[kvm-unit-tests PATCH v2 0/3] s390x: add tests for SIGP call orders in enabled wait

[kvm-unit-tests PATCH v2 0/3] s390x: add tests for SIGP call orders in enabled wait

[Nico Boehr]

v1->v2:
---
* rebase to latest master to align with Claudio's SMP changes, drop
  patch which adds the ext int clean up since it is already in Claudio's
  series
* make sure ctl0 register bit is cleared

When a CPU is in enabled wait, it can still receive SIGP calls from
other CPUs.

Since this requires some special handling in KVM, we should have tests
for it. This has already revealed a KVM bug with ecall under PV, which
is why this test currently fails there.

Some refactoring is done as part of this series to reduce code
duplication.

Nico Boehr (3):
  s390x: smp: move sigp calls with invalid cpu address to array
  s390x: smp: use an array for sigp calls
  s390x: smp: add tests for calls in wait state

 s390x/smp.c | 190 +++++++++++++++++++++++++++++-----------------------
 1 file changed, 108 insertions(+), 82 deletions(-)

-- 
2.36.1

[kvm-unit-tests PATCH v2 1/3] s390x: smp: move sigp calls with invalid cpu address to array

[Nico Boehr]

We have the nice array to test SIGP calls with invalid CPU addresses.
Move the SIGP cases there to eliminate some of the duplicated code in
test_emcall and test_cond_emcall.

Since adding coverage for invalid CPU addresses in the ecall case is now
trivial, do that as well.

Signed-off-by: Nico Boehr <nrb@linux.ibm.com>
---
 s390x/smp.c | 18 +++---------------
 1 file changed, 3 insertions(+), 15 deletions(-)

diff --git a/s390x/smp.c b/s390x/smp.c
index 0df4751f9cee..34ae91c3fe12 100644
--- a/s390x/smp.c
+++ b/s390x/smp.c
@@ -30,6 +30,9 @@ static const struct sigp_invalid_cases cases_invalid_cpu_addr[] = {
 	{ SIGP_STOP,                  "stop with invalid CPU address" },
 	{ SIGP_START,                 "start with invalid CPU address" },
 	{ SIGP_CPU_RESET,             "reset with invalid CPU address" },
+	{ SIGP_COND_EMERGENCY_SIGNAL, "conditional emcall with invalid CPU address" },
+	{ SIGP_EMERGENCY_SIGNAL,      "emcall with invalid CPU address" },
+	{ SIGP_EXTERNAL_CALL,         "ecall with invalid CPU address" },
 	{ INVALID_ORDER_CODE,         "invalid order code and CPU address" },
 	{ SIGP_SENSE,                 "sense with invalid CPU address" },
 	{ SIGP_STOP_AND_STORE_STATUS, "stop and store status with invalid CPU address" },
@@ -329,7 +332,6 @@ static void emcall(void)
 static void test_emcall(void)
 {
 	struct psw psw;
-	int cc;
 	psw.mask = extract_psw_mask();
 	psw.addr = (unsigned long)emcall;
 
@@ -343,13 +345,6 @@ static void test_emcall(void)
 	wait_for_flag();
 	smp_cpu_stop(1);
 
-	report_prefix_push("invalid CPU address");
-
-	cc = sigp(INVALID_CPU_ADDRESS, SIGP_EMERGENCY_SIGNAL, 0, NULL);
-	report(cc == 3, "CC = 3");
-
-	report_prefix_pop();
-
 	report_prefix_pop();
 }
 
@@ -368,13 +363,6 @@ static void test_cond_emcall(void)
 		goto out;
 	}
 
-	report_prefix_push("invalid CPU address");
-
-	cc = sigp(INVALID_CPU_ADDRESS, SIGP_COND_EMERGENCY_SIGNAL, 0, NULL);
-	report(cc == 3, "CC = 3");
-
-	report_prefix_pop();
-
 	report_prefix_push("success");
 	set_flag(0);
 
-- 
2.36.1

[kvm-unit-tests PATCH v2 2/3] s390x: smp: use an array for sigp calls

[Nico Boehr]

Tests for the SIGP calls are quite similar, so we have a lot of code
duplication right now. Since upcoming changes will add more cases,
refactor the code to iterate over an array, similarily as we already do
for test_invalid().

The receiving CPU is disabled for IO interrupts. This makes sure the
conditional emergency signal is accepted and doesn't hurt the other
orders.

Signed-off-by: Nico Boehr <nrb@linux.ibm.com>
---
 s390x/smp.c | 127 +++++++++++++++++++---------------------------------
 1 file changed, 45 insertions(+), 82 deletions(-)

diff --git a/s390x/smp.c b/s390x/smp.c
index 34ae91c3fe12..683b0e618a48 100644
--- a/s390x/smp.c
+++ b/s390x/smp.c
@@ -43,6 +43,20 @@ static const struct sigp_invalid_cases cases_valid_cpu_addr[] = {
 
 static uint32_t cpu1_prefix;
 
+struct sigp_call_cases {
+	char name[20];
+	int call;
+	uint16_t ext_int_expected_type;
+	uint32_t cr0_bit;
+	bool supports_pv;
+};
+static const struct sigp_call_cases cases_sigp_call[] = {
+	{ "emcall",      SIGP_EMERGENCY_SIGNAL,      0x1201, CTL0_EMERGENCY_SIGNAL, true },
+	{ "cond emcall", SIGP_COND_EMERGENCY_SIGNAL, 0x1201, CTL0_EMERGENCY_SIGNAL, false },
+	{ "ecall",       SIGP_EXTERNAL_CALL,         0x1202, CTL0_EXTERNAL_CALL,    true },
+};
+static const struct sigp_call_cases *current_sigp_call_case;
+
 static void test_invalid(void)
 {
 	const struct sigp_invalid_cases *c;
@@ -289,97 +303,48 @@ static void test_set_prefix(void)
 
 }
 
-static void ecall(void)
+static void call_received(void)
 {
 	expect_ext_int();
-	ctl_set_bit(0, CTL0_EXTERNAL_CALL);
-	psw_mask_set_bits(PSW_MASK_EXT);
-	set_flag(1);
-	while (lowcore.ext_int_code != 0x1202) { mb(); }
-	report_pass("received");
-	set_flag(1);
-}
-
-static void test_ecall(void)
-{
-	struct psw psw;
-	psw.mask = extract_psw_mask();
-	psw.addr = (unsigned long)ecall;
-
-	report_prefix_push("ecall");
-	set_flag(0);
-
-	smp_cpu_start(1, psw);
-	wait_for_flag();
-	set_flag(0);
-	smp_sigp(1, SIGP_EXTERNAL_CALL, 0, NULL);
-	wait_for_flag();
-	smp_cpu_stop(1);
-	report_prefix_pop();
-}
+	ctl_set_bit(0, current_sigp_call_case->cr0_bit);
+	/* make sure conditional emergency is accepted by disabling IO interrupts */
+	psw_mask_clear_and_set_bits(PSW_MASK_IO, PSW_MASK_EXT);
 
-static void emcall(void)
-{
-	expect_ext_int();
-	ctl_set_bit(0, CTL0_EMERGENCY_SIGNAL);
-	psw_mask_set_bits(PSW_MASK_EXT);
 	set_flag(1);
-	while (lowcore.ext_int_code != 0x1201) { mb(); }
+	while (lowcore.ext_int_code != current_sigp_call_case->ext_int_expected_type)
+		mb();
 	report_pass("received");
+	ctl_clear_bit(0, current_sigp_call_case->cr0_bit);
 	set_flag(1);
 }
 
-static void test_emcall(void)
+static void test_calls(void)
 {
+	int i;
 	struct psw psw;
-	psw.mask = extract_psw_mask();
-	psw.addr = (unsigned long)emcall;
-
-	report_prefix_push("emcall");
-	set_flag(0);
-
-	smp_cpu_start(1, psw);
-	wait_for_flag();
-	set_flag(0);
-	smp_sigp(1, SIGP_EMERGENCY_SIGNAL, 0, NULL);
-	wait_for_flag();
-	smp_cpu_stop(1);
-
-	report_prefix_pop();
-}
-
-static void test_cond_emcall(void)
-{
-	uint32_t status = 0;
-	struct psw psw;
-	int cc;
-	psw.mask = extract_psw_mask() & ~PSW_MASK_IO;
-	psw.addr = (unsigned long)emcall;
 
-	report_prefix_push("conditional emergency call");
-
-	if (uv_os_is_guest()) {
-		report_skip("unsupported under PV");
-		goto out;
+	for (i = 0; i < ARRAY_SIZE(cases_sigp_call); i++) {
+		current_sigp_call_case = &cases_sigp_call[i];
+
+		report_prefix_push(current_sigp_call_case->name);
+		if (!current_sigp_call_case->supports_pv && uv_os_is_guest()) {
+			report_skip("Not supported under PV");
+			report_prefix_pop();
+			continue;
+		}
+
+		set_flag(0);
+		psw.mask = extract_psw_mask();
+		psw.addr = (unsigned long)call_received;
+
+		smp_cpu_start(1, psw);
+		wait_for_flag();
+		set_flag(0);
+		smp_sigp(1, current_sigp_call_case->call, 0, NULL);
+		wait_for_flag();
+		smp_cpu_stop(1);
+		report_prefix_pop();
 	}
-
-	report_prefix_push("success");
-	set_flag(0);
-
-	smp_cpu_start(1, psw);
-	wait_for_flag();
-	set_flag(0);
-	cc = smp_sigp(1, SIGP_COND_EMERGENCY_SIGNAL, 0, &status);
-	report(!cc, "CC = 0");
-
-	wait_for_flag();
-	smp_cpu_stop(1);
-
-	report_prefix_pop();
-
-out:
-	report_prefix_pop();
-
 }
 
 static void test_sense_running(void)
@@ -499,9 +464,7 @@ int main(void)
 	test_stop_store_status();
 	test_store_status();
 	test_set_prefix();
-	test_ecall();
-	test_emcall();
-	test_cond_emcall();
+	test_calls();
 	test_sense_running();
 	test_reset();
 	test_reset_initial();
-- 
2.36.1

[kvm-unit-tests PATCH v2 3/3] s390x: smp: add tests for calls in wait state

[Nico Boehr]

Under PV, SIGP ecall requires some special handling by the hypervisor
when the receiving CPU is in enabled wait. Hence, we should have
coverage for the various SIGP call orders when the receiving CPU is in
enabled wait.

The ecall test currently fails under PV due to a KVM bug under
investigation.

Signed-off-by: Nico Boehr <nrb@linux.ibm.com>
---
 s390x/smp.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 75 insertions(+)

diff --git a/s390x/smp.c b/s390x/smp.c
index 683b0e618a48..eed7aa3564de 100644
--- a/s390x/smp.c
+++ b/s390x/smp.c
@@ -347,6 +347,80 @@ static void test_calls(void)
 	}
 }
 
+static void call_in_wait_ext_int_fixup(struct stack_frame_int *stack)
+{
+	/* leave wait after returning */
+	lowcore.ext_old_psw.mask &= ~PSW_MASK_WAIT;
+
+	stack->crs[0] &= ~current_sigp_call_case->cr0_bit;
+}
+
+static void call_in_wait_setup(void)
+{
+	expect_ext_int();
+	ctl_set_bit(0, current_sigp_call_case->cr0_bit);
+	register_ext_cleanup_func(call_in_wait_ext_int_fixup);
+
+	set_flag(1);
+}
+
+static void call_in_wait_received(void)
+{
+	report(lowcore.ext_int_code == current_sigp_call_case->ext_int_expected_type, "received");
+	ctl_clear_bit(0, current_sigp_call_case->cr0_bit);
+	register_ext_cleanup_func(NULL);
+
+	set_flag(1);
+}
+
+static void test_calls_in_wait(void)
+{
+	int i;
+	struct psw psw;
+
+	report_prefix_push("psw wait");
+	for (i = 0; i < ARRAY_SIZE(cases_sigp_call); i++) {
+		current_sigp_call_case = &cases_sigp_call[i];
+
+		report_prefix_push(current_sigp_call_case->name);
+		if (!current_sigp_call_case->supports_pv && uv_os_is_guest()) {
+			report_skip("Not supported under PV");
+			report_prefix_pop();
+			continue;
+		}
+
+		set_flag(0);
+		psw.mask = extract_psw_mask();
+		psw.addr = (unsigned long)call_in_wait_setup;
+		smp_cpu_start(1, psw);
+		wait_for_flag();
+		set_flag(0);
+
+		/*
+		 * To avoid races, we need to know that the secondary CPU has entered wait,
+		 * but the architecture provides no way to check whether the secondary CPU
+		 * is in wait.
+		 *
+		 * But since a waiting CPU is considered operating, simply stop the CPU, set
+		 * up the restart new PSW mask in wait, send the restart interrupt and then
+		 * wait until the CPU becomes operating (done by smp_cpu_start).
+		 */
+		smp_cpu_stop(1);
+		expect_ext_int();
+		psw.mask = extract_psw_mask() | PSW_MASK_EXT | PSW_MASK_WAIT;
+		psw.addr = (unsigned long)call_in_wait_received;
+		smp_cpu_start(1, psw);
+
+		smp_sigp(1, current_sigp_call_case->call, 0, NULL);
+
+		wait_for_flag();
+		smp_cpu_stop(1);
+
+		report_prefix_pop();
+	}
+	report_prefix_pop();
+}
+
 static void test_sense_running(void)
 {
 	report_prefix_push("sense_running");
@@ -465,6 +539,7 @@ int main(void)
 	test_store_status();
 	test_set_prefix();
 	test_calls();
+	test_calls_in_wait();
 	test_sense_running();
 	test_reset();
 	test_reset_initial();
-- 
2.36.1

Re: [kvm-unit-tests PATCH v2 1/3] s390x: smp: move sigp calls with invalid cpu address to array

[Janosch Frank]

On 7/22/22 09:20, Nico Boehr wrote:
> We have the nice array to test SIGP calls with invalid CPU addresses.
> Move the SIGP cases there to eliminate some of the duplicated code in
> test_emcall and test_cond_emcall.
> 
> Since adding coverage for invalid CPU addresses in the ecall case is now
> trivial, do that as well.
> 
> Signed-off-by: Nico Boehr <nrb@linux.ibm.com>

Reviewed-by: Janosch Frank <frankja@linux.ibm.com>

> ---
>   s390x/smp.c | 18 +++---------------
>   1 file changed, 3 insertions(+), 15 deletions(-)
> 
> diff --git a/s390x/smp.c b/s390x/smp.c
> index 0df4751f9cee..34ae91c3fe12 100644
> --- a/s390x/smp.c
> +++ b/s390x/smp.c
> @@ -30,6 +30,9 @@ static const struct sigp_invalid_cases cases_invalid_cpu_addr[] = {
>   	{ SIGP_STOP,                  "stop with invalid CPU address" },
>   	{ SIGP_START,                 "start with invalid CPU address" },
>   	{ SIGP_CPU_RESET,             "reset with invalid CPU address" },
> +	{ SIGP_COND_EMERGENCY_SIGNAL, "conditional emcall with invalid CPU address" },
> +	{ SIGP_EMERGENCY_SIGNAL,      "emcall with invalid CPU address" },
> +	{ SIGP_EXTERNAL_CALL,         "ecall with invalid CPU address" },
>   	{ INVALID_ORDER_CODE,         "invalid order code and CPU address" },
>   	{ SIGP_SENSE,                 "sense with invalid CPU address" },
>   	{ SIGP_STOP_AND_STORE_STATUS, "stop and store status with invalid CPU address" },
> @@ -329,7 +332,6 @@ static void emcall(void)
>   static void test_emcall(void)
>   {
>   	struct psw psw;
> -	int cc;
>   	psw.mask = extract_psw_mask();
>   	psw.addr = (unsigned long)emcall;
>   
> @@ -343,13 +345,6 @@ static void test_emcall(void)
>   	wait_for_flag();
>   	smp_cpu_stop(1);
>   
> -	report_prefix_push("invalid CPU address");
> -
> -	cc = sigp(INVALID_CPU_ADDRESS, SIGP_EMERGENCY_SIGNAL, 0, NULL);
> -	report(cc == 3, "CC = 3");
> -
> -	report_prefix_pop();
> -
>   	report_prefix_pop();
>   }
>   
> @@ -368,13 +363,6 @@ static void test_cond_emcall(void)
>   		goto out;
>   	}
>   
> -	report_prefix_push("invalid CPU address");
> -
> -	cc = sigp(INVALID_CPU_ADDRESS, SIGP_COND_EMERGENCY_SIGNAL, 0, NULL);
> -	report(cc == 3, "CC = 3");
> -
> -	report_prefix_pop();
> -
>   	report_prefix_push("success");
>   	set_flag(0);
>   

Re: [kvm-unit-tests PATCH v2 2/3] s390x: smp: use an array for sigp calls

[Janosch Frank]

On 7/22/22 09:20, Nico Boehr wrote:
> Tests for the SIGP calls are quite similar, so we have a lot of code
> duplication right now. Since upcoming changes will add more cases,
> refactor the code to iterate over an array, similarily as we already do
> for test_invalid().
> 
> The receiving CPU is disabled for IO interrupts. This makes sure the
> conditional emergency signal is accepted and doesn't hurt the other
> orders.
> 
> Signed-off-by: Nico Boehr <nrb@linux.ibm.com>

I'd like a few comments to make it more understandable. Other than that:
Reviewed-by: Janosch Frank <frankja@linux.ibm.com>


Thanks for improving this

> ---
>   s390x/smp.c | 127 +++++++++++++++++++---------------------------------
>   1 file changed, 45 insertions(+), 82 deletions(-)
> 
> diff --git a/s390x/smp.c b/s390x/smp.c
> index 34ae91c3fe12..683b0e618a48 100644
> --- a/s390x/smp.c
> +++ b/s390x/smp.c
> @@ -43,6 +43,20 @@ static const struct sigp_invalid_cases cases_valid_cpu_addr[] = {
>   
>   static uint32_t cpu1_prefix;
>   
> +struct sigp_call_cases {
> +	char name[20];
> +	int call;
> +	uint16_t ext_int_expected_type;
> +	uint32_t cr0_bit;
> +	bool supports_pv;
> +};
> +static const struct sigp_call_cases cases_sigp_call[] = {
> +	{ "emcall",      SIGP_EMERGENCY_SIGNAL,      0x1201, CTL0_EMERGENCY_SIGNAL, true },
> +	{ "cond emcall", SIGP_COND_EMERGENCY_SIGNAL, 0x1201, CTL0_EMERGENCY_SIGNAL, false },
> +	{ "ecall",       SIGP_EXTERNAL_CALL,         0x1202, CTL0_EXTERNAL_CALL,    true },
> +};

/* Indicate that we're ready to receive the call */

>   	set_flag(1);
> -	while (lowcore.ext_int_code != 0x1201) { mb(); }
> +	while (lowcore.ext_int_code != current_sigp_call_case->ext_int_expected_type)
> +		mb();
>   	report_pass("received");
> +	ctl_clear_bit(0, current_sigp_call_case->cr0_bit);

/* Indicate that we're done */

>   	set_flag(1);
>   }
[..]

>   
> -static void test_cond_emcall(void)
> -{
> -	uint32_t status = 0;
> -	struct psw psw;
> -	int cc;
> -	psw.mask = extract_psw_mask() & ~PSW_MASK_IO;
> -	psw.addr = (unsigned long)emcall;
>   
> -	report_prefix_push("conditional emergency call");
> -
> -	if (uv_os_is_guest()) {
> -		report_skip("unsupported under PV");
> -		goto out;
> +	for (i = 0; i < ARRAY_SIZE(cases_sigp_call); i++) {
> +		current_sigp_call_case = &cases_sigp_call[i];
> +
> +		report_prefix_push(current_sigp_call_case->name);
> +		if (!current_sigp_call_case->supports_pv && uv_os_is_guest()) {
> +			report_skip("Not supported under PV");
> +			report_prefix_pop();
> +			continue;
> +		}
> +
> +		set_flag(0);
> +		psw.mask = extract_psw_mask();
> +		psw.addr = (unsigned long)call_received;
> +
> +		smp_cpu_start(1, psw);

/* Wait until the receiver has finished setup */

> +		wait_for_flag();
> +		set_flag(0);
> +		smp_sigp(1, current_sigp_call_case->call, 0, NULL);

/* Wait until the receiver has handled the call */

> +		wait_for_flag();
> +		smp_cpu_stop(1);
> +		report_prefix_pop();
>   	}
> -
> -	report_prefix_push("success");
> -	set_flag(0);
> -
> -	smp_cpu_start(1, psw);
> -	wait_for_flag();
> -	set_flag(0);
> -	cc = smp_sigp(1, SIGP_COND_EMERGENCY_SIGNAL, 0, &status);
> -	report(!cc, "CC = 0");
> -
> -	wait_for_flag();
> -	smp_cpu_stop(1);
> -
> -	report_prefix_pop();
> -
> -out:
> -	report_prefix_pop();
> -
>   }
>   
>   static void test_sense_running(void)
> @@ -499,9 +464,7 @@ int main(void)
>   	test_stop_store_status();
>   	test_store_status();
>   	test_set_prefix();
> -	test_ecall();
> -	test_emcall();
> -	test_cond_emcall();
> +	test_calls();
>   	test_sense_running();
>   	test_reset();
>   	test_reset_initial();

Re: [kvm-unit-tests PATCH v2 3/3] s390x: smp: add tests for calls in wait state

[Janosch Frank]

On 7/22/22 09:20, Nico Boehr wrote:
> Under PV, SIGP ecall requires some special handling by the hypervisor
> when the receiving CPU is in enabled wait. Hence, we should have
> coverage for the various SIGP call orders when the receiving CPU is in
> enabled wait.

When the SIGP interpretation facility is in use a SIGP external call to 
a waiting CPU will result in an exit of the calling cpu. For non-pv 
guests it's a code 56 (partial execution) exit otherwise its a code 108 
(secure instruction notification) exit. Those exits are handled 
differently from a normal SIGP instruction intercept that happens 
without interpretation and hence need to be tested.

> 
> The ecall test currently fails under PV due to a KVM bug under
> investigation.

That shouldn't be true anymore

> 
> Signed-off-by: Nico Boehr <nrb@linux.ibm.com>
> ---
>   s390x/smp.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++++++
>   1 file changed, 75 insertions(+)
> 
> diff --git a/s390x/smp.c b/s390x/smp.c
> index 683b0e618a48..eed7aa3564de 100644
> --- a/s390x/smp.c
> +++ b/s390x/smp.c
> @@ -347,6 +347,80 @@ static void test_calls(void)
>   	}
>   }
>   
> +static void call_in_wait_ext_int_fixup(struct stack_frame_int *stack)
> +{
> +	/* leave wait after returning */

Clear wait bit so we don't immediately wait again after the fixup

> +	lowcore.ext_old_psw.mask &= ~PSW_MASK_WAIT;
> +
> +	stack->crs[0] &= ~current_sigp_call_case->cr0_bit;

You need a mask but have a bit, no?

~BIT(current_sigp_call_case->cr0_bit)

> +}
> +
> +static void call_in_wait_setup(void)
> +{
> +	expect_ext_int();
> +	ctl_set_bit(0, current_sigp_call_case->cr0_bit);
> +	register_ext_cleanup_func(call_in_wait_ext_int_fixup);
> +
> +	set_flag(1);
> +}
> +
> +static void call_in_wait_received(void)
> +{
> +	report(lowcore.ext_int_code == current_sigp_call_case->ext_int_expected_type, "received");
> +	ctl_clear_bit(0, current_sigp_call_case->cr0_bit);
> +	register_ext_cleanup_func(NULL);
> +
> +	set_flag(1);
> +}
> +
> +static void test_calls_in_wait(void)
> +{
> +	int i;
> +	struct psw psw;
> +
> +	report_prefix_push("psw wait");
> +	for (i = 0; i < ARRAY_SIZE(cases_sigp_call); i++) {
> +		current_sigp_call_case = &cases_sigp_call[i];
> +
> +		report_prefix_push(current_sigp_call_case->name);
> +		if (!current_sigp_call_case->supports_pv && uv_os_is_guest()) {
> +			report_skip("Not supported under PV");
> +			report_prefix_pop();
> +			continue;
> +		}
> +
/* Let the secondary CPU setup the external mask and the external 
interrupt cleanup function */
> +		set_flag(0);
> +		psw.mask = extract_psw_mask();
> +		psw.addr = (unsigned long)call_in_wait_setup;
> +		smp_cpu_start(1, psw);
> +		wait_for_flag();
> +		set_flag(0);
> +
> +		/*
> +		 * To avoid races, we need to know that the secondary CPU has entered wait,
> +		 * but the architecture provides no way to check whether the secondary CPU
> +		 * is in wait.
> +		 *
> +		 * But since a waiting CPU is considered operating, simply stop the CPU, set
> +		 * up the restart new PSW mask in wait, send the restart interrupt and then
> +		 * wait until the CPU becomes operating (done by smp_cpu_start).
> +		 */
> +		smp_cpu_stop(1);
> +		expect_ext_int();
> +		psw.mask = extract_psw_mask() | PSW_MASK_EXT | PSW_MASK_WAIT;
> +		psw.addr = (unsigned long)call_in_wait_received;
> +		smp_cpu_start(1, psw);
> +
> +		smp_sigp(1, current_sigp_call_case->call, 0, NULL);
> +
> +		wait_for_flag();
> +		smp_cpu_stop(1);
> +
> +		report_prefix_pop();
> +	}
> +	report_prefix_pop();
> +}
> +
>   static void test_sense_running(void)
>   {
>   	report_prefix_push("sense_running");
> @@ -465,6 +539,7 @@ int main(void)
>   	test_store_status();
>   	test_set_prefix();
>   	test_calls();
> +	test_calls_in_wait();
>   	test_sense_running();
>   	test_reset();
>   	test_reset_initial();

Re: [kvm-unit-tests PATCH v2 3/3] s390x: smp: add tests for calls in wait state

[Nico Boehr]

Quoting Janosch Frank (2022-07-22 10:30:46)
> On 7/22/22 09:20, Nico Boehr wrote:
> > Under PV, SIGP ecall requires some special handling by the hypervisor
> > when the receiving CPU is in enabled wait. Hence, we should have
> > coverage for the various SIGP call orders when the receiving CPU is in
> > enabled wait.
> 
> When the SIGP interpretation facility is in use a SIGP external call to 
> a waiting CPU will result in an exit of the calling cpu. For non-pv 
> guests it's a code 56 (partial execution) exit otherwise its a code 108 
> (secure instruction notification) exit. Those exits are handled 
> differently from a normal SIGP instruction intercept that happens 
> without interpretation and hence need to be tested.

Changed.

> > 
> > The ecall test currently fails under PV due to a KVM bug under
> > investigation.
> 
> That shouldn't be true anymore

Yeah, it's not yet in mainline, but is soon gonna be. I will remove this.

> > 
> > Signed-off-by: Nico Boehr <nrb@linux.ibm.com>
> > ---
> >   s390x/smp.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++++++
> >   1 file changed, 75 insertions(+)
> > 
> > diff --git a/s390x/smp.c b/s390x/smp.c
> > index 683b0e618a48..eed7aa3564de 100644
> > --- a/s390x/smp.c
> > +++ b/s390x/smp.c
> > @@ -347,6 +347,80 @@ static void test_calls(void)
> >       }
> >   }
> >   
> > +static void call_in_wait_ext_int_fixup(struct stack_frame_int *stack)
> > +{
> > +     /* leave wait after returning */
> 
> Clear wait bit so we don't immediately wait again after the fixup

Changed.

> 
> > +     lowcore.ext_old_psw.mask &= ~PSW_MASK_WAIT;
> > +
> > +     stack->crs[0] &= ~current_sigp_call_case->cr0_bit;
> 
> You need a mask but have a bit, no?
> 
> ~BIT(current_sigp_call_case->cr0_bit)

Oopsie, thanks, good find.

This reminds me the ctl_clear_bit() I added in call_in_wait_received() is completely useless, since I handle it here. So, I will remove it there as well.

[...]
> > +static void test_calls_in_wait(void)
> > +{
> > +     int i;
> > +     struct psw psw;
> > +
> > +     report_prefix_push("psw wait");
> > +     for (i = 0; i < ARRAY_SIZE(cases_sigp_call); i++) {
> > +             current_sigp_call_case = &cases_sigp_call[i];
> > +
> > +             report_prefix_push(current_sigp_call_case->name);
> > +             if (!current_sigp_call_case->supports_pv && uv_os_is_guest()) {
> > +                     report_skip("Not supported under PV");
> > +                     report_prefix_pop();
> > +                     continue;
> > +             }
> > +
> /* Let the secondary CPU setup the external mask and the external 
> interrupt cleanup function */

Changed.