***SPAM*** RE: [Xen-devel] Re: [ofa-general] Fwd: pciback module not working

Jiang, Yunhong yunhong.jiang at intel.com
Thu Feb 12 00:01:27 PST 2009


Can you please share more information how will the ib_mthca do QUERY_FW? Through config space access? Through MMIO access? I think more information will be helpful. The only thing seems strange to me is, from "Memory at fdc00000 (64-bit, non-prefetchable) [disabled] [size=1M]" , seems the MMIO is disabled?

Thanks
Yunhong Jiang

________________________________
From: subbu kl [mailto:subbukl at gmail.com]
Sent: 2009年2月12日 15:46
To: Jiang, Yunhong
Cc: David Brown; xen-devel at lists.xensource.com; general at lists.openfabrics.org
Subject: Re: [Xen-devel] Re: [ofa-general] Fwd: pciback module not working

so back to square one ?
Why QUERY_FW should fail in domU ?

~subbu

On Thu, Feb 12, 2009 at 12:30 PM, Jiang, Yunhong <yunhong.jiang at intel.com<mailto:yunhong.jiang at intel.com>> wrote:
DomU access config space through pcibackend, so that message is ok.

________________________________
From: subbu kl [mailto:subbukl at gmail.com<mailto:subbukl at gmail.com>]
Sent: 2009年2月12日 14:59

To: Jiang, Yunhong
Cc: David Brown; xen-devel at lists.xensource.com<mailto:xen-devel at lists.xensource.com>; general at lists.openfabrics.org<mailto:general at lists.openfabrics.org>
Subject: Re: [Xen-devel] Re: [ofa-general] Fwd: pciback module not working

So getting PCI config space access in domU will solve the problem ? if so how can I achieve that ?

~subbu

On Thu, Feb 12, 2009 at 12:26 PM, Jiang, Yunhong <yunhong.jiang at intel.com<mailto:yunhong.jiang at intel.com>> wrote:
Sorry that seems the original mail has tried the permissive already :$
How will So how will the card do the QEUREY_FW command?Through config space or through MMIO? Following information is something strange, why all the MMIO range is disabled?

      Memory at fdc00000 (64-bit, non-prefetchable) [disabled] [size=1M]
      Memory at fd000000 (64-bit, prefetchable) [disabled] [size=8M]

As for the following information, I think it should be harmless since domU has no method of config spacess access method.
 PCI: Fatal: No PCI config space access function found

Thanks
Yunhong Jiang

________________________________
From: subbu kl [mailto:subbukl at gmail.com<mailto:subbukl at gmail.com>]
Sent: 2009年2月12日 14:43

To: Jiang, Yunhong
Cc: David Brown; xen-devel at lists.xensource.com<mailto:xen-devel at lists.xensource.com>; general at lists.openfabrics.org<mailto:general at lists.openfabrics.org>
Subject: Re: [Xen-devel] Re: [ofa-general] Fwd: pciback module not working

oops missed it,

well now I dont see that enable permissive...message. here goes the messages what I got in dom0 while booting domU

tap tap-1-51712: 2 getting info
pciback: vpci: 0000:0e:00.0: assign to virtual slot 0
device vif1.0 entered promiscuous mode
ADDRCONF(NETDEV_UP): vif1.0: link is not ready
blktap: ring-ref 9, event-channel 9, protocol 1 (x86_64-abi)
PCI: Enabling device 0000:0e:00.0 (0000 -> 0002)
ACPI: PCI Interrupt 0000:0e:00.0[A] -> GSI 16 (level, low) -> IRQ 16
PCI: Setting latency timer of device 0000:0e:00.0 to 64
ACPI: PCI interrupt for device 0000:0e:00.0 disabled
ADDRCONF(NETDEV_CHANGE): vif1.0: link becomes ready
xenbr0: topology change detected, propagating
xenbr0: port 3(vif1.0) entering forwarding state

any suspicious message ?
any Idea why I get that :
 PCI: Fatal: No PCI config space access function found
rtc: IRQ 8 is not free.

message in domU bootup message ?

~subbu

On Thu, Feb 12, 2009 at 11:50 AM, Jiang, Yunhong <yunhong.jiang at intel.com<mailto:yunhong.jiang at intel.com>> wrote:
So any changes in dom0's dmesg?


________________________________
From: subbu kl [mailto:subbukl at gmail.com<mailto:subbukl at gmail.com>]
Sent: 2009年2月12日 13:52
To: Jiang, Yunhong
Cc: David Brown; xen-devel at lists.xensource.com<mailto:xen-devel at lists.xensource.com>; general at lists.openfabrics.org<mailto:general at lists.openfabrics.org>
Subject: Re: [Xen-devel] Re: [ofa-general] Fwd: pciback module not working

no luck !
 dmesg in XEN PV guest shows :

ib_mthca: Mellanox InfiniBand HCA driver v1.0 (April 4, 2008)
ib_mthca: Initializing 0000:00:00.0
PCI: Enabling device 0000:00:00.0 (0000 -> 0002)
PCI: Setting latency timer of device 0000:00:00.0 to 64
ib_mthca 0000:00:00.0: QUERY_FW command failed, aborting.
ib_mthca: probe of 0000:00:00.0 failed with error -11

even after executingh the following in dom0:

#echo -n 0000:0e:00.0 > /sys/bus/pci/drivers/pciback/permissive

I am getting the follwing messages on the console as part of the initial bootup messages of the guest:

Started domain rhel52_64_3
PCI: Fatal: No PCI config space access function found
rtc: IRQ 8 is not free.
i8042.c: No controller found.

after executing the following in dom0 :
#xm create -c rhel52_64_3


so, problem persisits,

~subbu


2009/2/12 Jiang, Yunhong <yunhong.jiang at intel.com<mailto:yunhong.jiang at intel.com>>
Seems it is because PCI frontend try to write some configuration space that PCIback has no config_field entry to support it.
I think you can firstly try to do as dom0's dmesg suggested: "see permissive attribute in sysfs" (it should be "set permissive attribute...", I think).

BTW, where you got following log? That seems suggest config space function not found.

PCI: Fatal: No PCI config space access function found
rtc: IRQ 8 is not free.
i8042.c: No controller found."

-- Yunhong Jiang
________________________________
From: xen-devel-bounces at lists.xensource.com<mailto:xen-devel-bounces at lists.xensource.com> [mailto:xen-devel-bounces at lists.xensource.com<mailto:xen-devel-bounces at lists.xensource.com>] On Behalf Of subbu kl
Sent: 2009年2月11日 22:18
To: David Brown
Cc: xen-devel at lists.xensource.com<mailto:xen-devel at lists.xensource.com>; general at lists.openfabrics.org<mailto:general at lists.openfabrics.org>
Subject: [Xen-devel] Re: [ofa-general] Fwd: pciback module not working

I am getting the same QUERY_FW failed on RHEL5.2 with xenxen paravirtualized guest with pciback module.

No one seems to have tried answering this question on the list, let me ping xen-devel and ofed people again.

after executing in dom0
echo -n 0000:0e:00.0 > /sys/bus/pci/drivers/ib_mthca/unbind
echo -n 0000:0e:00.0 > /sys/bus/pci/drivers/pciback/new_slot
echo -n 0000:0e:00.0 > /sys/bus/pci/drivers/pciback/bind

#dmesg
ACPI: PCI interrupt for device 0000:0e:00.0 disabled
tap tap-1-51712: 2 getting info
tap tap-2-51712: 2 getting info
pciback 0000:0e:00.0: seizing device
PCI: Enabling device 0000:0e:00.0 (0140 -> 0142)
ACPI: PCI Interrupt 0000:0e:00.0[A] -> GSI 16 (level, low) -> IRQ 16
ACPI: PCI interrupt for device 0000:0e:00.0 disabled

#xm create -c rhel52_64_3

PCI: Fatal: No PCI config space access function found
rtc: IRQ 8 is not free.
i8042.c: No controller found.


GUEST dmesg:

ib_mthca: Mellanox InfiniBand HCA driver v1.0 (April 4, 2008)
ib_mthca: Initializing 0000:00:00.0
PCI: Enabling device 0000:00:00.0 (0000 -> 0002)
PCI: Setting latency timer of device 0000:00:00.0 to 64
ib_mthca 0000:00:00.0: QUERY_FW command failed, aborting.
ib_mthca: probe of 0000:00:00.0 failed with error -11

in dom0:
Feb 11 19:44:37 p128 kernel: tap tap-3-51712: 2 getting info
Feb 11 19:44:37 p128 kernel: pciback: vpci: 0000:0e:00.0: assign to virtual slot 0
Feb 11 19:44:37 p128 kernel: device vif3.0 entered promiscuous mode
Feb 11 19:44:37 p128 kernel: ADDRCONF(NETDEV_UP): vif3.0: link is not ready
Feb 11 19:44:39 p128 kernel: blktap: ring-ref 9, event-channel 9, protocol 1 (x86_64-abi)
Feb 11 19:44:48 p128 kernel: pciback 0000:0e:00.0: Driver tried to write to a read-only configuration space field at offset 0x44, size 2. This may be harmless, but if you have problems with your device:
Feb 11 19:44:48 p128 kernel: 1) see permissive attribute in sysfs
Feb 11 19:44:48 p128 kernel: 2) report problems to the xen-devel mailing list along with details of your device obtained from lspci.
Feb 11 19:44:48 p128 kernel: PCI: Enabling device 0000:0e:00.0 (0000 -> 0002)
Feb 11 19:44:48 p128 kernel: ACPI: PCI Interrupt 0000:0e:00.0[A] -> GSI 16 (level, low) -> IRQ 16
Feb 11 19:44:49 p128 kernel: ACPI: PCI interrupt for device 0000:0e:00.0 disabled



some more details - [root at p128 ~]# rpm -qa | grep xen
kernel-xen-2.6.18-92.1.22.el5
xen-3.0.3-64.el5_2.9
xen-libs-3.0.3-64.el5_2.9
xen-libs-3.0.3-64.el5_2.9

[root at p128 ~]# ibv_devinfo
hca_id: mthca0
        fw_ver:                         5.3.0
        node_guid:                      0002:c902:0022:cd48
        sys_image_guid:                 0002:c902:0022:cd4b
        vendor_id:                      0x02c9
        vendor_part_id:                 25218
        hw_ver:                         0x20
        board_id:                       MT_0370130002
        phys_port_cnt:                  2
                port:   1
                        state:                  PORT_INIT (2)
                        max_mtu:                2048 (4)
                        active_mtu:             512 (2)
                        sm_lid:                 0
                        port_lid:               0
                        port_lmc:               0x00

                port:   2
                        state:                  PORT_DOWN (1)
                        max_mtu:                2048 (4)
                        active_mtu:             512 (2)
                        sm_lid:                 0
                        port_lid:               0
                        port_lmc:               0x00


any help greatly appreciated.

~subbu

On Sat, Oct 18, 2008 at 4:54 AM, David Brown <dmlb2000 at gmail.com<mailto:dmlb2000 at gmail.com>> wrote:
Okay so my question to the openfabrics guys is, why would the OFED
drivers fail to read the firmware?

Any thoughts?

Thanks,
- David Brown


---------- Forwarded message ----------
From: David Brown <dmlb2000 at gmail.com<mailto:dmlb2000 at gmail.com>>
Date: Thu, Sep 11, 2008 at 2:24 PM
Subject: pciback module not working
To: xen-users at lists.xensource.com<mailto:xen-users at lists.xensource.com>, xen-devel at lists.xensource.com<mailto:xen-devel at lists.xensource.com>


This issue was brought up about a year and a half ago. So I'll bring
it up again and see if anything happens.

I've got an infiniband network and am attempting to pass the
infiniband card through the host and give it to the guest.
I'm working with standard CentOS 5.2 on both guest and host with their
provided xen (3.0.3 ish). I've also attempted to install the newest
Xen 3.3 and use their standard host kernel and that did the same
thing. The guest dmesg output in the guest is similar on both
permissive and normal mode.

I'm getting issues with detecting the firmware on the card for some reason...

Any help would be appreciated.

Thanks,
- David Brown

=== GUEST dmesg output ===
ib_mthca: Mellanox InfiniBand HCA driver v1.0 (February 28, 2008)
ib_mthca: Initializing 0000:00:00.0
PCI: Enabling device 0000:00:00.0 (0000 -> 0002)
PCI: Setting latency timer of device 0000:00:00.0 to 64
ib_mthca 0000:00:00.0: QUERY_FW command failed, aborting.
ib_mthca: probe of 0000:00:00.0 failed with error -11
=======================

=== Host modprobe.conf ===
alias eth0 bnx2
alias eth1 bnx2
alias scsi_hostadapter cciss
options pciback hide=(41:00.0)
=====================

=== Host lspci output ===
# lspci -vs 41:00.0
41:00.0 InfiniBand: Mellanox Technologies MT25204 [InfiniHost III Lx
HCA] (rev 20)
      Subsystem: Hewlett-Packard Company Unknown device 170a
      Flags: fast devsel, IRQ 16
      Memory at fdc00000 (64-bit, non-prefetchable) [disabled] [size=1M]
      Memory at fd000000 (64-bit, prefetchable) [disabled] [size=8M]
      Capabilities: [40] Power Management version 2
      Capabilities: [48] Vital Product Data
      Capabilities: [90] Message Signalled Interrupts: 64bit+ Queue=0/5 Enable-
      Capabilities: [84] MSI-X: Enable- Mask- TabSize=32
      Capabilities: [60] Express Endpoint IRQ 0
=====================

This makes sure it get loaded first off before anything else.
=== Host mkinitrd cmd ===
# mkinitrd -f --with=pciback --preload pciback
/boot/initrd-2.6.18-92.1.10.el5xen.img 2.6.18-92.1.10.el5xen
====================

=== Host pciback dmesg ===
pciback 0000:41:00.0: Driver tried to write to a read-only
configuration space field at offset 0x44, size 2. This may be
harmless, but if you have problems with your device:
1) see permissive attribute in sysfs
2) report problems to the xen-devel mailing list along with details of
your device obtained from lspci.
PCI: Enabling device 0000:41:00.0 (0000 -> 0002)
ACPI: PCI Interrupt 0000:41:00.0[A] -> GSI 16 (level, low) -> IRQ 16
PCI: Setting latency timer of device 0000:41:00.0 to 64
ACPI: PCI interrupt for device 0000:41:00.0 disabled
======================

=== Host pciback dmesg (after setting it permissive) ===
pciback 0000:41:00.0: enabling permissive mode configuration space accesses!
pciback 0000:41:00.0: permissive mode is potentially unsafe!
pciback: vpci: 0000:41:00.0: assign to virtual slot 0
device vif1.0 entered promiscuous mode
ADDRCONF(NETDEV_UP): vif1.0: link is not ready
blkback: ring-ref 9, event-channel 28, protocol 1 (x86_64-abi)
PCI: Enabling device 0000:41:00.0 (0000 -> 0002)
ACPI: PCI Interrupt 0000:41:00.0[A] -> GSI 16 (level, low) -> IRQ 16
PCI: Setting latency timer of device 0000:41:00.0 to 64
ACPI: PCI interrupt for device 0000:41:00.0 disabled
=========================================

=== Guest lspci output ===
# lspci -v
00:00.0 InfiniBand: Mellanox Technologies MT25204 [InfiniHost III Lx
HCA] (rev 20)
      Subsystem: Hewlett-Packard Company Unknown device 170a
      Flags: fast devsel, IRQ 16
      Memory at fdc00000 (64-bit, non-prefetchable) [disabled] [size=1M]
      Memory at fd000000 (64-bit, prefetchable) [disabled] [size=8M]
      Capabilities: [40] Power Management version 2
      Capabilities: [48] Vital Product Data
      Capabilities: [90] Message Signalled Interrupts: 64bit+
Queue=0/5 Enable-
      Capabilities: [84] MSI-X: Enable- Mask- TabSize=32
      Capabilities: [60] Express Endpoint IRQ 0
=====================
_______________________________________________
general mailing list
general at lists.openfabrics.org<mailto:general at lists.openfabrics.org>
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general



--
. . . s u b b u
"You've got to be original, because if you're like someone else, what do they need you for?"



--
. . . s u b b u
"You've got to be original, because if you're like someone else, what do they need you for?"



--
. . . s u b b u
"You've got to be original, because if you're like someone else, what do they need you for?"



--
. . . s u b b u
"You've got to be original, because if you're like someone else, what do they need you for?"



--
. . . s u b b u
"You've got to be original, because if you're like someone else, what do they need you for?"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/general/attachments/20090212/9f4e8808/attachment.html>


More information about the general mailing list