[v3.11][3.12][Regression] ACPI / processor: Use common hotplug infrastructure

[v3.11][3.12][Regression] ACPI / processor: Use common hotplug infrastructure

[Joseph Salisbury]

Hi Rafael,

A kernel bug affecting several users was opened against Ubuntu[0]. 
After a bisect, it was found the following commit introduced the regression:

commit ac212b6980d8d5eda705864fc5a8ecddc6d6eacc
Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Date: Fri May 3 00:26:22 2013 +0200

    ACPI / processor: Use common hotplug infrastructure

The regression was introduced as of v3.11-rc1.

This commit doesn't revert cleanly in 3.12, and requires the revert of
other commits as well.  I wanted to get your feedback since you are the
author.

   
Thanks,
   
Joe

http://pad.lv/1237336

Re: [v3.11][3.12][Regression] ACPI / processor: Use common hotplug infrastructure

[Rafael J. Wysocki]

On Friday, November 15, 2013 12:44:11 PM Joseph Salisbury wrote:
> Hi Rafael,
> 
> A kernel bug affecting several users was opened against Ubuntu[0]. 
> After a bisect, it was found the following commit introduced the regression:
> 
> commit ac212b6980d8d5eda705864fc5a8ecddc6d6eacc
> Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> Date: Fri May 3 00:26:22 2013 +0200
> 
>     ACPI / processor: Use common hotplug infrastructure
> 
> The regression was introduced as of v3.11-rc1.
> 
> This commit doesn't revert cleanly in 3.12, and requires the revert of
> other commits as well.  I wanted to get your feedback since you are the
> author.

Well, that commit has nothing to do with PCMCIA, so I'm not sure how the
breakage is related to it.

Moreover, comment #8 reports 3.11.0-11.17 as working and the change you're
asking about should be in that kernel as well.

Thanks!

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [v3.11][3.12][Regression] ACPI / processor: Use common hotplug infrastructure

[Rafael J. Wysocki]

On Friday, November 15, 2013 11:55:10 PM Rafael J. Wysocki wrote:
> On Friday, November 15, 2013 12:44:11 PM Joseph Salisbury wrote:
> > Hi Rafael,
> > 
> > A kernel bug affecting several users was opened against Ubuntu[0]. 
> > After a bisect, it was found the following commit introduced the regression:
> > 
> > commit ac212b6980d8d5eda705864fc5a8ecddc6d6eacc
> > Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > Date: Fri May 3 00:26:22 2013 +0200
> > 
> >     ACPI / processor: Use common hotplug infrastructure
> > 
> > The regression was introduced as of v3.11-rc1.
> > 
> > This commit doesn't revert cleanly in 3.12, and requires the revert of
> > other commits as well.  I wanted to get your feedback since you are the
> > author.
> 
> Well, that commit has nothing to do with PCMCIA, so I'm not sure how the
> breakage is related to it.
> 
> Moreover, comment #8 reports 3.11.0-11.17 as working and the change you're
> asking about should be in that kernel as well.

Ah, so that was bogus.

Well, I'm still not sure how the commit in question may affect PCMCIA.  I have
a test machine with PCMCIA and it boots correctly with 3.11+.

Thanks!

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [v3.11][3.12][Regression] ACPI / processor: Use common hotplug infrastructure

[Rafael J. Wysocki]

On Saturday, November 16, 2013 12:00:50 AM Rafael J. Wysocki wrote:
> On Friday, November 15, 2013 11:55:10 PM Rafael J. Wysocki wrote:
> > On Friday, November 15, 2013 12:44:11 PM Joseph Salisbury wrote:
> > > Hi Rafael,
> > > 
> > > A kernel bug affecting several users was opened against Ubuntu[0]. 
> > > After a bisect, it was found the following commit introduced the regression:
> > > 
> > > commit ac212b6980d8d5eda705864fc5a8ecddc6d6eacc
> > > Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > Date: Fri May 3 00:26:22 2013 +0200
> > > 
> > >     ACPI / processor: Use common hotplug infrastructure
> > > 
> > > The regression was introduced as of v3.11-rc1.
> > > 
> > > This commit doesn't revert cleanly in 3.12, and requires the revert of
> > > other commits as well.  I wanted to get your feedback since you are the
> > > author.
> > 
> > Well, that commit has nothing to do with PCMCIA, so I'm not sure how the
> > breakage is related to it.
> > 
> > Moreover, comment #8 reports 3.11.0-11.17 as working and the change you're
> > asking about should be in that kernel as well.
> 
> Ah, so that was bogus.
> 
> Well, I'm still not sure how the commit in question may affect PCMCIA.  I have
> a test machine with PCMCIA and it boots correctly with 3.11+.

The bisect could just trip over a different bug in that commit that was fixed
later.

Moreover, comment #70 indicates that 3.11.0-11-generic works correctly for
someone.

I'm also unsure what comment #43 means.

And I wonder if the reported who "bisected" this can boot without PCMCIA?

Rafael

Re: [v3.11][3.12][Regression] ACPI / processor: Use common hotplug infrastructure

[Rafael J. Wysocki]

On Saturday, November 16, 2013 12:16:05 AM Rafael J. Wysocki wrote:
> On Saturday, November 16, 2013 12:00:50 AM Rafael J. Wysocki wrote:
> > On Friday, November 15, 2013 11:55:10 PM Rafael J. Wysocki wrote:
> > > On Friday, November 15, 2013 12:44:11 PM Joseph Salisbury wrote:
> > > > Hi Rafael,
> > > > 
> > > > A kernel bug affecting several users was opened against Ubuntu[0]. 
> > > > After a bisect, it was found the following commit introduced the regression:
> > > > 
> > > > commit ac212b6980d8d5eda705864fc5a8ecddc6d6eacc
> > > > Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > > Date: Fri May 3 00:26:22 2013 +0200
> > > > 
> > > >     ACPI / processor: Use common hotplug infrastructure
> > > > 
> > > > The regression was introduced as of v3.11-rc1.
> > > > 
> > > > This commit doesn't revert cleanly in 3.12, and requires the revert of
> > > > other commits as well.  I wanted to get your feedback since you are the
> > > > author.
> > > 
> > > Well, that commit has nothing to do with PCMCIA, so I'm not sure how the
> > > breakage is related to it.
> > > 
> > > Moreover, comment #8 reports 3.11.0-11.17 as working and the change you're
> > > asking about should be in that kernel as well.
> > 
> > Ah, so that was bogus.
> > 
> > Well, I'm still not sure how the commit in question may affect PCMCIA.  I have
> > a test machine with PCMCIA and it boots correctly with 3.11+.
> 
> The bisect could just trip over a different bug in that commit that was fixed
> later.
> 
> Moreover, comment #70 indicates that 3.11.0-11-generic works correctly for
> someone.
> 
> I'm also unsure what comment #43 means.
> 
> And I wonder if the reported who "bisected" this can boot without PCMCIA?

One more thing.  Comment #32 says that the reporter could not reproduce the
problem without "the -extras package which contains the PCMCIA drivers."  So
I'm wondering why exactly the PCMCIA drivers are in that package?

Rafael