* [1/1] s390x/smp: fix detection of "running"
@ 2020-03-27 16:33 Christian Borntraeger
2020-03-27 16:34 ` Christian Borntraeger
2020-03-27 16:55 ` David Hildenbrand
0 siblings, 2 replies; 6+ messages in thread
From: Christian Borntraeger @ 2020-03-27 16:33 UTC (permalink / raw)
To: David Hildenbrand, Janosch Frank
Cc: Cornelia Huck, kvm, Christian Borntraeger
On s390x hosts with a single CPU, the smp test case hangs (loops).
The check is our restart has finished is wrong.
Sigp sense running status checks if the CPU is currently backed by a
real CPU. This means that on single CPU hosts a sigp sense running
will never claim that a target is running. We need to check for not
being stopped instead.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
lib/s390x/smp.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/s390x/smp.c b/lib/s390x/smp.c
index 2555bf4..5ed8b7b 100644
--- a/lib/s390x/smp.c
+++ b/lib/s390x/smp.c
@@ -128,7 +128,7 @@ static int smp_cpu_restart_nolock(uint16_t addr, struct psw *psw)
* The order has been accepted, but the actual restart may not
* have been performed yet, so wait until the cpu is running.
*/
- while (!smp_cpu_running(addr))
+ while (smp_cpu_stopped(addr))
mb();
cpu->active = true;
return 0;
--
2.25.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [1/1] s390x/smp: fix detection of "running"
2020-03-27 16:33 [1/1] s390x/smp: fix detection of "running" Christian Borntraeger
@ 2020-03-27 16:34 ` Christian Borntraeger
2020-03-27 16:55 ` David Hildenbrand
1 sibling, 0 replies; 6+ messages in thread
From: Christian Borntraeger @ 2020-03-27 16:34 UTC (permalink / raw)
To: David Hildenbrand, Janosch Frank; +Cc: Cornelia Huck, kvm
[1/1] was supposed to be kvm-unit-tests....
On 27.03.20 17:33, Christian Borntraeger wrote:
> On s390x hosts with a single CPU, the smp test case hangs (loops).
> The check is our restart has finished is wrong.
> Sigp sense running status checks if the CPU is currently backed by a
> real CPU. This means that on single CPU hosts a sigp sense running
> will never claim that a target is running. We need to check for not
> being stopped instead.
>
> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
> ---
> lib/s390x/smp.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/s390x/smp.c b/lib/s390x/smp.c
> index 2555bf4..5ed8b7b 100644
> --- a/lib/s390x/smp.c
> +++ b/lib/s390x/smp.c
> @@ -128,7 +128,7 @@ static int smp_cpu_restart_nolock(uint16_t addr, struct psw *psw)
> * The order has been accepted, but the actual restart may not
> * have been performed yet, so wait until the cpu is running.
> */
> - while (!smp_cpu_running(addr))
> + while (smp_cpu_stopped(addr))
> mb();
> cpu->active = true;
> return 0;
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [1/1] s390x/smp: fix detection of "running"
2020-03-27 16:33 [1/1] s390x/smp: fix detection of "running" Christian Borntraeger
2020-03-27 16:34 ` Christian Borntraeger
@ 2020-03-27 16:55 ` David Hildenbrand
2020-03-27 18:01 ` Christian Borntraeger
1 sibling, 1 reply; 6+ messages in thread
From: David Hildenbrand @ 2020-03-27 16:55 UTC (permalink / raw)
To: Christian Borntraeger, Janosch Frank; +Cc: Cornelia Huck, kvm
On 27.03.20 17:33, Christian Borntraeger wrote:
> On s390x hosts with a single CPU, the smp test case hangs (loops).
> The check is our restart has finished is wrong.
> Sigp sense running status checks if the CPU is currently backed by a
> real CPU. This means that on single CPU hosts a sigp sense running
> will never claim that a target is running. We need to check for not
> being stopped instead.
>
> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
> ---
> lib/s390x/smp.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/s390x/smp.c b/lib/s390x/smp.c
> index 2555bf4..5ed8b7b 100644
> --- a/lib/s390x/smp.c
> +++ b/lib/s390x/smp.c
> @@ -128,7 +128,7 @@ static int smp_cpu_restart_nolock(uint16_t addr, struct psw *psw)
> * The order has been accepted, but the actual restart may not
> * have been performed yet, so wait until the cpu is running.
> */
> - while (!smp_cpu_running(addr))
> + while (smp_cpu_stopped(addr))
> mb();
> cpu->active = true;
> return 0;
>
Yeah, same as the other issue we fixed before. There is no trusting on
SIGP SENSE RUNNING STATUS.
Can you please get rid of smp_cpu_running()? (looks like this was the
last user)
--
Thanks,
David / dhildenb
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [1/1] s390x/smp: fix detection of "running"
2020-03-27 16:55 ` David Hildenbrand
@ 2020-03-27 18:01 ` Christian Borntraeger
2020-03-27 18:05 ` David Hildenbrand
2020-03-30 7:37 ` Janosch Frank
0 siblings, 2 replies; 6+ messages in thread
From: Christian Borntraeger @ 2020-03-27 18:01 UTC (permalink / raw)
To: David Hildenbrand, Janosch Frank; +Cc: Cornelia Huck, kvm
On 27.03.20 17:55, David Hildenbrand wrote:
> On 27.03.20 17:33, Christian Borntraeger wrote:
>> On s390x hosts with a single CPU, the smp test case hangs (loops).
>> The check is our restart has finished is wrong.
>> Sigp sense running status checks if the CPU is currently backed by a
>> real CPU. This means that on single CPU hosts a sigp sense running
>> will never claim that a target is running. We need to check for not
>> being stopped instead.
>>
>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
>> ---
>> lib/s390x/smp.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/lib/s390x/smp.c b/lib/s390x/smp.c
>> index 2555bf4..5ed8b7b 100644
>> --- a/lib/s390x/smp.c
>> +++ b/lib/s390x/smp.c
>> @@ -128,7 +128,7 @@ static int smp_cpu_restart_nolock(uint16_t addr, struct psw *psw)
>> * The order has been accepted, but the actual restart may not
>> * have been performed yet, so wait until the cpu is running.
>> */
>> - while (!smp_cpu_running(addr))
>> + while (smp_cpu_stopped(addr))
>> mb();
>> cpu->active = true;
>> return 0;
>>
>
> Yeah, same as the other issue we fixed before. There is no trusting on
> SIGP SENSE RUNNING STATUS.
>
> Can you please get rid of smp_cpu_running()? (looks like this was the
> last user)
I think we should keep it and rename it to smp_cpu_running_status. This bug
actually showed that we should be able to test this feature (which is used
for spinlocks in Linux) somehow.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [1/1] s390x/smp: fix detection of "running"
2020-03-27 18:01 ` Christian Borntraeger
@ 2020-03-27 18:05 ` David Hildenbrand
2020-03-30 7:37 ` Janosch Frank
1 sibling, 0 replies; 6+ messages in thread
From: David Hildenbrand @ 2020-03-27 18:05 UTC (permalink / raw)
To: Christian Borntraeger, Janosch Frank; +Cc: Cornelia Huck, kvm
On 27.03.20 19:01, Christian Borntraeger wrote:
>
>
> On 27.03.20 17:55, David Hildenbrand wrote:
>> On 27.03.20 17:33, Christian Borntraeger wrote:
>>> On s390x hosts with a single CPU, the smp test case hangs (loops).
>>> The check is our restart has finished is wrong.
>>> Sigp sense running status checks if the CPU is currently backed by a
>>> real CPU. This means that on single CPU hosts a sigp sense running
>>> will never claim that a target is running. We need to check for not
>>> being stopped instead.
>>>
>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
>>> ---
>>> lib/s390x/smp.c | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/lib/s390x/smp.c b/lib/s390x/smp.c
>>> index 2555bf4..5ed8b7b 100644
>>> --- a/lib/s390x/smp.c
>>> +++ b/lib/s390x/smp.c
>>> @@ -128,7 +128,7 @@ static int smp_cpu_restart_nolock(uint16_t addr, struct psw *psw)
>>> * The order has been accepted, but the actual restart may not
>>> * have been performed yet, so wait until the cpu is running.
>>> */
>>> - while (!smp_cpu_running(addr))
>>> + while (smp_cpu_stopped(addr))
>>> mb();
>>> cpu->active = true;
>>> return 0;
>>>
>>
>> Yeah, same as the other issue we fixed before. There is no trusting on
>> SIGP SENSE RUNNING STATUS.
>>
>> Can you please get rid of smp_cpu_running()? (looks like this was the
>> last user)
>
> I think we should keep it and rename it to smp_cpu_running_status. This bug
> actually showed that we should be able to test this feature (which is used
> for spinlocks in Linux) somehow.
AFAIK, there is no trusting on SIGP SENSE RUNNING STATUS at all (I
discussed this with Janosch back then). And I don't see a way for a
reasonable test either. But if you have plans to add a test, then yes,
we can keep it.
--
Thanks,
David / dhildenb
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [1/1] s390x/smp: fix detection of "running"
2020-03-27 18:01 ` Christian Borntraeger
2020-03-27 18:05 ` David Hildenbrand
@ 2020-03-30 7:37 ` Janosch Frank
1 sibling, 0 replies; 6+ messages in thread
From: Janosch Frank @ 2020-03-30 7:37 UTC (permalink / raw)
To: Christian Borntraeger, David Hildenbrand; +Cc: Cornelia Huck, kvm
[-- Attachment #1.1: Type: text/plain, Size: 1836 bytes --]
On 3/27/20 7:01 PM, Christian Borntraeger wrote:
>
>
> On 27.03.20 17:55, David Hildenbrand wrote:
>> On 27.03.20 17:33, Christian Borntraeger wrote:
>>> On s390x hosts with a single CPU, the smp test case hangs (loops).
>>> The check is our restart has finished is wrong.
>>> Sigp sense running status checks if the CPU is currently backed by a
>>> real CPU. This means that on single CPU hosts a sigp sense running
>>> will never claim that a target is running. We need to check for not
>>> being stopped instead.
>>>
>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
>>> ---
>>> lib/s390x/smp.c | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/lib/s390x/smp.c b/lib/s390x/smp.c
>>> index 2555bf4..5ed8b7b 100644
>>> --- a/lib/s390x/smp.c
>>> +++ b/lib/s390x/smp.c
>>> @@ -128,7 +128,7 @@ static int smp_cpu_restart_nolock(uint16_t addr, struct psw *psw)
>>> * The order has been accepted, but the actual restart may not
>>> * have been performed yet, so wait until the cpu is running.
>>> */
>>> - while (!smp_cpu_running(addr))
>>> + while (smp_cpu_stopped(addr))
>>> mb();
>>> cpu->active = true;
>>> return 0;
>>>
>>
>> Yeah, same as the other issue we fixed before. There is no trusting on
>> SIGP SENSE RUNNING STATUS.
>>
>> Can you please get rid of smp_cpu_running()? (looks like this was the
>> last user)
>
> I think we should keep it and rename it to smp_cpu_running_status. This bug
> actually showed that we should be able to test this feature (which is used
> for spinlocks in Linux) somehow.
>
We need to move it to the smp test case and at least test it but not use
it for any library stuff. When I do firmware testing with the unit tests
I need to be able to exercise as much sigp orders as possible.
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-03-30 7:37 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-03-27 16:33 [1/1] s390x/smp: fix detection of "running" Christian Borntraeger
2020-03-27 16:34 ` Christian Borntraeger
2020-03-27 16:55 ` David Hildenbrand
2020-03-27 18:01 ` Christian Borntraeger
2020-03-27 18:05 ` David Hildenbrand
2020-03-30 7:37 ` Janosch Frank
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.