Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-09 Thread Peter Zijlstra
On Wed, Apr 09, 2014 at 11:14:38AM +0200, Stefan Bader wrote: > On 08.04.2014 14:21, Peter Zijlstra wrote: > > On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: > >> On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: > >> > >>> v3.14-10353-g2b3a8fd works fine AFAICS >

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-09 Thread Stefan Bader
On 08.04.2014 14:21, Peter Zijlstra wrote: > On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: >> On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: >> >>> v3.14-10353-g2b3a8fd works fine AFAICS >>> (BTW the fix is stable material, right ?) >> >> I'm fairly sure its not;

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-09 Thread Stefan Bader
On 08.04.2014 14:21, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is stable material, right ?) I'm fairly sure its not; its a rather

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-09 Thread Peter Zijlstra
On Wed, Apr 09, 2014 at 11:14:38AM +0200, Stefan Bader wrote: On 08.04.2014 14:21, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/08/2014 02:21 PM, Peter Zijlstra wrote: > On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: >> On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: >> >>> v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is stable

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Michele Ballabio
On 08/04/2014 21:14, Michele Ballabio wrote: > On 08/04/2014 14:21, Peter Zijlstra wrote: >> On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: >>> On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: >>> v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Michele Ballabio
On 08/04/2014 14:21, Peter Zijlstra wrote: > On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: >> On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: >> >>> v3.14-10353-g2b3a8fd works fine AFAICS >>> (BTW the fix is stable material, right ?) >> >> I'm fairly sure its not;

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Peter Zijlstra
On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: > On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: > > > v3.14-10353-g2b3a8fd works fine AFAICS > > (BTW the fix is stable material, right ?) > > I'm fairly sure its not; its a rather invasive series; see: > >

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Peter Zijlstra
On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is stable material, right ?) I'm fairly sure its not; its a rather invasive series; see: 2432e1364bbe

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Michele Ballabio
On 08/04/2014 14:21, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is stable material, right ?) I'm fairly sure its not; its a rather

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Michele Ballabio
On 08/04/2014 21:14, Michele Ballabio wrote: On 08/04/2014 14:21, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is stable material,

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-08 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/08/2014 02:21 PM, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:56:58PM +0200, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is stable

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Frederic Weisbecker
On Mon, Apr 07, 2014 at 09:57:00PM +0200, Toralf Förster wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA256 > > On 04/07/2014 08:59 PM, Frederic Weisbecker wrote: > > On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: > >> -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 > >>

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/07/2014 08:59 PM, Frederic Weisbecker wrote: > On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: >> -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 >> >> On 04/07/2014 05:07 PM, Peter Zijlstra wrote: >>> On Mon, Apr 07, 2014 at

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Michele Ballabio
On 07/04/2014 17:07, Peter Zijlstra wrote: > On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: >> So what I suspect at this point is that because i386 and x86_64 have a >> difference in current_thread_info() (i386 is stack based), we end up >> setting the TIF_NEED_RESCHED bit on the

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Frederic Weisbecker
On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA256 > > On 04/07/2014 05:07 PM, Peter Zijlstra wrote: > > On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: > >> So what I suspect at this point is that because i386 and

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Peter Zijlstra
On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: > v3.14-10353-g2b3a8fd works fine AFAICS > (BTW the fix is stable material, right ?) I'm fairly sure its not; its a rather invasive series; see: 2432e1364bbe x86: Nuke the supervisor_stack field in i386 thread_info b807902a88c4

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/07/2014 05:07 PM, Peter Zijlstra wrote: > On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: >> So what I suspect at this point is that because i386 and x86_64 >> have a difference in current_thread_info() (i386 is stack based),

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Peter Zijlstra
On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: > So what I suspect at this point is that because i386 and x86_64 have a > difference in current_thread_info() (i386 is stack based), we end up > setting the TIF_NEED_RESCHED bit on the wrong stack. > > Now I have some vague memories

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Peter Zijlstra
On Sun, Apr 06, 2014 at 05:19:27PM +0200, Michele Ballabio wrote: > Toralf Förster reported this in > http://article.gmane.org/gmane.linux.kernel/1662567 > http://article.gmane.org/gmane.linux.kernel/1658422 > http://article.gmane.org/gmane.linux.kernel/1657962 > > "The issue happens here

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Peter Zijlstra
On Sun, Apr 06, 2014 at 05:19:27PM +0200, Michele Ballabio wrote: Toralf Förster reported this in http://article.gmane.org/gmane.linux.kernel/1662567 http://article.gmane.org/gmane.linux.kernel/1658422 http://article.gmane.org/gmane.linux.kernel/1657962 The issue happens here at a

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Peter Zijlstra
On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: So what I suspect at this point is that because i386 and x86_64 have a difference in current_thread_info() (i386 is stack based), we end up setting the TIF_NEED_RESCHED bit on the wrong stack. Now I have some vague memories of

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/07/2014 05:07 PM, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: So what I suspect at this point is that because i386 and x86_64 have a difference in current_thread_info() (i386 is stack based), we

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Peter Zijlstra
On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: v3.14-10353-g2b3a8fd works fine AFAICS (BTW the fix is stable material, right ?) I'm fairly sure its not; its a rather invasive series; see: 2432e1364bbe x86: Nuke the supervisor_stack field in i386 thread_info b807902a88c4 x86:

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Frederic Weisbecker
On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/07/2014 05:07 PM, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: So what I suspect at this point is that because i386 and x86_64

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Michele Ballabio
On 07/04/2014 17:07, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 05:03:37PM +0200, Peter Zijlstra wrote: So what I suspect at this point is that because i386 and x86_64 have a difference in current_thread_info() (i386 is stack based), we end up setting the TIF_NEED_RESCHED bit on the wrong

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/07/2014 08:59 PM, Frederic Weisbecker wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/07/2014 05:07 PM, Peter Zijlstra wrote: On Mon, Apr 07, 2014 at 05:03:37PM

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-07 Thread Frederic Weisbecker
On Mon, Apr 07, 2014 at 09:57:00PM +0200, Toralf Förster wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/07/2014 08:59 PM, Frederic Weisbecker wrote: On Mon, Apr 07, 2014 at 08:16:24PM +0200, Toralf Förster wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-06 Thread Michele Ballabio
On 06/04/2014 17:52, Toralf Förster wrote: >> > Reverting either one of them solves the problem reported with kvm, >> > but revert is probably not the correct answer. >> > >> > I wonder if the solution is as simple as this: >> > >> > --->8--- >> > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-06 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/06/2014 05:19 PM, Michele Ballabio wrote: > Toralf Förster reported this in > http://article.gmane.org/gmane.linux.kernel/1662567 > http://article.gmane.org/gmane.linux.kernel/1658422 > http://article.gmane.org/gmane.linux.kernel/1657962

Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-06 Thread Michele Ballabio
Toralf Förster reported this in http://article.gmane.org/gmane.linux.kernel/1662567 http://article.gmane.org/gmane.linux.kernel/1658422 http://article.gmane.org/gmane.linux.kernel/1657962 "The issue happens here at a 32 bit stable Gentoo Linux if I try to start a KVM image. Kernels

Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-06 Thread Michele Ballabio
Toralf Förster reported this in http://article.gmane.org/gmane.linux.kernel/1662567 http://article.gmane.org/gmane.linux.kernel/1658422 http://article.gmane.org/gmane.linux.kernel/1657962 The issue happens here at a 32 bit stable Gentoo Linux if I try to start a KVM image. Kernels

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-06 Thread Toralf Förster
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 04/06/2014 05:19 PM, Michele Ballabio wrote: Toralf Förster reported this in http://article.gmane.org/gmane.linux.kernel/1662567 http://article.gmane.org/gmane.linux.kernel/1658422 http://article.gmane.org/gmane.linux.kernel/1657962

Re: Bisected KVM hang on x86-32 between v3.12 and v3.13

2014-04-06 Thread Michele Ballabio
On 06/04/2014 17:52, Toralf Förster wrote: Reverting either one of them solves the problem reported with kvm, but revert is probably not the correct answer. I wonder if the solution is as simple as this: ---8--- diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index