behind bridge:
fffffffffff00000-00000000000fffff
Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort-
<TAbort- <MAbort- <SERR- <PERR-
BridgeCtl: Parity+ SERR+ NoISA- VGA- MAbort- >Reset- FastB2B-
PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
Capabilities: [40] Subsystem: Intel Corporation 5520/5500/X58
I/O Hub PCI Express Root Port 1
Capabilities: [60] Message Signalled Interrupts: Mask+ 64bit-
Count=1/2 Enable+
Address: fee00000 Data: 4046
Masking: 00000002 Pending: 00000000
Capabilities: [90] Express (v2) Root Port (Slot-), MSI 00
DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s
<64ns, L1 <1us
ExtTag+ RBE+ FLReset-
DevCtl: Report errors: Correctable+ Non-Fatal+ Fatal+
Unsupported+
RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop-
MaxPayload 128 bytes, MaxReadReq 128 bytes
DevSta: CorrErr- UncorrErr- FatalErr- UnsuppReq-
AuxPwr- TransPend-
LnkCap: Port #0, Speed 5GT/s, Width x2, ASPM L0s L1,
Latency L0 <512ns, L1 <64us
ClockPM- Suprise+ LLActRep+ BwNot+
LnkCtl: ASPM Disabled; RCB 64 bytes Disabled- Retrain-
CommClk-
ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
LnkSta: Speed 2.5GT/s, Width x1, TrErr- Train- SlotClk+
DLActive+ BWMgmt- ABWMgmt-
RootCtl: ErrCorrectable- ErrNon-Fatal- ErrFatal-
PMEIntEna+ CRSVisible-
RootCap: CRSVisible-
RootSta: PME ReqID 0000, PMEStatus- PMEPending-
DevCap2: Completion Timeout: Range BCD, TimeoutDis+ ARIFwd+
DevCtl2: Completion Timeout: 260ms to 900ms,
TimeoutDis- ARIFwd-
LnkCtl2: Target Link Speed: 5GT/s, EnterCompliance-
SpeedDis-, Selectable De-emphasis: -3.5dB
Transmit Margin: Normal Operating Range,
EnterModifiedCompliance- ComplianceSOS-
Compliance De-emphasis: -6dB
LnkSta2: Current De-emphasis Level: -3.5dB
Capabilities: [e0] Power Management version 3
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA
PME(D0+,D1-,D2-,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Capabilities: [100] Advanced Error Reporting
UESta: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt-
UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
UEMsk: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt-
UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
UESvrt: DLP+ SDES+ TLP+ FCP+ CmpltTO- CmpltAbrt-
UnxCmplt- RxOF+ MalfTLP+ ECRC- UnsupReq+ ACSViol-
CESta: RxErr- BadTLP- BadDLLP- Rollover- Timeout-
NonFatalErr-
CEMsk: RxErr- BadTLP- BadDLLP- Rollover- Timeout-
NonFatalErr+
AERCap: First Error Pointer: 00, GenCap- CGenEn-
ChkCap- ChkEn-
Capabilities: [150] Access Control Services
ACSCap: SrcValid+ TransBlk+ ReqRedir+ CmpltRedir+
UpstreamFwd+ EgressCtrl- DirectTrans-
ACSCtl: SrcValid- TransBlk- ReqRedir- CmpltRedir-
UpstreamFwd- EgressCtrl- DirectTrans-
Capabilities: [160] Vendor Specific Information <?>
Kernel driver in use: pcieport
Kernel modules: shpchp
Thanks
ZhenHua
On 07/10/2013 12:49 AM, Bjorn Helgaas wrote:
On Mon, Jul 8, 2013 at 11:42 PM, Li, Zhen-Hua <zhen-hual@xxxxxx> wrote:
On some IA64 platforms with intel PCI bridge, for example, HP BL890c i2
with Intel Corporation 5520/5500/X58 I/O Hub PCI Express Root Port,
when kernel tries to disable the mmio decoding on the PCI bridge devices,
kernel may crash.
And in the comment of function quirk_mmio_always_on, it also says:
"But doing so (disable the mmio decoding) may cause problems on host bridge
and perhaps other key system devices"
So, for this PCI bridge, dev->mmio_always_on bit should be set to 1.
To avoid affecting the use of quirk_mmio_always_on, a new function is created.
Signed-off-by: Li, Zhen-Hua <zhen-hual@xxxxxx>
---
drivers/pci/quirks.c | 17 +++++++++++++++++
include/linux/pci_ids.h | 1 +
2 files changed, 18 insertions(+)
diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
index e85d230..665af3e 100644
--- a/drivers/pci/quirks.c
+++ b/drivers/pci/quirks.c
@@ -44,6 +44,23 @@ static void quirk_mmio_always_on(struct pci_dev *dev)
DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_ANY_ID, PCI_ANY_ID,
PCI_CLASS_BRIDGE_HOST, 8, quirk_mmio_always_on);
+#ifdef CONFIG_IA64
+/*
+ * On some IA64 platforms, for some intel PCI bridge devices, for example,
+ * the Intel Corporation 5520/5500/X58 I/O Hub PCI Express Root Port,
+ * disable the mmio decoding on this device may cause system crash.
+ * So dev->mmio_always_on bit should be set to 1.
+ */
+static void quirk_mmio_on_intel_pcibridge(struct pci_dev *dev)
+{
+ dev->mmio_always_on = 1;
+}
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL,
+ PCI_DEVICE_ID_INTEL_5520_5550_X58,
+ PCI_CLASS_BRIDGE_PCI,
+ 8, quirk_mmio_on_intel_pcibridge);
+#endif
+
/* The Mellanox Tavor device gives false positive parity errors
* Mark this device with a broken_parity_status, to allow
* PCI scanning code to "skip" this now blacklisted device.
diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h
index 3bed2e8..d8c60b7 100644
--- a/include/linux/pci_ids.h
+++ b/include/linux/pci_ids.h
@@ -2742,6 +2742,7 @@
#define PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH2_RANK_REV2 0x2db2
#define PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH2_TC_REV2 0x2db3
#define PCI_DEVICE_ID_INTEL_82855PM_HB 0x3340
+#define PCI_DEVICE_ID_INTEL_5520_5550_X58 0x3408
#define PCI_DEVICE_ID_INTEL_IOAT_TBG4 0x3429
#define PCI_DEVICE_ID_INTEL_IOAT_TBG5 0x342a
#define PCI_DEVICE_ID_INTEL_IOAT_TBG6 0x342b
--
1.7.10.4
You need to figure out what the problem is, not just avoid it. It's
very unlikely that the problem is something unique to ia64. In fact,
I think it's very doubtful that the problem is even something unique
to the 5520 root ports. My guess is there's something special about
the system you're testing.
Evidently you have traffic going to a device behind the root port at
the same time as we're trying to read the root port's BARs. Linux
should not generate traffic like that while we're enumerating the root
port. Does the problem happen on a root port with an iLO behind it?
Can you collect "lspci -vvv" output and identify the root port where
the problem occurs?
Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
