Re: [bisected] system hang after boot

2017-11-27 Thread Peter Zijlstra
On Mon, Nov 27, 2017 at 01:10:31PM +, Will Deacon wrote: > break_lock should be annotated (at least) with READ_ONCE/WRITE_ONCE, which > should prevent that from happening. This code long predates all that new fangled stuff ;-)

Re: [bisected] system hang after boot

2017-11-27 Thread Peter Zijlstra
On Mon, Nov 27, 2017 at 02:00:28PM +0100, Martin Schwidefsky wrote: > I would opt for removing it entirely. Like said; I'm not opposed to that. I was just explaining where it came from (before my time might I add) and how its supposed to 'work' :-)

Re: [bisected] system hang after boot

2017-11-27 Thread Will Deacon
On Mon, Nov 27, 2017 at 02:00:28PM +0100, Martin Schwidefsky wrote: > On Mon, 27 Nov 2017 12:54:56 + > Will Deacon wrote: > > On Mon, Nov 27, 2017 at 01:49:18PM +0100, Martin Schwidefsky wrote: > > > On Mon, 27 Nov 2017 11:49:48 + > > > Will Deacon wrote: > > > > On Wed, Nov 22, 2017 at

Re: [bisected] system hang after boot

2017-11-27 Thread Sebastian Ott
On Mon, 27 Nov 2017, Will Deacon wrote: > Sebastian: could you try the diff below, please? If that fixes s390, then > we can debate the merits of GENERIC_LOCKBREAK independently of fixing this > issue. > > Thanks, > > Will > > --->8 > > diff --git a/kernel/locking/spinlock.c b/kernel/locking/sp

Re: [bisected] system hang after boot

2017-11-27 Thread Martin Schwidefsky
On Mon, 27 Nov 2017 12:54:56 + Will Deacon wrote: > Hi Martin, > > On Mon, Nov 27, 2017 at 01:49:18PM +0100, Martin Schwidefsky wrote: > > On Mon, 27 Nov 2017 11:49:48 + > > Will Deacon wrote: > > > On Wed, Nov 22, 2017 at 09:22:17PM +0100, Peter Zijlstra wrote: > > > > On Wed, Nov

Re: [bisected] system hang after boot

2017-11-27 Thread Will Deacon
Hi Martin, On Mon, Nov 27, 2017 at 01:49:18PM +0100, Martin Schwidefsky wrote: > On Mon, 27 Nov 2017 11:49:48 + > Will Deacon wrote: > > On Wed, Nov 22, 2017 at 09:22:17PM +0100, Peter Zijlstra wrote: > > > On Wed, Nov 22, 2017 at 06:26:59PM +, Will Deacon wrote: > > > > > > > Now, I c

Re: [bisected] system hang after boot

2017-11-27 Thread Martin Schwidefsky
On Mon, 27 Nov 2017 11:49:48 + Will Deacon wrote: > Hi Peter, > > On Wed, Nov 22, 2017 at 09:22:17PM +0100, Peter Zijlstra wrote: > > On Wed, Nov 22, 2017 at 06:26:59PM +, Will Deacon wrote: > > > > > Now, I can't see what the break_lock is doing here other than causing > > > problems

Re: [bisected] system hang after boot

2017-11-27 Thread Will Deacon
Me again... On Mon, Nov 27, 2017 at 11:49:47AM +, Will Deacon wrote: > On Wed, Nov 22, 2017 at 09:22:17PM +0100, Peter Zijlstra wrote: > > On Wed, Nov 22, 2017 at 06:26:59PM +, Will Deacon wrote: > > > > > Now, I can't see what the break_lock is doing here other than causing > > > problem

Re: [bisected] system hang after boot

2017-11-27 Thread Will Deacon
Hi Peter, On Wed, Nov 22, 2017 at 09:22:17PM +0100, Peter Zijlstra wrote: > On Wed, Nov 22, 2017 at 06:26:59PM +, Will Deacon wrote: > > > Now, I can't see what the break_lock is doing here other than causing > > problems. Is there a good reason for it, or can you just try removing it > > alt

Re: [bisected] system hang after boot

2017-11-22 Thread Peter Zijlstra
On Wed, Nov 22, 2017 at 06:26:59PM +, Will Deacon wrote: > Now, I can't see what the break_lock is doing here other than causing > problems. Is there a good reason for it, or can you just try removing it > altogether? Patch below. The main use is spin_is_contended(), which in turn ends up use

Re: [bisected] system hang after boot

2017-11-22 Thread Will Deacon
Hi Sebastian, On Wed, Nov 22, 2017 at 07:54:54PM +0100, Sebastian Ott wrote: > On Wed, 22 Nov 2017, Will Deacon wrote: > > Now, I can't see what the break_lock is doing here other than causing > > problems. Is there a good reason for it, or can you just try removing it > > altogether? Patch below.

Re: [bisected] system hang after boot

2017-11-22 Thread Sebastian Ott
Hello Will, On Wed, 22 Nov 2017, Will Deacon wrote: > Now, I can't see what the break_lock is doing here other than causing > problems. Is there a good reason for it, or can you just try removing it > altogether? Patch below. With your patch applied the system is able to boot again. I did some qu

Re: [bisected] system hang after boot

2017-11-22 Thread Will Deacon
Hi Sebastian, Thanks for the report. On Wed, Nov 22, 2017 at 06:46:01PM +0100, Sebastian Ott wrote: > One of my test systems (s390) hangs after boot. One of the cpus is idle > the other is in arch_spin_relax. Bisect pointed to this one: > > a8a217c22116eff6c120d753c9934089fb229af0 is the first b

[bisected] system hang after boot

2017-11-22 Thread Sebastian Ott
Hi, One of my test systems (s390) hangs after boot. One of the cpus is idle the other is in arch_spin_relax. Bisect pointed to this one: a8a217c22116eff6c120d753c9934089fb229af0 is the first bad commit commit a8a217c22116eff6c120d753c9934089fb229af0 Author: Will Deacon Date: Tue Oct 3 19:25:27