[PATCH] runstate: ignore finishmigrate -> prelaunch transition

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

[PATCH] runstate: ignore finishmigrate -> prelaunch transition

Laurent Vivier-5
Commit 1bd71dce4bf2 tries to prevent a finishmigrate -> prelaunch
transition by exiting at the beginning of the main_loop_should_exit()
function if the state is already finishmigrate.

As the finishmigrate state is set in the migration thread it can
happen concurrently to the function. The migration thread and the
function are normally protected by the iothread mutex and thus the
state should no evolve between the start of the function and its end.

Unfortunately during the function life the lock is released by
pause_all_vcpus() just before the point we need to be sure we are
not in finishmigrate state and if the migration thread is waiting
for the lock it will take the opportunity to change the state
to finishmigrate.

The only way to be sure we are not in the finishmigrate state when
we need is to check the state after the pause_all_vcpus() function.

Fixes: 1bd71dce4bf2 ("runstate: ignore exit request in finish migrate state")
Signed-off-by: Laurent Vivier <[hidden email]>
---
 vl.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/vl.c b/vl.c
index 6a65a64bfd64..bf0a6345d239 100644
--- a/vl.c
+++ b/vl.c
@@ -1745,9 +1745,6 @@ static bool main_loop_should_exit(void)
     RunState r;
     ShutdownCause request;
 
-    if (runstate_check(RUN_STATE_FINISH_MIGRATE)) {
-        return false;
-    }
     if (preconfig_exit_requested) {
         if (runstate_check(RUN_STATE_PRECONFIG)) {
             runstate_set(RUN_STATE_PRELAUNCH);
@@ -1776,8 +1773,13 @@ static bool main_loop_should_exit(void)
         pause_all_vcpus();
         qemu_system_reset(request);
         resume_all_vcpus();
+        /*
+         * runstate can change in pause_all_vcpus()
+         * as iothread mutex is unlocked
+         */
         if (!runstate_check(RUN_STATE_RUNNING) &&
-                !runstate_check(RUN_STATE_INMIGRATE)) {
+                !runstate_check(RUN_STATE_INMIGRATE) &&
+                !runstate_check(RUN_STATE_FINISH_MIGRATE)) {
             runstate_set(RUN_STATE_PRELAUNCH);
         }
     }
--
2.23.0


Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] runstate: ignore finishmigrate -> prelaunch transition

Dr. David Alan Gilbert (git)
* Laurent Vivier ([hidden email]) wrote:

> Commit 1bd71dce4bf2 tries to prevent a finishmigrate -> prelaunch
> transition by exiting at the beginning of the main_loop_should_exit()
> function if the state is already finishmigrate.
>
> As the finishmigrate state is set in the migration thread it can
> happen concurrently to the function. The migration thread and the
> function are normally protected by the iothread mutex and thus the
> state should no evolve between the start of the function and its end.
>
> Unfortunately during the function life the lock is released by
> pause_all_vcpus() just before the point we need to be sure we are
> not in finishmigrate state and if the migration thread is waiting
> for the lock it will take the opportunity to change the state
> to finishmigrate.

Ewww.
I hate those short wakeups for pause_all_vcpus; I'm sure there are loads
more corners that break.

Still, I _think_ this is an improvement, so:

Reviewed-by: Dr. David Alan Gilbert <[hidden email]>

> The only way to be sure we are not in the finishmigrate state when
> we need is to check the state after the pause_all_vcpus() function.
>
> Fixes: 1bd71dce4bf2 ("runstate: ignore exit request in finish migrate state")
> Signed-off-by: Laurent Vivier <[hidden email]>
> ---
>  vl.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/vl.c b/vl.c
> index 6a65a64bfd64..bf0a6345d239 100644
> --- a/vl.c
> +++ b/vl.c
> @@ -1745,9 +1745,6 @@ static bool main_loop_should_exit(void)
>      RunState r;
>      ShutdownCause request;
>  
> -    if (runstate_check(RUN_STATE_FINISH_MIGRATE)) {
> -        return false;
> -    }
>      if (preconfig_exit_requested) {
>          if (runstate_check(RUN_STATE_PRECONFIG)) {
>              runstate_set(RUN_STATE_PRELAUNCH);
> @@ -1776,8 +1773,13 @@ static bool main_loop_should_exit(void)
>          pause_all_vcpus();
>          qemu_system_reset(request);
>          resume_all_vcpus();
> +        /*
> +         * runstate can change in pause_all_vcpus()
> +         * as iothread mutex is unlocked
> +         */
>          if (!runstate_check(RUN_STATE_RUNNING) &&
> -                !runstate_check(RUN_STATE_INMIGRATE)) {
> +                !runstate_check(RUN_STATE_INMIGRATE) &&
> +                !runstate_check(RUN_STATE_FINISH_MIGRATE)) {
>              runstate_set(RUN_STATE_PRELAUNCH);
>          }
>      }
> --
> 2.23.0
>
--
Dr. David Alan Gilbert / [hidden email] / Manchester, UK


Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] runstate: ignore finishmigrate -> prelaunch transition

Laurent Vivier-5
On 06/12/2019 20:52, Dr. David Alan Gilbert wrote:

> * Laurent Vivier ([hidden email]) wrote:
>> Commit 1bd71dce4bf2 tries to prevent a finishmigrate -> prelaunch
>> transition by exiting at the beginning of the main_loop_should_exit()
>> function if the state is already finishmigrate.
>>
>> As the finishmigrate state is set in the migration thread it can
>> happen concurrently to the function. The migration thread and the
>> function are normally protected by the iothread mutex and thus the
>> state should no evolve between the start of the function and its end.
>>
>> Unfortunately during the function life the lock is released by
>> pause_all_vcpus() just before the point we need to be sure we are
>> not in finishmigrate state and if the migration thread is waiting
>> for the lock it will take the opportunity to change the state
>> to finishmigrate.
>
> Ewww.
> I hate those short wakeups for pause_all_vcpus; I'm sure there are loads
> more corners that break.
>
> Still, I _think_ this is an improvement, so:
>
> Reviewed-by: Dr. David Alan Gilbert <[hidden email]>
>

Who volunteers to take this in his queue?

Thanks,
Laurent


Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] runstate: ignore finishmigrate -> prelaunch transition

Laurent Vivier-5
Ping?

Thanks,
Laurent

On 12/12/2019 20:40, Laurent Vivier wrote:

> On 06/12/2019 20:52, Dr. David Alan Gilbert wrote:
>> * Laurent Vivier ([hidden email]) wrote:
>>> Commit 1bd71dce4bf2 tries to prevent a finishmigrate -> prelaunch
>>> transition by exiting at the beginning of the main_loop_should_exit()
>>> function if the state is already finishmigrate.
>>>
>>> As the finishmigrate state is set in the migration thread it can
>>> happen concurrently to the function. The migration thread and the
>>> function are normally protected by the iothread mutex and thus the
>>> state should no evolve between the start of the function and its end.
>>>
>>> Unfortunately during the function life the lock is released by
>>> pause_all_vcpus() just before the point we need to be sure we are
>>> not in finishmigrate state and if the migration thread is waiting
>>> for the lock it will take the opportunity to change the state
>>> to finishmigrate.
>>
>> Ewww.
>> I hate those short wakeups for pause_all_vcpus; I'm sure there are loads
>> more corners that break.
>>
>> Still, I _think_ this is an improvement, so:
>>
>> Reviewed-by: Dr. David Alan Gilbert <[hidden email]>
>>
>
> Who volunteers to take this in his queue?
>
> Thanks,
> Laurent
>
>


Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] runstate: ignore finishmigrate -> prelaunch transition

Dr. David Alan Gilbert (git)
* Laurent Vivier ([hidden email]) wrote:
> Ping?

Juan says he's doing a migration pull soon - this seems like a good one
for it.

Dave

> Thanks,
> Laurent
>
> On 12/12/2019 20:40, Laurent Vivier wrote:
> > On 06/12/2019 20:52, Dr. David Alan Gilbert wrote:
> >> * Laurent Vivier ([hidden email]) wrote:
> >>> Commit 1bd71dce4bf2 tries to prevent a finishmigrate -> prelaunch
> >>> transition by exiting at the beginning of the main_loop_should_exit()
> >>> function if the state is already finishmigrate.
> >>>
> >>> As the finishmigrate state is set in the migration thread it can
> >>> happen concurrently to the function. The migration thread and the
> >>> function are normally protected by the iothread mutex and thus the
> >>> state should no evolve between the start of the function and its end.
> >>>
> >>> Unfortunately during the function life the lock is released by
> >>> pause_all_vcpus() just before the point we need to be sure we are
> >>> not in finishmigrate state and if the migration thread is waiting
> >>> for the lock it will take the opportunity to change the state
> >>> to finishmigrate.
> >>
> >> Ewww.
> >> I hate those short wakeups for pause_all_vcpus; I'm sure there are loads
> >> more corners that break.
> >>
> >> Still, I _think_ this is an improvement, so:
> >>
> >> Reviewed-by: Dr. David Alan Gilbert <[hidden email]>
> >>
> >
> > Who volunteers to take this in his queue?
> >
> > Thanks,
> > Laurent
> >
> >
>
--
Dr. David Alan Gilbert / [hidden email] / Manchester, UK


Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] runstate: ignore finishmigrate -> prelaunch transition

Juan Quintela
In reply to this post by Laurent Vivier-5
Laurent Vivier <[hidden email]> wrote:

> Ping?
>
> Thanks,
> Laurent
>
> On 12/12/2019 20:40, Laurent Vivier wrote:
>> On 06/12/2019 20:52, Dr. David Alan Gilbert wrote:
>>> * Laurent Vivier ([hidden email]) wrote:
>>>> Commit 1bd71dce4bf2 tries to prevent a finishmigrate -> prelaunch
>>>> transition by exiting at the beginning of the main_loop_should_exit()
>>>> function if the state is already finishmigrate.
>>>>
>>>> As the finishmigrate state is set in the migration thread it can
>>>> happen concurrently to the function. The migration thread and the
>>>> function are normally protected by the iothread mutex and thus the
>>>> state should no evolve between the start of the function and its end.
>>>>
>>>> Unfortunately during the function life the lock is released by
>>>> pause_all_vcpus() just before the point we need to be sure we are
>>>> not in finishmigrate state and if the migration thread is waiting
>>>> for the lock it will take the opportunity to change the state
>>>> to finishmigrate.
>>>
>>> Ewww.
>>> I hate those short wakeups for pause_all_vcpus; I'm sure there are loads
>>> more corners that break.
>>>
>>> Still, I _think_ this is an improvement, so:
>>>
>>> Reviewed-by: Dr. David Alan Gilbert <[hidden email]>
>>>
>>
>> Who volunteers to take this in his queue?

Reviewed-by: Juan Quintela <[hidden email]>

queued