Bug 13580 - /var/log/message contains non-printable character
Summary: /var/log/message contains non-printable character
Status: CLOSED CODE_FIX
Alias: None
Product: ACPI
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: acpi_other
URL: https://bugs.gentoo.org/show_bug.cgi?...
Keywords:
Depends on:
Blocks:
 
Reported: 2009-06-19 11:55 UTC by Toralf Förster
Modified: 2009-11-25 05:44 UTC (History)
3 users (show)

See Also:
Kernel Version: 2.6.29.5
Subsystem:
Regression: No
Bisected commit-id:

Attachments
log line (104 bytes, text/plain)
2009-06-19 11:55 UTC, Toralf Förster 		Details
dmesg (34.53 KB, text/plain)
2009-06-19 18:53 UTC, Toralf Förster 		Details
/tmp/acpidump (314.24 KB, text/plain)
2009-06-19 18:57 UTC, Toralf Förster 		Details
patch: support non-printable character in ACPI table header (1.75 KB, patch)
2009-06-22 02:15 UTC, Zhang Rui 		Details | Diff
patch from Bob (4.75 KB, patch)
2009-06-29 07:41 UTC, Lin Ming 		Details | Diff
Add an attachment (proposed patch, testcase, etc.)

Description Toralf Förster 2009-06-19 11:55:51 UTC 
Created attachment 22007 [details]
log line

The attached file is a line which I have within my /var/log/messages of my
ThinkPad T400. Unfortunateley that's why less won't show that file as text :-(
Comment 1 Robert Moore 2009-06-19 18:05:51 UTC 
Please attach the full message log and the acpidump for the machine.
Comment 2 Toralf Förster 2009-06-19 18:53:53 UTC 
Created attachment 22011 [details]
dmesg 

attached is the dmesg output, I can't find acpidump within Gentoo's repository.
Comment 3 Toralf Förster 2009-06-19 18:57:20 UTC 
Created attachment 22012 [details]
/tmp/acpidump

Gentoo package name was pmtools ...
Comment 4 Robert Moore 2009-06-19 19:31:45 UTC 
From the acpidump, your ACPI DMAR table has an unusual string in the header:

DMAR @ 0x7c706000
  0000: 44 4d 41 52 20 01 00 00 01 46 20 20 20 20 20 20  DMAR ....F      
  0010: 01 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00  ................
  0020: 00 00 00 00 23 00 00 00 00 00 00 00 00 00 00 00  ....#...........

The value starting at offset 0010 is the "OEM Table ID". It starts with 0x01, which is not a printable ascii character.

The function that dumps the table headers should probably handle non-printable characters.
Comment 5 Zhang Rui 2009-06-22 01:52:21 UTC 
IMO, this is clearly a BIOS bug to me.
Bob, do you mean you want to workaround it in ACPICA core?
Comment 6 Zhang Rui 2009-06-22 02:15:13 UTC 
Created attachment 22040 [details]
patch: support non-printable character in ACPI table header

Bob, do you mean a patch like this?
Toralf, can you please try this patch and see if it helps?
Comment 7 Robert Moore 2009-06-22 02:25:58 UTC 
This comes up every once in a while. Enough so that I think we should just workaround it.

I have a fix for it already. I take a copy of the header and stuff a '?' for every non-printable character.
Comment 8 Zhang Rui 2009-06-22 02:30:59 UTC 
that's great.
can you attach it here and close the bug report please? :)
Comment 9 Robert Moore 2009-06-22 02:33:19 UTC 
It is against the ACPICA core, so it has to wait until Linux integration.
Comment 10 Lin Ming 2009-06-29 07:41:34 UTC 
Created attachment 22136 [details]
patch from Bob
Comment 11 Zhang Rui 2009-06-29 07:49:25 UTC 
great.
toralf,
The patch should be shipped in the next ACPICA release and soon hit upsteam kernl.
please re-open it if the patch doesn't work for you.
Comment 12 Len Brown 2009-08-29 22:31:49 UTC 
this patch is in the queue for 2.6.32
cf02cd47d4747abf8ff0617e15fc05a00202e6d5

    ACPICA: Dump table header - suppress output of non-printable characters
    
    Function acpi_tb_print_table_header. Some ACPI tables contain
    non-printable characters in one of the string fields of the the
    header - Signature, OemId, OemTableId, or CompilerId. Invalid
    characters are replaced by '?'. ACPICA BZ 788.
    
    http://acpica.org/bugzilla/show_bug.cgi?id=788
Comment 13 Len Brown 2009-11-25 03:02:51 UTC 
patch above shipped in 2.6.32-rc2

closed.
Note You need to log in before you can comment on or make changes to this bug.