[#5468] traps test 20 prefetch l1_instruction fails on trunk head
Submitted By: Mingquan Pan
Open Date
2009-08-27 01:19:25
Priority:
Low Assignee:
Sonic Zhang
Status:
Open Fixed In Release:
N/A
Found In Release:
2010R1 Release:
Category:
N/A Board:
N/A
Processor:
BF537 Silicon Revision:
Is this bug repeatable?:
Yes Resolution:
Assigned (Not Start)
Uboot version or rev.:
Toolchain version or rev.:
09r1-rc9
App binary format:
N/A
Summary: traps test 20 prefetch l1_instruction fails on trunk head
Details:
traps test 20 prefetch l1_instruction fails on trunk head.
bfin> bootm
## Booting kernel from Legacy Image at 02000000 ...
Image Name: bf537-2.6.30.5-ADI-2010R1-pre-sv
Created: 2009-08-27 15:07:59 UTC
Image Type: Blackfin Linux Kernel Image (gzip compressed)
Data Size: 4731161 Bytes = 4.5 MB
Load Address: 00001000
Entry Point: 0018b834
Verifying Checksum ... OK
Uncompressing Kernel Image ... OK
Starting Kernel at = 0018b834
Linux version 2.6.30.5-ADI-2010R1-pre-svn7234 (test@uclinux55-bf537-cf-spi) (gcc version 4.1.2 (ADI svn)) #4 Thu Aug 27 09:07:45 MDT 2009
bootconsole [early_shadow0] enabled
bootconsole [early_BFuart0] enabled
early printk enabled on early_BFuart0
Limiting kernel memory to 56MB due to anomaly 05000263
Board Memory: 64MB
Kernel Managed Memory: 64MB
Memory map:
fixedcode = 0x00000400-0x00000490
text = 0x00001000-0x00105f50
rodata = 0x00105f50-0x0015d900
bss = 0x0015e000-0x0016eacc
data = 0x0016eacc-0x00180000
stack = 0x0017e000-0x00180000
init = 0x00180000-0x008c2000
available = 0x008c2000-0x037ff000
DMA Zone = 0x03f00000-0x04000000
Hardware Trace Active and Enabled
Boot Mode: 0
...
BusyBox v1.14.3 (2009-08-27 07:06:43 MDT) hush - the humble shell
root:/> versioPHY: 0:01 - Link is Up - 100/Full
n
kernel: Linux release 2.6.30.5-ADI-2010R1-pre-svn7234, build #4 Thu Aug 27 09:07:45 MDT 2009
toolchain: bfin-uclinux-gcc release gcc version 4.1.2 (ADI svn)
user-dist: release svn-8772, build #43 Thu Aug 27 09:06:53 MDT 2009
root:/> successful boot attempt
************** STEP 3: Start testing.
uname -a
Linux blackfin 2.6.30.5-ADI-2010R1-pre-svn7234 #4 Thu Aug 27 09:07:45 MDT 2009 blackfin GNU/Linux
root:/>
It's not a SMP kernel.
cd /bin
root:/bin> ./traps_test #
94
rootTotal case num is 95.
:/bin> ./traps_test -1
Running test 0 for exception 0x00: flush scratch pad
... PASS (test completed 1/1 times, as expected by signal 0: Unknown signal 0)
...
Running test 20 for exception 0x00: prefetch l1_instruction
... FAIL (test failed, but not with the right signal)
(We expected 4 'Illegal instruction' but instead we got 11 'Segmentation fault')
FAIL (test completed properly 0/1 times)
Running test 21 for exception 0x00: prefetch _l1_non
... PASS (test completed 1/1 times, as expected by signal 0: Unknown signal 0)
Follow-ups
--- Robin Getz 2009-09-01 07:11:30
This is a new instruction anomaly -- which the silicon folks have confirmed, but
not root caused, or provided a workaround, or added to the anomaly sheets.
As soon as that happens, we will need to change things - until then -- there
isn't much to do...???
-Robin
--- Mingquan Pan 2009-09-22 02:53:39
On bf548 ezkit and bf518f-ezkit, there is one more fail test 21. It is not found
failing on other boards.
Running test 19 for exception 0x00: prefetch l1_dataB^M
... PASS (test completed 1/1 times, as expected by signal 0: Unknown signal
0)^M
^M
Running test 20 for exception 0x00: prefetch l1_instruction^M
... FAIL (test exited sucessfully, and we expected 4 Illegal instruction
failure^M
FAIL (test completed properly 0/1 times)^M
^M
Running test 21 for exception 0x00: prefetch _l1_non^M
... FAIL (test failed, but not with the right signal)^M
(We expected 0 'Unknown signal 0' but instead we got 11 'Segmentation
fault')^M
FAIL (test completed properly 0/1 times)^M
^M
--- Mingquan Pan 2009-10-16 04:25:18
It is strange that when the testing target board is switching to bf527 ezkit 0.2
on Sep 22, case 21 also fails. While this case used to be passed on bf527 ezkit
0.0.
Running test 21 for exception 0x00: prefetch _l1_non^M
... FAIL (test failed, but not with the right signal)^M
(We expected 0 'Unknown signal 0' but instead we got 11 'Segmentation
fault')^M
FAIL (test completed properly 0/1 times)^M
The revision changes are:
failed revision: uclinux-dist 8880 linux 7402
passed revision: uclinux-dist 8871 linux 7391
--- Sonic Zhang 2009-11-11 05:05:36
I move the fail of traps test 21 to a new bug
5674 traps test 21 prefetch _l1_no fails on trunk head On bf548 and bf518 and
bf527
Barry will take a look at that first.
--- Robin Getz 2009-11-12 11:44:13
There are a bunch of known (and undocumented) silicon errata (anomalies) which
have to do with prefetch.
Until these are figured out -- it might be best to just to temporarily remove
all the prefetch tests in traps tests.
-Robin
--- Sonic Zhang 2010-04-16 05:39:03
Vivi, please remove all prefetch tests in traps tests.
--- Vivi Li 2010-04-19 00:00:10
Traps test passed.
So close this bug.
--- Vivi Li 2010-06-07 05:04:20
The problem still exists. Test script is modified to ignore the error of running
prefetch test in traps.
Files
Changes
Commits
Dependencies
Duplicates
Associations
Tags
File Name File Type File Size Posted By
No Files Were Found