diff mbox series

[kvm-unit-tests,PULL,1/2] powerpc: Allow for a custom decr value to be specified to load on decr excp

Message ID 20190517130305.32123-2-lvivier@redhat.com (mailing list archive)
State New, archived
Headers show
Series [kvm-unit-tests,PULL,1/2] powerpc: Allow for a custom decr value to be specified to load on decr excp | expand

Commit Message

Laurent Vivier May 17, 2019, 1:03 p.m. UTC 
From: Suraj Jitindar Singh <sjitindarsingh@gmail.com>

Currently the handler for a decrementer exception will simply reload the
maximum value (0x7FFFFFFF), which will take ~4 seconds to expire again.
This means that if a vcpu cedes, it will be ~4 seconds between wakeups.

The h_cede_tm test is testing a known breakage when a guest cedes while
suspended. To be sure we cede 500 times to check for the bug. However
since it takes ~4 seconds to be woken up once we've ceded, we only get
through ~20 iterations before we reach the 90 seconds timeout and the
test appears to fail.

Add an option when registering the decrementer handler to specify the
value which should be reloaded by the handler, allowing the timeout to be
chosen.

Modify the spr test to use the max timeout to preserve existing
behaviour.
Modify the h_cede_tm test to use a 10ms timeout to ensure we can perform
500 iterations before hitting the 90 second time limit for a test.

This means the h_cede_tm test now succeeds rather than timing out.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
[lv: reset initial value to 0x3FFFFFFF]
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
---
 lib/powerpc/handlers.c | 7 ++++---
 powerpc/sprs.c         | 3 ++-
 powerpc/tm.c           | 4 +++-
 3 files changed, 9 insertions(+), 5 deletions(-)

Comments

Thomas Huth May 17, 2019, 1:10 p.m. UTC | #1 
On 17/05/2019 15.03, Laurent Vivier wrote:
> From: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
> 
> Currently the handler for a decrementer exception will simply reload the
> maximum value (0x7FFFFFFF), which will take ~4 seconds to expire again.
> This means that if a vcpu cedes, it will be ~4 seconds between wakeups.
> 
> The h_cede_tm test is testing a known breakage when a guest cedes while
> suspended. To be sure we cede 500 times to check for the bug. However
> since it takes ~4 seconds to be woken up once we've ceded, we only get
> through ~20 iterations before we reach the 90 seconds timeout and the
> test appears to fail.
> 
> Add an option when registering the decrementer handler to specify the
> value which should be reloaded by the handler, allowing the timeout to be
> chosen.
> 
> Modify the spr test to use the max timeout to preserve existing
> behaviour.
> Modify the h_cede_tm test to use a 10ms timeout to ensure we can perform
> 500 iterations before hitting the 90 second time limit for a test.
> 
> This means the h_cede_tm test now succeeds rather than timing out.
> 
> Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
> Reviewed-by: Thomas Huth <thuth@redhat.com>
> Reviewed-by: Laurent Vivier <lvivier@redhat.com>
> [lv: reset initial value to 0x3FFFFFFF]

Looks like something went wrong here? There is still the 0x7FFFFFFF in
the hunk below...

> diff --git a/powerpc/sprs.c b/powerpc/sprs.c
> index 6744bd8d8049..3c2d98c9ca99 100644
> --- a/powerpc/sprs.c
> +++ b/powerpc/sprs.c
> @@ -253,6 +253,7 @@ int main(int argc, char **argv)
>  		0x1234567890ABCDEFULL, 0xFEDCBA0987654321ULL,
>  		-1ULL,
>  	};
> +	static uint64_t decr = 0x7FFFFFFF; /* Max value */
>  
>  	for (i = 1; i < argc; i++) {
>  		if (!strcmp(argv[i], "-w")) {
> @@ -288,7 +289,7 @@ int main(int argc, char **argv)
>  		(void) getchar();
>  	} else {
>  		puts("Sleeping...\n");
> -		handle_exception(0x900, &dec_except_handler, NULL);
> +		handle_exception(0x900, &dec_except_handler, &decr);
>  		asm volatile ("mtdec %0" : : "r" (0x3FFFFFFF));
>  		hcall(H_CEDE);
>  	}

 Thomas
Suraj Jitindar Singh May 20, 2019, 4:47 a.m. UTC | #2 
On Fri, 2019-05-17 at 15:10 +0200, Thomas Huth wrote:
> On 17/05/2019 15.03, Laurent Vivier wrote:
> > From: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
> > 
> > Currently the handler for a decrementer exception will simply
> > reload the
> > maximum value (0x7FFFFFFF), which will take ~4 seconds to expire
> > again.
> > This means that if a vcpu cedes, it will be ~4 seconds between
> > wakeups.
> > 
> > The h_cede_tm test is testing a known breakage when a guest cedes
> > while
> > suspended. To be sure we cede 500 times to check for the bug.
> > However
> > since it takes ~4 seconds to be woken up once we've ceded, we only
> > get
> > through ~20 iterations before we reach the 90 seconds timeout and
> > the
> > test appears to fail.
> > 
> > Add an option when registering the decrementer handler to specify
> > the
> > value which should be reloaded by the handler, allowing the timeout
> > to be
> > chosen.
> > 
> > Modify the spr test to use the max timeout to preserve existing
> > behaviour.
> > Modify the h_cede_tm test to use a 10ms timeout to ensure we can
> > perform
> > 500 iterations before hitting the 90 second time limit for a test.
> > 
> > This means the h_cede_tm test now succeeds rather than timing out.
> > 
> > Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
> > Reviewed-by: Thomas Huth <thuth@redhat.com>
> > Reviewed-by: Laurent Vivier <lvivier@redhat.com>
> > [lv: reset initial value to 0x3FFFFFFF]
> 
> Looks like something went wrong here? There is still the 0x7FFFFFFF
> in
> the hunk below...

No, I think this is correct.
Max value is ox7FFFFFFF, but the initial value we load via mtdec is the
original 0x3FFFFFFF.

> 
> > diff --git a/powerpc/sprs.c b/powerpc/sprs.c
> > index 6744bd8d8049..3c2d98c9ca99 100644
> > --- a/powerpc/sprs.c
> > +++ b/powerpc/sprs.c
> > @@ -253,6 +253,7 @@ int main(int argc, char **argv)
> >  		0x1234567890ABCDEFULL, 0xFEDCBA0987654321ULL,
> >  		-1ULL,
> >  	};
> > +	static uint64_t decr = 0x7FFFFFFF; /* Max value */
> >  
> >  	for (i = 1; i < argc; i++) {
> >  		if (!strcmp(argv[i], "-w")) {
> > @@ -288,7 +289,7 @@ int main(int argc, char **argv)
> >  		(void) getchar();
> >  	} else {
> >  		puts("Sleeping...\n");
> > -		handle_exception(0x900, &dec_except_handler,
> > NULL);
> > +		handle_exception(0x900, &dec_except_handler,
> > &decr);
> >  		asm volatile ("mtdec %0" : : "r" (0x3FFFFFFF));
> >  		hcall(H_CEDE);
> >  	}
> 
>  Thomas
Laurent Vivier May 20, 2019, 6:12 a.m. UTC | #3 
On 20/05/2019 06:47, Suraj Jitindar Singh wrote:
> On Fri, 2019-05-17 at 15:10 +0200, Thomas Huth wrote:
>> On 17/05/2019 15.03, Laurent Vivier wrote:
>>> From: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
>>>
>>> Currently the handler for a decrementer exception will simply
>>> reload the
>>> maximum value (0x7FFFFFFF), which will take ~4 seconds to expire
>>> again.
>>> This means that if a vcpu cedes, it will be ~4 seconds between
>>> wakeups.
>>>
>>> The h_cede_tm test is testing a known breakage when a guest cedes
>>> while
>>> suspended. To be sure we cede 500 times to check for the bug.
>>> However
>>> since it takes ~4 seconds to be woken up once we've ceded, we only
>>> get
>>> through ~20 iterations before we reach the 90 seconds timeout and
>>> the
>>> test appears to fail.
>>>
>>> Add an option when registering the decrementer handler to specify
>>> the
>>> value which should be reloaded by the handler, allowing the timeout
>>> to be
>>> chosen.
>>>
>>> Modify the spr test to use the max timeout to preserve existing
>>> behaviour.
>>> Modify the h_cede_tm test to use a 10ms timeout to ensure we can
>>> perform
>>> 500 iterations before hitting the 90 second time limit for a test.
>>>
>>> This means the h_cede_tm test now succeeds rather than timing out.
>>>
>>> Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
>>> Reviewed-by: Laurent Vivier <lvivier@redhat.com>
>>> [lv: reset initial value to 0x3FFFFFFF]
>>
>> Looks like something went wrong here? There is still the 0x7FFFFFFF
>> in
>> the hunk below...
> 
> No, I think this is correct.
> Max value is ox7FFFFFFF, but the initial value we load via mtdec is the
> original 0x3FFFFFFF.

Yes, that's it.

Thanks,
Laurent
Thomas Huth May 20, 2019, 7:03 a.m. UTC | #4 
On 20/05/2019 08.12, Laurent Vivier wrote:
> On 20/05/2019 06:47, Suraj Jitindar Singh wrote:
>> On Fri, 2019-05-17 at 15:10 +0200, Thomas Huth wrote:
>>> On 17/05/2019 15.03, Laurent Vivier wrote:
>>>> From: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
>>>>
>>>> Currently the handler for a decrementer exception will simply
>>>> reload the
>>>> maximum value (0x7FFFFFFF), which will take ~4 seconds to expire
>>>> again.
>>>> This means that if a vcpu cedes, it will be ~4 seconds between
>>>> wakeups.
>>>>
>>>> The h_cede_tm test is testing a known breakage when a guest cedes
>>>> while
>>>> suspended. To be sure we cede 500 times to check for the bug.
>>>> However
>>>> since it takes ~4 seconds to be woken up once we've ceded, we only
>>>> get
>>>> through ~20 iterations before we reach the 90 seconds timeout and
>>>> the
>>>> test appears to fail.
>>>>
>>>> Add an option when registering the decrementer handler to specify
>>>> the
>>>> value which should be reloaded by the handler, allowing the timeout
>>>> to be
>>>> chosen.
>>>>
>>>> Modify the spr test to use the max timeout to preserve existing
>>>> behaviour.
>>>> Modify the h_cede_tm test to use a 10ms timeout to ensure we can
>>>> perform
>>>> 500 iterations before hitting the 90 second time limit for a test.
>>>>
>>>> This means the h_cede_tm test now succeeds rather than timing out.
>>>>
>>>> Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
>>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
>>>> Reviewed-by: Laurent Vivier <lvivier@redhat.com>
>>>> [lv: reset initial value to 0x3FFFFFFF]
>>>
>>> Looks like something went wrong here? There is still the 0x7FFFFFFF
>>> in
>>> the hunk below...
>>
>> No, I think this is correct.
>> Max value is ox7FFFFFFF, but the initial value we load via mtdec is the
>> original 0x3FFFFFFF.
> 
> Yes, that's it.

Ok, since we're only calling H_CEDE once here, we should be fine,
indeed. Sorry for the confusion.

 Thomas
diff mbox series

Patch

diff --git a/lib/powerpc/handlers.c b/lib/powerpc/handlers.c
index be8226a2e573..c8721e0a11ae 100644
--- a/lib/powerpc/handlers.c
+++ b/lib/powerpc/handlers.c
@@ -12,11 +12,12 @@ 
 
 /*
  * Generic handler for decrementer exceptions (0x900)
- * Just reset the decrementer back to its maximum value (0x7FFFFFFF)
+ * Just reset the decrementer back to the value specified when registering the
+ * handler
  */
-void dec_except_handler(struct pt_regs *regs __unused, void *data __unused)
+void dec_except_handler(struct pt_regs *regs __unused, void *data)
 {
-	uint32_t dec = 0x7FFFFFFF;
+	uint64_t dec = *((uint64_t *) data);
 
 	asm volatile ("mtdec %0" : : "r" (dec));
 }
diff --git a/powerpc/sprs.c b/powerpc/sprs.c
index 6744bd8d8049..3c2d98c9ca99 100644
--- a/powerpc/sprs.c
+++ b/powerpc/sprs.c
@@ -253,6 +253,7 @@  int main(int argc, char **argv)
 		0x1234567890ABCDEFULL, 0xFEDCBA0987654321ULL,
 		-1ULL,
 	};
+	static uint64_t decr = 0x7FFFFFFF; /* Max value */
 
 	for (i = 1; i < argc; i++) {
 		if (!strcmp(argv[i], "-w")) {
@@ -288,7 +289,7 @@  int main(int argc, char **argv)
 		(void) getchar();
 	} else {
 		puts("Sleeping...\n");
-		handle_exception(0x900, &dec_except_handler, NULL);
+		handle_exception(0x900, &dec_except_handler, &decr);
 		asm volatile ("mtdec %0" : : "r" (0x3FFFFFFF));
 		hcall(H_CEDE);
 	}
diff --git a/powerpc/tm.c b/powerpc/tm.c
index bd56baa5b3d8..c588985352f4 100644
--- a/powerpc/tm.c
+++ b/powerpc/tm.c
@@ -95,11 +95,13 @@  static bool enable_tm(void)
 static void test_h_cede_tm(int argc, char **argv)
 {
 	int i;
+	static uint64_t decr = 0x3FFFFF; /* ~10ms */
 
 	if (argc > 2)
 		report_abort("Unsupported argument: '%s'", argv[2]);
 
-	handle_exception(0x900, &dec_except_handler, NULL);
+	handle_exception(0x900, &dec_except_handler, &decr);
+	asm volatile ("mtdec %0" : : "r" (decr));
 
 	if (!start_all_cpus(halt, 0))
 		report_abort("Failed to start secondary cpus");