Bug 729538

Summary: kernel-3.0.1-3.fc16.i686 boot process waits for about 20 sec before it starts
Product: [Fedora] Fedora Reporter: sd.domrep
Component: kernelAssignee: Kernel Maintainer List <kernel-maint>
Status: CLOSED WORKSFORME QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: medium Docs Contact:
Priority: unspecified    
Version: 16CC: aquini, gansalmon, harald, itamar, jonathan, jwboyer, kernel-maint, madhu.chinakonda
Target Milestone: ---   
Target Release: ---   
Hardware: i686   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-08-15 11:34:10 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
dmesg-kernel-3.0.1-2.txt
none
dmesg-kernel-3.0.1-3.txt
none
lspci.txt
none
boot video 22 sec none

Description sd.domrep 2011-08-10 06:05:03 UTC 
Description of problem:
The problem is - with kernel-3.0.1-3.fc16.i686 there is a halt for about 20 sec before it starts to boot everything.
I see just black screen for ~ 20 sec.

Kernel-3.0.1-2.fc16 starts to boot without any delays. 


Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:
Comment 1 Josh Boyer 2011-08-10 11:45:11 UTC 
Can you attach /var/log/messages from both 3.0.1-2.fc16 and 3.0.1-3.fc16?

What kind of machine?

The only difference between those two kernels is a small utrace update.
Comment 2 sd.domrep 2011-08-10 13:45:24 UTC 
No problem.
I want to say again that this halt occurs before I see "Probing EDD ..." on my screen.
This is laptop Lenovo 3000 C100

I attached dmesges and "lspci"

Thanks.
Comment 3 sd.domrep 2011-08-10 13:46:25 UTC 
Created attachment 517609 [details]
dmesg-kernel-3.0.1-2.txt
Comment 4 sd.domrep 2011-08-10 13:47:31 UTC 
Created attachment 517610 [details]
dmesg-kernel-3.0.1-3.txt
Comment 5 sd.domrep 2011-08-10 13:48:38 UTC 
Created attachment 517611 [details]
lspci.txt
Comment 6 sd.domrep 2011-08-11 13:00:16 UTC 
kernel-3.0.1-5.fc16 - there is no difference, the same ~ 20 sec delay before boot starts.

Thanks.
Comment 7 Josh Boyer 2011-08-11 13:49:25 UTC 
You seem to have i915.i915_enable_fbc=1 set on your kernel command line for 3.0.1-3.  Could you try removing that and seeing if the behavior matches 3.0.1-2?

I don't see much else that is different in the boot output from the attached dmesgs.  It seems both boots progress at approximately the same speed, so perhaps the console device simply isn't being updated.
Comment 8 sd.domrep 2011-08-11 15:14:01 UTC 
It does nothing to kernel - it is for i915 driver.
And I removed it in kernel-3.0.1-5.fc16 anyway.
You wont see anything in dmesg, because the 20 sec pause occur before any logging starts.
This halt occurs before I see "Probing EDD ..." on my screen. Immediately after I choose which kernel to boot through grub.
Later I can take short video if one do not believe me.
Comment 9 Josh Boyer 2011-08-11 18:37:11 UTC 
(In reply to comment #8)
> It does nothing to kernel - it is for i915 driver.
> And I removed it in kernel-3.0.1-5.fc16 anyway.
> You wont see anything in dmesg, because the 20 sec pause occur before any
> logging starts.
> This halt occurs before I see "Probing EDD ..." on my screen. Immediately after
> I choose which kernel to boot through grub.
> Later I can take short video if one do not believe me.

Sorry, I missed the "before Probing EDD" part.  That's called extremely early in the boot, and there isn't much before it.  Adding 'debug' to the kernel command line might help show if it's hung even before we get to validating the CPU or not, but that is going to be of somewhat limited value.
Comment 10 sd.domrep 2011-08-12 06:49:21 UTC 
Ok, I took short (for 22 sec) video, there is no much to see, sorry.
At the end of the halt period with "debug" there is only one new line appears:
"early console in setup code".
Comment 11 sd.domrep 2011-08-12 06:53:58 UTC 
Created attachment 517969 [details]
boot video 22 sec
Comment 12 Josh Boyer 2011-08-12 12:39:28 UTC 
(In reply to comment #10)
> Ok, I took short (for 22 sec) video, there is no much to see, sorry.
> At the end of the halt period with "debug" there is only one new line appears:
> "early console in setup code".

OK, that means it's likely either the bootloader or the firmware are adding the additional delay.  There are basically only a handful of lines of assembly code that are run before that "early console in setup code" is printed.
Comment 13 sd.domrep 2011-08-13 07:47:41 UTC 
Good news.

In the last few updates there was anaconda-16.14.4-1.fc16 and dracut-013-1.fc16 update.
I did those updates and then reinstalled the latest kernel from koji - kernel-3.0.1-5.fc16.
And there is no halts any more - problem solved.

Thanks.
Comment 14 Josh Boyer 2011-08-15 11:34:10 UTC 
Glad to hear it.
Comment 15 Harald Hoyer 2011-08-22 08:58:28 UTC 
(In reply to comment #3)
> Created attachment 517609 [details]
> dmesg-kernel-3.0.1-2.txt

[    5.936077] dracut: dracut-011-15.git20110720
Comment 16 Harald Hoyer 2011-08-22 08:59:14 UTC 
(In reply to comment #4)
> Created attachment 517610 [details]
> dmesg-kernel-3.0.1-3.txt

[    5.937923] dracut: dracut-011-15.git20110720