Re: MPC8641D PCI-Express error

8 messages, 4 authors, 2008-02-22 · open the first message on its own page

Re: MPC8641D PCI-Express error

From: Marco Stornelli <hidden>
Date: 2008-02-20 16:12:53

Kumar Gala wrote:
quoted
Marco Stornelli wrote:
quoted
No, but I can try to backport the PCI-E code from 2.6.24 to 2.6.18
if it could help. What do you think about it? Do you think this
problem could be not present in 2.6.24?
I have no idea there, honestly.  Sorry.
As Jon said, try 2.6.24 and see if it has an issue, if so we can look
at helping.  if not, you know you need to back port the fixes.
- k
I've "backported" the PCI-Express code from 2.6.24 to 2.6.18, but it
still doesn't work, I have the same problem (sigh), could you give me
any suggestions?

Thanks.

Marco

Re: MPC8641D PCI-Express error

From: Kumar Gala <hidden>
Date: 2008-02-21 02:21:09

On Feb 20, 2008, at 10:13 AM, Marco Stornelli wrote:
Kumar Gala wrote:
quoted
quoted
Marco Stornelli wrote:
quoted
No, but I can try to backport the PCI-E code from 2.6.24 to 2.6.18
if it could help. What do you think about it? Do you think this
problem could be not present in 2.6.24?
I have no idea there, honestly.  Sorry.
quoted
As Jon said, try 2.6.24 and see if it has an issue, if so we can look
at helping.  if not, you know you need to back port the fixes.
quoted
- k
I've "backported" the PCI-Express code from 2.6.24 to 2.6.18, but it
still doesn't work, I have the same problem (sigh), could you give me
any suggestions?
did 2.6.24 work for you or not?

- k

Re: MPC8641D PCI-Express error

From: Marco Stornelli <hidden>
Date: 2008-02-21 08:01:07

Kumar Gala ha scritto:
On Feb 20, 2008, at 10:13 AM, Marco Stornelli wrote:
quoted
Kumar Gala wrote:
quoted
quoted
Marco Stornelli wrote:
quoted
No, but I can try to backport the PCI-E code from 2.6.24 to 2.6.18
if it could help. What do you think about it? Do you think this
problem could be not present in 2.6.24?
I have no idea there, honestly.  Sorry.
quoted
As Jon said, try 2.6.24 and see if it has an issue, if so we can look
at helping.  if not, you know you need to back port the fixes.
quoted
- k
I've "backported" the PCI-Express code from 2.6.24 to 2.6.18, but it
still doesn't work, I have the same problem (sigh), could you give me
any suggestions?
did 2.6.24 work for you or not?

- k
No, it didn't. I have the same problem even with the 2.6.18 plus the
2.6.24 PCI-Express code. I performed this action because I can't change
kernel version but I can modify it.

Marco

Re: MPC8641D PCI-Express error

From: Timur Tabi <hidden>
Date: 2008-02-21 15:21:41

Marco Stornelli wrote:
No, it didn't. I have the same problem even with the 2.6.18 plus the
2.6.24 PCI-Express code. I performed this action because I can't change
kernel version but I can modify it.
Please try 2.6.24 (or even better, 2.6.25-rc2).  You may have done something 
wrong in back-porting the code to 2.6.18.  There's no reason why you can't at 
least try the latest version of the code for testing purposes, even if you can't 
use it in production.

-- 
Timur Tabi
Linux kernel developer at Freescale

Re: MPC8641D PCI-Express error

From: Jon Loeliger <hidden>
Date: 2008-02-21 15:27:10

Timur Tabi wrote:
Marco Stornelli wrote:
quoted
No, it didn't. I have the same problem even with the 2.6.18 plus the
2.6.24 PCI-Express code. I performed this action because I can't change
kernel version but I can modify it.
Please try 2.6.24 (or even better, 2.6.25-rc2).  You may have done something 
wrong in back-porting the code to 2.6.18.  There's no reason why you can't at 
least try the latest version of the code for testing purposes, even if you can't 
use it in production.
Also, once you back-ported it and all the DTS changes, and all
the fsl_soc.c changes needed and , and, and otherwise made it
look _like_, 2.6.24 but still called it 2.6.18, after all that,
can you remind us what was the error or failure mode that you saw?

jdl

Re: MPC8641D PCI-Express error

From: Marco Stornelli <hidden>
Date: 2008-02-21 15:36:47

Jon Loeliger ha scritto:
Timur Tabi wrote:
quoted
Marco Stornelli wrote:
quoted
No, it didn't. I have the same problem even with the 2.6.18 plus the
2.6.24 PCI-Express code. I performed this action because I can't change
kernel version but I can modify it.
Please try 2.6.24 (or even better, 2.6.25-rc2).  You may have done
something wrong in back-porting the code to 2.6.18.  There's no reason
why you can't at least try the latest version of the code for testing
purposes, even if you can't use it in production.
Also, once you back-ported it and all the DTS changes, and all
the fsl_soc.c changes needed and , and, and otherwise made it
look _like_, 2.6.24 but still called it 2.6.18, after all that,
can you remind us what was the error or failure mode that you saw?

jdl
When I try to read some register from my FPGA (virtex5) I have this bus
error:

Machine check in kernel mode.
Caused by (from SRR1=149030): Transfer error ack signal
Oops: Machine check, sig: 7 [#1]
PREEMPT SMP NR_CPUS=2
Modules linked in: virtex5
LTT NESTING LEVEL : 0
NIP: F108019C LR: F1080198 CTR: 00000001
REGS: c044dd60 TRAP: 0200   Not tainted      (2.6.18-mpc8641d_hpcn)
MSR: 00149030 <EE,ME,IR,DR>  CR: 22000222  XER: 00000000
TASK = c20e9990[568] 'insmod' THREAD: c044c000 CPU: 0
GPR00: F1080198 C044DE10 C20E9990 0000002E 80000000 FFFFFFFF 00008000
00002EA3
GPR08: C20E9990 00000000 C04C0220 C044C000 22000222 1001956C 00000000
00000000
GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
GPR24: 3000EAA0 7FD6FDC0 00000000 C045FCC0 F107E594 F10A0000 00000000
C2036000
NIP [F108019C] virtex5_probe+0x130/0x1c4 [virtex5]
LR [F1080198] virtex5_probe+0x12c/0x1c4 [virtex5]
Call Trace:
[C044DE10] [F1080198] virtex5_probe+0x12c/0x1c4 [virtex5] (unreliable)
[C044DE30] [C01D2A58] pci_device_probe+0x84/0xbc
[C044DE50] [C0213754] driver_probe_device+0x60/0x118
[C044DE70] [C0213890] __driver_attach+0x84/0x88
[C044DE90] [C02130F4] bus_for_each_dev+0x58/0x94
[C044DEC0] [C02135D4] driver_attach+0x24/0x34
[C044DED0] [C0212AC8] bus_add_driver+0x88/0x164
[C044DEF0] [C021397C] driver_register+0x70/0xb8
[C044DF00] [C01D284C] __pci_register_driver+0x64/0x98
[C044DF10] [F1080030] init_module+0x30/0x6c [virtex5]
[C044DF20] [C004CBFC] sys_init_module+0xc8/0x25c
[C044DF40] [C0011358] ret_from_syscall+0x0/0x38
--- Exception: c00 at 0xff6de0c
    LR = 0x10000de4
Instruction dump:
40820060 3c60f108 3863cea0 48000311 807f0238 3c800001 48000395 7c7d1b78
3c60f108 3863ced4 480002f5 809d0000 <3c60f108> 3863cf04 480002e5 3c60f108

Re: MPC8641D PCI-Express error

From: Jon Loeliger <hidden>
Date: 2008-02-21 15:58:43

Marco Stornelli wrote:
When I try to read some register from my FPGA (virtex5) I have this bus
error:

Hmmm....  OK, so if we've eliminated PCI-E as a source
for this issue, it really must be in the Virtex support
somewhere then.  Unfortunately, I know nothing about those,
and am not going to be much direct help there.

Sorry.

jdl

quoted hunk
Machine check in kernel mode.
Caused by (from SRR1=149030): Transfer error ack signal
Oops: Machine check, sig: 7 [#1]
PREEMPT SMP NR_CPUS=2
Modules linked in: virtex5
LTT NESTING LEVEL : 0
NIP: F108019C LR: F1080198 CTR: 00000001
REGS: c044dd60 TRAP: 0200   Not tainted      (2.6.18-mpc8641d_hpcn)
MSR: 00149030 <EE,ME,IR,DR>  CR: 22000222  XER: 00000000
TASK = c20e9990[568] 'insmod' THREAD: c044c000 CPU: 0
GPR00: F1080198 C044DE10 C20E9990 0000002E 80000000 FFFFFFFF 00008000
00002EA3
GPR08: C20E9990 00000000 C04C0220 C044C000 22000222 1001956C 00000000
00000000
GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
GPR24: 3000EAA0 7FD6FDC0 00000000 C045FCC0 F107E594 F10A0000 00000000
C2036000
NIP [F108019C] virtex5_probe+0x130/0x1c4 [virtex5]
LR [F1080198] virtex5_probe+0x12c/0x1c4 [virtex5]
Call Trace:
[C044DE10] [F1080198] virtex5_probe+0x12c/0x1c4 [virtex5] (unreliable)
[C044DE30] [C01D2A58] pci_device_probe+0x84/0xbc
[C044DE50] [C0213754] driver_probe_device+0x60/0x118
[C044DE70] [C0213890] __driver_attach+0x84/0x88
[C044DE90] [C02130F4] bus_for_each_dev+0x58/0x94
[C044DEC0] [C02135D4] driver_attach+0x24/0x34
[C044DED0] [C0212AC8] bus_add_driver+0x88/0x164
[C044DEF0] [C021397C] driver_register+0x70/0xb8
[C044DF00] [C01D284C] __pci_register_driver+0x64/0x98
[C044DF10] [F1080030] init_module+0x30/0x6c [virtex5]
[C044DF20] [C004CBFC] sys_init_module+0xc8/0x25c
[C044DF40] [C0011358] ret_from_syscall+0x0/0x38
--- Exception: c00 at 0xff6de0c
    LR = 0x10000de4
Instruction dump:
40820060 3c60f108 3863cea0 48000311 807f0238 3c800001 48000395 7c7d1b78
3c60f108 3863ced4 480002f5 809d0000 <3c60f108> 3863cf04 480002e5 3c60f108

Re: MPC8641D PCI-Express error

From: Marco Stornelli <hidden>
Date: 2008-02-22 14:18:08

Jon Loeliger ha scritto:
Marco Stornelli wrote:
quoted
When I try to read some register from my FPGA (virtex5) I have this bus
error:

Hmmm....  OK, so if we've eliminated PCI-E as a source
for this issue, it really must be in the Virtex support
somewhere then.  Unfortunately, I know nothing about those,
and am not going to be much direct help there.

Sorry.

jdl

quoted
Machine check in kernel mode.
Caused by (from SRR1=149030): Transfer error ack signal
Oops: Machine check, sig: 7 [#1]
PREEMPT SMP NR_CPUS=2
Modules linked in: virtex5
LTT NESTING LEVEL : 0
NIP: F108019C LR: F1080198 CTR: 00000001
REGS: c044dd60 TRAP: 0200   Not tainted      (2.6.18-mpc8641d_hpcn)
MSR: 00149030 <EE,ME,IR,DR>  CR: 22000222  XER: 00000000
TASK = c20e9990[568] 'insmod' THREAD: c044c000 CPU: 0
GPR00: F1080198 C044DE10 C20E9990 0000002E 80000000 FFFFFFFF 00008000
00002EA3
GPR08: C20E9990 00000000 C04C0220 C044C000 22000222 1001956C 00000000
00000000
GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
GPR24: 3000EAA0 7FD6FDC0 00000000 C045FCC0 F107E594 F10A0000 00000000
C2036000
NIP [F108019C] virtex5_probe+0x130/0x1c4 [virtex5]
LR [F1080198] virtex5_probe+0x12c/0x1c4 [virtex5]
Call Trace:
[C044DE10] [F1080198] virtex5_probe+0x12c/0x1c4 [virtex5] (unreliable)
[C044DE30] [C01D2A58] pci_device_probe+0x84/0xbc
[C044DE50] [C0213754] driver_probe_device+0x60/0x118
[C044DE70] [C0213890] __driver_attach+0x84/0x88
[C044DE90] [C02130F4] bus_for_each_dev+0x58/0x94
[C044DEC0] [C02135D4] driver_attach+0x24/0x34
[C044DED0] [C0212AC8] bus_add_driver+0x88/0x164
[C044DEF0] [C021397C] driver_register+0x70/0xb8
[C044DF00] [C01D284C] __pci_register_driver+0x64/0x98
[C044DF10] [F1080030] init_module+0x30/0x6c [virtex5]
[C044DF20] [C004CBFC] sys_init_module+0xc8/0x25c
[C044DF40] [C0011358] ret_from_syscall+0x0/0x38
--- Exception: c00 at 0xff6de0c
    LR = 0x10000de4
Instruction dump:
40820060 3c60f108 3863cea0 48000311 807f0238 3c800001 48000395 7c7d1b78
3c60f108 3863ced4 480002f5 809d0000 <3c60f108> 3863cf04 480002e5 3c60f108
_______________________________________________
Linuxppc-embedded mailing list
Linuxppc-embedded@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
I tried to use 2.6.24 but I have the same error. It looks like this
problem happens when I try to read something in my ML555 board (Virtex5
evaluation board), but if try to write everything works fine. I tried to
change the settings about inbound/outbound PCI-E windows but I have the
same problem. Have you got any suggestions?

Thanks.

Marco
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help