Bug 3516 - Problems when compiled with gcc3
Summary: Problems when compiled with gcc3
Status: REJECTED INSUFFICIENT_DATA
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Battery (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: Luming Yu
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-10-05 14:09 UTC by Hans Fugal
Modified: 2006-11-26 07:54 UTC (History)
3 users (show)

See Also:
Kernel Version: 2.6.8.1
Subsystem:
Regression: ---
Bisected commit-id:


Attachments
Compaq Evo N150 acpidump (21.24 KB, application/x-bzip2)
2005-11-14 09:05 UTC, Federico Nu
Details
dmesg with kernel 2.6.14rc4 compiled with gcc2.95 here battery and thermal status work. (4.51 KB, application/x-bzip2)
2005-11-14 09:07 UTC, Federico Nu
Details
dmesg with kernel 2.6.14rc4 compiled with gcc3.4 here acpi don't work. (3.15 KB, application/x-bzip2)
2005-11-14 09:12 UTC, Federico Nu
Details

Description Hans Fugal 2004-10-05 14:09:12 UTC
Distribution: Debian Sarge
Hardware Environment: Compaq Evo N150 Laptop
Software Environment:
Problem Description:
I was having problems that seemed like DSDT problems. (useless ACPI info such as
battery, thermal, etc.) As it turns out, the problems disappear when I compile
with gcc-2.95.

I'm sorry I don't have more of an idea as to _what_ is wrong, but the difference
is night and day. With gcc3, acpi reports that my cpu is 71 degC always, 2%
battery, inaccurate A/C status, etc. With gcc-2.95, acpi works perfectly. I
'fixed' the DSDT with absolutely no improvement in gcc3.

Steps to reproduce:


I would be happy to help nail down the problem, but I've done what I can without
instruction/ideas from the developers.
Comment 1 Len Brown 2004-11-10 20:31:18 UTC
still an issue with 2.6.9?

I'm running gcc 3.3 and gcc 3.3.3 w/o problems.
Comment 2 Hans Fugal 2004-11-15 08:39:17 UTC
Sorry for the delay. 2.6.9 is exactly the same, it is not fixed.
Comment 3 Len Brown 2004-11-17 12:59:45 UTC
If you compile the 2.6.9 kernel and modules with gcc-2.95 and run it,
you have no problems.  But if you compile it with a newer compiler
you have problems?

Exactly what compiler are you running and where did you get it?
What does gcc --version say?

Can you be more specific about the problems you do see with the
new compiler?

Are the problems with the new compiler only seen in the ACPI
sub-system, or are they seen elsewhere too?

I don't understand the reference to fixing the DSDT.
Are you running the latest BIOS from Compaq for this box,
or have you made modifications to the DSDT?
Comment 4 Hans Fugal 2004-11-17 13:34:37 UTC
I haven't tested 2.6.9 with gcc-2.9.5 yet, I only tested gcc 3.3.4 and verified
that it did the same thing. I will try it with gcc-2.95 for thoroughness asap.

fugalh@arawn:~$ gcc --version
gcc (GCC) 3.3.4 (Debian 1:3.3.4-13)
Copyright (C) 2003 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

The problem I see is that much ACPI information reported in /proc/acpi is
completely wrong. The most obvious example is that the CPU temperature is
reported as 71 degrees celsius no matter whether the laptop has been on for
hours or just barely turned on after taking it out of the refridgerator. A/C
status and battery are also wrong.

I have only noticed the problems in the ACPI subsystem. I haven't noticed any
troubles in the rest of the kernel, on this laptop or on any other machines
running the 2.6 series compiled with gcc 3.

The DSDT reference is because the problem I'm seeing sounded like something DSDT
hacking would fix. I hacked the DSDT to get rid of the compiler errors and
warnings, but it didn't change anything. I am presently running the BIOS's DSDT.

I am running the latest BIOS from Compaq, and it was a painful process to get
it. (Had to install Windows to do it.)
Comment 5 Len Brown 2005-01-03 18:29:33 UTC
where did you get gcc 3.3.4 - did it come with Debian?
is it i386 or x86_64?
Comment 6 Hans Fugal 2005-01-04 07:06:14 UTC
Yes, from Debian. i386. FWIW I tried it again just the other day with 2.6.10,
same results.
Comment 7 Luming Yu 2005-08-21 00:11:39 UTC
Please reopen the bug if you fail with solution described at bug 3974.
I suspect same brand have same issue.

*** This bug has been marked as a duplicate of 3974 ***
Comment 8 Hans Fugal 2005-08-22 14:55:06 UTC
This may or may not be the same thing. The symptoms there don't sound exactly
the same as what I was experiencing, but they may be related. In any case I
unfortunately don't have access to this laptop anymore and can't be any more
help as a tester.
Comment 9 Federico Nu 2005-10-19 08:59:39 UTC
I have the same laptop (Evo N150) and the same problem (compiling with gcc-2.95 
make acpi work, gcc-3 >= don't) 
 
I have test linux-2.6.14-rc4 with ec_burst=1 (I try 0 and 1)  like Luming 
suggest to me but nothing change here. 
The acpi continue not working property. 
 
bye, fede 
Comment 10 Luming Yu 2005-10-20 06:43:41 UTC
Please post acpidump output.
Comment 11 Luming Yu 2005-10-23 00:00:49 UTC
Please also try latest gcc?
Comment 12 Federico Nu 2005-10-25 19:27:38 UTC
I send all my homeworks!!
bye.
Comment 13 Federico Nu 2005-11-14 09:05:35 UTC
Created attachment 6583 [details]
Compaq Evo N150 acpidump

The original acpidump with all from factory.
Comment 14 Federico Nu 2005-11-14 09:07:17 UTC
Created attachment 6584 [details]
dmesg with kernel 2.6.14rc4 compiled with gcc2.95 here battery and thermal status work.
Comment 15 Federico Nu 2005-11-14 09:12:15 UTC
Created attachment 6585 [details]
dmesg with kernel 2.6.14rc4 compiled with gcc3.4 here acpi don't  work.
Comment 16 Luming Yu 2005-11-28 05:32:59 UTC
3.04 dmesg:
[17179621.180000] ACPI: Battery Slot [BAT2] (battery present)
2.95 dmesg:
[17179616.392000] ACPI: Battery Slot [BAT2] (battery absent)

I assume you were testing with only one battery. So, I think this is a clue.
Comment 17 Len Brown 2006-08-10 02:08:56 UTC
is this still an issue?

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