Re: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Joerg Roedel]

On Wed, Jul 26, 2017 at 03:25:05PM +0200, Artem Savkov wrote:
> On Wed, Jul 26, 2017 at 02:26:14PM +0200, Joerg Roedel wrote:
> > Yes, that should fix it, but I think its better to just move the
> > register_syscore_ops() call to a later initialization step, like in the
> > patch below. I tested it an will queue it to my iommu/fixes branch.
> 
> Checked it as well just in case, didn't see any issues. Thank you.
> 
> Reported-and-tested-by: Artem Savkov 

Thanks for testing it! I added your's and Thomas' tags and applied the
patch to my tree. It should go upstream this week.


Joerg

Re: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Joerg Roedel]

On Wed, Jul 26, 2017 at 03:25:05PM +0200, Artem Savkov wrote:
> On Wed, Jul 26, 2017 at 02:26:14PM +0200, Joerg Roedel wrote:
> > Yes, that should fix it, but I think its better to just move the
> > register_syscore_ops() call to a later initialization step, like in the
> > patch below. I tested it an will queue it to my iommu/fixes branch.
> 
> Checked it as well just in case, didn't see any issues. Thank you.
> 
> Reported-and-tested-by: Artem Savkov 

Thanks for testing it! I added your's and Thomas' tags and applied the
patch to my tree. It should go upstream this week.


Joerg

Re: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Artem Savkov]

On Wed, Jul 26, 2017 at 02:26:14PM +0200, Joerg Roedel wrote:
> Hi Artem, Thomas,
> 
> On Wed, Jul 26, 2017 at 12:42:49PM +0200, Thomas Gleixner wrote:
> > On Tue, 25 Jul 2017, Artem Savkov wrote:
> > 
> > > Hi,
> > > 
> > > Commit 1c3c5ea "sched/core: Enable might_sleep() and smp_processor_id()
> > > checks early" seem to have uncovered an issue with amd-iommu/x2apic.
> > > 
> > > Starting with that commit the following warning started to show up on AMD
> > > systems during boot:
> >  
> > > [0.16] BUG: sleeping function called from invalid context at 
> > > kernel/locking/mutex.c:747 
> > 
> > > [0.16]  mutex_lock_nested+0x1b/0x20 
> > > [0.16]  register_syscore_ops+0x1d/0x70 
> > > [0.16]  state_next+0x119/0x910 
> > > [0.16]  iommu_go_to_state+0x29/0x30 
> > > [0.16]  amd_iommu_enable+0x13/0x23 
> > > [0.16]  irq_remapping_enable+0x1b/0x39 
> > > [0.16]  enable_IR_x2apic+0x91/0x196 
> > > [0.16]  default_setup_apic_routing+0x16/0x6e 
> > > [0.16]  native_smp_prepare_cpus+0x257/0x2d5
> 
> Thanks for the report!
> 
> > --- a/drivers/iommu/amd_iommu_init.c
> > +++ b/drivers/iommu/amd_iommu_init.c
> > @@ -2440,7 +2440,6 @@ static int __init state_next(void)
> > break;
> > case IOMMU_ACPI_FINISHED:
> > early_enable_iommus();
> > -   register_syscore_ops(_iommu_syscore_ops);
> > x86_platform.iommu_shutdown = disable_iommus;
> > init_state = IOMMU_ENABLED;
> > break;
> > @@ -2559,6 +2558,8 @@ static int __init amd_iommu_init(void)
> > for_each_iommu(iommu)
> > iommu_flush_all_caches(iommu);
> > }
> > +   } else {
> > +   register_syscore_ops(_iommu_syscore_ops);
> > }
> >  
> > return ret;
> 
> Yes, that should fix it, but I think its better to just move the
> register_syscore_ops() call to a later initialization step, like in the
> patch below. I tested it an will queue it to my iommu/fixes branch.

Checked it as well just in case, didn't see any issues. Thank you.

Reported-and-tested-by: Artem Savkov 

-- 
Regards,
  Artem

Re: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Artem Savkov]

On Wed, Jul 26, 2017 at 02:26:14PM +0200, Joerg Roedel wrote:
> Hi Artem, Thomas,
> 
> On Wed, Jul 26, 2017 at 12:42:49PM +0200, Thomas Gleixner wrote:
> > On Tue, 25 Jul 2017, Artem Savkov wrote:
> > 
> > > Hi,
> > > 
> > > Commit 1c3c5ea "sched/core: Enable might_sleep() and smp_processor_id()
> > > checks early" seem to have uncovered an issue with amd-iommu/x2apic.
> > > 
> > > Starting with that commit the following warning started to show up on AMD
> > > systems during boot:
> >  
> > > [0.16] BUG: sleeping function called from invalid context at 
> > > kernel/locking/mutex.c:747 
> > 
> > > [0.16]  mutex_lock_nested+0x1b/0x20 
> > > [0.16]  register_syscore_ops+0x1d/0x70 
> > > [0.16]  state_next+0x119/0x910 
> > > [0.16]  iommu_go_to_state+0x29/0x30 
> > > [0.16]  amd_iommu_enable+0x13/0x23 
> > > [0.16]  irq_remapping_enable+0x1b/0x39 
> > > [0.16]  enable_IR_x2apic+0x91/0x196 
> > > [0.16]  default_setup_apic_routing+0x16/0x6e 
> > > [0.16]  native_smp_prepare_cpus+0x257/0x2d5
> 
> Thanks for the report!
> 
> > --- a/drivers/iommu/amd_iommu_init.c
> > +++ b/drivers/iommu/amd_iommu_init.c
> > @@ -2440,7 +2440,6 @@ static int __init state_next(void)
> > break;
> > case IOMMU_ACPI_FINISHED:
> > early_enable_iommus();
> > -   register_syscore_ops(_iommu_syscore_ops);
> > x86_platform.iommu_shutdown = disable_iommus;
> > init_state = IOMMU_ENABLED;
> > break;
> > @@ -2559,6 +2558,8 @@ static int __init amd_iommu_init(void)
> > for_each_iommu(iommu)
> > iommu_flush_all_caches(iommu);
> > }
> > +   } else {
> > +   register_syscore_ops(_iommu_syscore_ops);
> > }
> >  
> > return ret;
> 
> Yes, that should fix it, but I think its better to just move the
> register_syscore_ops() call to a later initialization step, like in the
> patch below. I tested it an will queue it to my iommu/fixes branch.

Checked it as well just in case, didn't see any issues. Thank you.

Reported-and-tested-by: Artem Savkov 

-- 
Regards,
  Artem

Re: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Thomas Gleixner]

On Wed, 26 Jul 2017, Joerg Roedel wrote:
> Yes, that should fix it, but I think its better to just move the
> register_syscore_ops() call to a later initialization step, like in the
> patch below. I tested it an will queue it to my iommu/fixes branch.

Fair enough. Acked-by-me.

Re: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Thomas Gleixner]

On Wed, 26 Jul 2017, Joerg Roedel wrote:
> Yes, that should fix it, but I think its better to just move the
> register_syscore_ops() call to a later initialization step, like in the
> patch below. I tested it an will queue it to my iommu/fixes branch.

Fair enough. Acked-by-me.

[PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Joerg Roedel]

Hi Artem, Thomas,

On Wed, Jul 26, 2017 at 12:42:49PM +0200, Thomas Gleixner wrote:
> On Tue, 25 Jul 2017, Artem Savkov wrote:
> 
> > Hi,
> > 
> > Commit 1c3c5ea "sched/core: Enable might_sleep() and smp_processor_id()
> > checks early" seem to have uncovered an issue with amd-iommu/x2apic.
> > 
> > Starting with that commit the following warning started to show up on AMD
> > systems during boot:
>  
> > [0.16] BUG: sleeping function called from invalid context at 
> > kernel/locking/mutex.c:747 
> 
> > [0.16]  mutex_lock_nested+0x1b/0x20 
> > [0.16]  register_syscore_ops+0x1d/0x70 
> > [0.16]  state_next+0x119/0x910 
> > [0.16]  iommu_go_to_state+0x29/0x30 
> > [0.16]  amd_iommu_enable+0x13/0x23 
> > [0.16]  irq_remapping_enable+0x1b/0x39 
> > [0.16]  enable_IR_x2apic+0x91/0x196 
> > [0.16]  default_setup_apic_routing+0x16/0x6e 
> > [0.16]  native_smp_prepare_cpus+0x257/0x2d5

Thanks for the report!

> --- a/drivers/iommu/amd_iommu_init.c
> +++ b/drivers/iommu/amd_iommu_init.c
> @@ -2440,7 +2440,6 @@ static int __init state_next(void)
>   break;
>   case IOMMU_ACPI_FINISHED:
>   early_enable_iommus();
> - register_syscore_ops(_iommu_syscore_ops);
>   x86_platform.iommu_shutdown = disable_iommus;
>   init_state = IOMMU_ENABLED;
>   break;
> @@ -2559,6 +2558,8 @@ static int __init amd_iommu_init(void)
>   for_each_iommu(iommu)
>   iommu_flush_all_caches(iommu);
>   }
> + } else {
> + register_syscore_ops(_iommu_syscore_ops);
>   }
>  
>   return ret;

Yes, that should fix it, but I think its better to just move the
register_syscore_ops() call to a later initialization step, like in the
patch below. I tested it an will queue it to my iommu/fixes branch.

>From 461242d7211c901b6ccdf349cc89235bd5da Mon Sep 17 00:00:00 2001
From: Joerg Roedel <jroe...@suse.de>
Date: Wed, 26 Jul 2017 14:17:55 +0200
Subject: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization
 code

The register_syscore_ops() function takes a mutex and might
sleep. In the IOMMU initialization code it is invoked during
irq-remapping setup already, where irqs are disabled.

This causes a schedule-while-atomic bug:

 BUG: sleeping function called from invalid context at 
kernel/locking/mutex.c:747
 in_atomic(): 0, irqs_disabled(): 1, pid: 1, name: swapper/0
 no locks held by swapper/0/1.
 irq event stamp: 304
 hardirqs last  enabled at (303): [] 
_raw_spin_unlock_irqrestore+0x36/0x60
 hardirqs last disabled at (304): [] 
enable_IR_x2apic+0x79/0x196
 softirqs last  enabled at (36): [] __do_softirq+0x35f/0x4ec
 softirqs last disabled at (31): [] irq_exit+0x105/0x120
 CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.13.0-rc2.1.el7a.test.x86_64.debug 
#1
 Hardware name:  PowerEdge C6145 /040N24, BIOS 3.5.0 10/28/2014
 Call Trace:
  dump_stack+0x85/0xca
  ___might_sleep+0x22a/0x260
  __might_sleep+0x4a/0x80
  __mutex_lock+0x58/0x960
  ? iommu_completion_wait.part.17+0xb5/0x160
  ? register_syscore_ops+0x1d/0x70
  ? iommu_flush_all_caches+0x120/0x150
  mutex_lock_nested+0x1b/0x20
  register_syscore_ops+0x1d/0x70
  state_next+0x119/0x910
  iommu_go_to_state+0x29/0x30
  amd_iommu_enable+0x13/0x23

Fix it by moving the register_syscore_ops() call to the next
initialization step, which runs with irqs enabled.

Signed-off-by: Joerg Roedel <jroe...@suse.de>
---
 drivers/iommu/amd_iommu_init.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index 5cc597b383c7..372303700566 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -2440,11 +2440,11 @@ static int __init state_next(void)
break;
case IOMMU_ACPI_FINISHED:
early_enable_iommus();
-   register_syscore_ops(_iommu_syscore_ops);
x86_platform.iommu_shutdown = disable_iommus;
init_state = IOMMU_ENABLED;
break;
case IOMMU_ENABLED:
+   register_syscore_ops(_iommu_syscore_ops);
ret = amd_iommu_init_pci();
init_state = ret ? IOMMU_INIT_ERROR : IOMMU_PCI_INIT;
enable_iommus_v2();
-- 
2.13.1

[PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization code

[Joerg Roedel]

Hi Artem, Thomas,

On Wed, Jul 26, 2017 at 12:42:49PM +0200, Thomas Gleixner wrote:
> On Tue, 25 Jul 2017, Artem Savkov wrote:
> 
> > Hi,
> > 
> > Commit 1c3c5ea "sched/core: Enable might_sleep() and smp_processor_id()
> > checks early" seem to have uncovered an issue with amd-iommu/x2apic.
> > 
> > Starting with that commit the following warning started to show up on AMD
> > systems during boot:
>  
> > [0.16] BUG: sleeping function called from invalid context at 
> > kernel/locking/mutex.c:747 
> 
> > [0.16]  mutex_lock_nested+0x1b/0x20 
> > [0.16]  register_syscore_ops+0x1d/0x70 
> > [0.16]  state_next+0x119/0x910 
> > [0.16]  iommu_go_to_state+0x29/0x30 
> > [0.16]  amd_iommu_enable+0x13/0x23 
> > [0.16]  irq_remapping_enable+0x1b/0x39 
> > [0.16]  enable_IR_x2apic+0x91/0x196 
> > [0.16]  default_setup_apic_routing+0x16/0x6e 
> > [0.16]  native_smp_prepare_cpus+0x257/0x2d5

Thanks for the report!

> --- a/drivers/iommu/amd_iommu_init.c
> +++ b/drivers/iommu/amd_iommu_init.c
> @@ -2440,7 +2440,6 @@ static int __init state_next(void)
>   break;
>   case IOMMU_ACPI_FINISHED:
>   early_enable_iommus();
> - register_syscore_ops(_iommu_syscore_ops);
>   x86_platform.iommu_shutdown = disable_iommus;
>   init_state = IOMMU_ENABLED;
>   break;
> @@ -2559,6 +2558,8 @@ static int __init amd_iommu_init(void)
>   for_each_iommu(iommu)
>   iommu_flush_all_caches(iommu);
>   }
> + } else {
> + register_syscore_ops(_iommu_syscore_ops);
>   }
>  
>   return ret;

Yes, that should fix it, but I think its better to just move the
register_syscore_ops() call to a later initialization step, like in the
patch below. I tested it an will queue it to my iommu/fixes branch.

>From 461242d7211c7777901b6ccdf349cc89235bd5da Mon Sep 17 00:00:00 2001
From: Joerg Roedel 
Date: Wed, 26 Jul 2017 14:17:55 +0200
Subject: [PATCH] iommu/amd: Fix schedule-while-atomic BUG in initialization
 code

The register_syscore_ops() function takes a mutex and might
sleep. In the IOMMU initialization code it is invoked during
irq-remapping setup already, where irqs are disabled.

This causes a schedule-while-atomic bug:

 BUG: sleeping function called from invalid context at 
kernel/locking/mutex.c:747
 in_atomic(): 0, irqs_disabled(): 1, pid: 1, name: swapper/0
 no locks held by swapper/0/1.
 irq event stamp: 304
 hardirqs last  enabled at (303): [] 
_raw_spin_unlock_irqrestore+0x36/0x60
 hardirqs last disabled at (304): [] 
enable_IR_x2apic+0x79/0x196
 softirqs last  enabled at (36): [] __do_softirq+0x35f/0x4ec
 softirqs last disabled at (31): [] irq_exit+0x105/0x120
 CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.13.0-rc2.1.el7a.test.x86_64.debug 
#1
 Hardware name:  PowerEdge C6145 /040N24, BIOS 3.5.0 10/28/2014
 Call Trace:
  dump_stack+0x85/0xca
  ___might_sleep+0x22a/0x260
  __might_sleep+0x4a/0x80
  __mutex_lock+0x58/0x960
  ? iommu_completion_wait.part.17+0xb5/0x160
  ? register_syscore_ops+0x1d/0x70
  ? iommu_flush_all_caches+0x120/0x150
  mutex_lock_nested+0x1b/0x20
  register_syscore_ops+0x1d/0x70
  state_next+0x119/0x910
  iommu_go_to_state+0x29/0x30
  amd_iommu_enable+0x13/0x23

Fix it by moving the register_syscore_ops() call to the next
initialization step, which runs with irqs enabled.

Signed-off-by: Joerg Roedel 
---
 drivers/iommu/amd_iommu_init.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index 5cc597b383c7..372303700566 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -2440,11 +2440,11 @@ static int __init state_next(void)
break;
case IOMMU_ACPI_FINISHED:
early_enable_iommus();
-   register_syscore_ops(_iommu_syscore_ops);
x86_platform.iommu_shutdown = disable_iommus;
init_state = IOMMU_ENABLED;
break;
case IOMMU_ENABLED:
+   register_syscore_ops(_iommu_syscore_ops);
ret = amd_iommu_init_pci();
init_state = ret ? IOMMU_INIT_ERROR : IOMMU_PCI_INIT;
enable_iommus_v2();
-- 
2.13.1