#3885 closed defect (fixed)
Context switch extension is broken in SMP configurations
Reported by: | Sebastian Huber | Owned by: | Sebastian Huber |
---|---|---|---|
Priority: | normal | Milestone: | 5.1 |
Component: | score | Version: | 5 |
Severity: | normal | Keywords: | qualification |
Cc: | Blocked By: | ||
Blocking: |
Description (last modified by Sebastian Huber)
The context switch extensions are called during _Thread_Do_dispatch():
void _Thread_Do_dispatch( Per_CPU_Control *cpu_self, ISR_Level level ) { Thread_Control *executing; executing = cpu_self->executing; ... do { Thread_Control *heir; heir = _Thread_Get_heir_and_make_it_executing( cpu_self ); ... _User_extensions_Thread_switch( executing, heir ); ... _Context_Switch( &executing->Registers, &heir->Registers ); ... } while ( cpu_self->dispatch_necessary ); ... }
In uniprocessor configurations, the context switch extensions are called for all thread switches except the very first thread switch to the initialization thread. However, in SMP configurations, the context switch may be invalidated and updated in the low-level _Context_Switch() routine. See:
In case such an update happens, a thread executes on the processor which was not visible to the context switch extensions. This can confuse for example event record consumers which use events generated by a context switch extension.
Fixing this is not straight forward. The context switch extensions call must move after the low-level context switch. The problem here is that we may end up in _Thread_Handler(). Adding the context switch extensions call to _Thread_Handler() covers now also the thread switch to the initialization thread. We also have to save the last executing thread of the processor. Registers or the stack cannot be used for this purpose. We have to add it to the per-processor information. Existing extensions may be affected, since now context switch extensions use the stack of the heir thread.
Calling the thread switch extensions in the low level context switch is difficult since at this point an intermediate stack is used which is only large enough to enable servicing of interrupts.
Change History (12)
comment:1 Changed on 02/25/20 at 18:34:30 by Joel Sherrill
comment:2 follow-up: 3 Changed on 02/26/20 at 02:55:06 by Chris Johns
Is this an issue that needs to be fixed for RTEMS 5?
comment:3 Changed on 02/26/20 at 08:47:44 by Sebastian Huber
Replying to Chris Johns:
Is this an issue that needs to be fixed for RTEMS 5?
Yes, I think that the thread switch extensions should work in SMP configurations.
comment:4 follow-up: 6 Changed on 02/26/20 at 08:49:11 by Sebastian Huber
Moving the thread switch extension calls into the context of the heir thread breaks for example the stack checker. It checks that the current stack pointer is in the stack area of a thread.
comment:5 Changed on 02/26/20 at 08:51:14 by Sebastian Huber
Description: | modified (diff) |
---|
comment:6 follow-up: 7 Changed on 02/27/20 at 00:23:43 by Chris Johns
Replying to Sebastian Huber:
Moving the thread switch extension calls into the context of the heir thread breaks for example the stack checker. It checks that the current stack pointer is in the stack area of a thread.
Is a patch to fix the stack checker being worked on?
I know of applications that have this running during development.
comment:7 Changed on 02/27/20 at 06:58:32 by Sebastian Huber
Replying to Chris Johns:
Replying to Sebastian Huber:
Moving the thread switch extension calls into the context of the heir thread breaks for example the stack checker. It checks that the current stack pointer is in the stack area of a thread.
Is a patch to fix the stack checker being worked on?
I know of applications that have this running during development.
This was easy to fix:
https://lists.rtems.org/pipermail/devel/2020-February/057759.html
I just wanted to highlight that this change can easily break existing extensions.
comment:9 Changed on 03/02/20 at 06:52:07 by Sebastian Huber <sebastian.huber@…>
In [changeset:"fcb11510c6b38a7a1f4da7205acc5461a4e18214/rtems" fcb11510/rtems]:
comment:10 Changed on 03/03/20 at 06:30:26 by Sebastian Huber <sebastian.huber@…>
In [changeset:"198c07e5a58413bb2b16c01a15c545875d144ad5/rtems" 198c07e5/rtems]:
comment:11 Changed on 03/03/20 at 12:13:32 by Sebastian Huber <sebastian.huber@…>
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
In [changeset:"8bd4e6a55b131a1429a082a33b48b2f56a1963be/rtems-docs" 8bd4e6a/rtems-docs]:
comment:12 Changed on 06/23/21 at 07:07:55 by Sebastian Huber
Keywords: | qualification added |
---|
As I recall, doing this in _Thread_Handler is similar to having to restore FPU. Are there other actions with this characteristic?