Bug 85212 – Intel 536EP driver is configured for power management that has been obsolete by kernel 2.6.11.

• Status: VERIFIED FIXED
• Product: Gentoo Linux

[Description Markus Gothe 2005-03-14 05:51:57 UTC]

The CONFIG_PM is on by default, my patch and updated ebuild turns this off and it compiles nicly on 2.6.11.

Reproducible: Always
Steps to Reproduce:
1. emerge intel-536ep on kernel >=2.6.9

Actual Results:  
the module had errors...

Expected Results:  
compile fine

[Comment 1 Markus Gothe 2005-03-14 05:53:47 UTC]

Created attachment 53416 [details]
Updated ebuild

[Comment 2 Markus Gothe 2005-03-14 05:54:37 UTC]

Created attachment 53417 [details, diff]
Patch-file!

[Comment 3 Markus Gothe 2005-03-14 05:57:13 UTC]

Okey here it is, the final frontier... This patch and ebuild solves the PM-probs!

[Comment 4 Markus Gothe 2005-03-14 06:02:12 UTC]

Reopened for assignment

[Comment 5 Markus Gothe 2005-03-14 06:05:25 UTC]

Reassigned bug to [e-mail].

[Comment 6 Alin Năstac (RETIRED) 2005-03-15 01:19:29 UTC]

What errors? On my 2.6.10 it builds without errors. It has some warnings though, but are not related to CONFIG_PM.

[Comment 7 Markus Gothe 2005-03-15 02:43:46 UTC]

/root/intel-536EP-2.56.76.0/coredrv/coredrv.c:295: error: `PM_SAVE_STATE' undeclared (first use in this function)

and when I comment PM_SAVE_STATE-case out it yells about this: 

*** Warning: "pm_access" [/root/intel-536EP-2.56.76.0/coredrv/Intel536.ko] undefined!

So by not using PM this problem is solved.

[Comment 8 Alin Năstac (RETIRED) 2005-03-15 03:12:30 UTC]

please comment the PM_SAVE_STATE case. 
also, you should comment the pm_access call.

[Comment 9 Markus Gothe 2005-03-15 04:49:19 UTC]

Well, I will not... It's THE ugliest way to solve this problem. It's way more correct to #undef CONFIG_PM to avoid future compability issues with the power management in linux.

[Comment 10 Alin Năstac (RETIRED) 2005-03-15 05:14:13 UTC]

the ugliest way will be to remove support for power management.

judging after changes suffered by power management I could say that:
a) pm_access does not exist in 2.6.11. However, in previous versions it was an empty function anyway.
b) PM_SAVE_STATE define has vanished, but this module did nothing but logging a message saying it doesn't support this.

either you test what I tell you to or I will mark this as CANTFIX, since you refuse to do it and I don't have the hardware.

btw, you should post patches as minimal as possible and with correct mime type.

[Comment 11 Markus Gothe 2005-03-15 05:27:56 UTC]

Oki... I'll try... Btw I had a great problem making the patch (i.e. diff -Naur'). Thanks for investigating this futher.

[Comment 12 Markus Gothe 2005-03-15 07:30:17 UTC]

Created attachment 53526 [details, diff]
New patch, patching the PM_SAVE_STATEMENT & pm_access-function

[Comment 13 Markus Gothe 2005-03-15 08:25:46 UTC]

Changed from 2.6.9->2.6.11

[Comment 14 Alin Năstac (RETIRED) 2005-03-15 09:17:51 UTC]

fixed in the same version, since compilation failed on affected systems.
if I've screwed up something, please reopen this bug.

thanks for your support.

[Comment 15 Markus Gothe 2005-03-15 11:47:15 UTC]

I verified it with 'emerge portage', worked out of the box.

[Comment 16 Markus Gothe 2005-03-15 11:49:23 UTC]

Closed it

[Comment 17 Markus Gothe 2005-03-15 12:09:42 UTC]

'emerge intel-536ep' of course