[kvm-unit-tests PATCH] s390x/intercept: Test invalid prefix argument to SET PREFIX

[kvm-unit-tests PATCH] s390x/intercept: Test invalid prefix argument to SET PREFIX

[Janis Schoetterl-Glausch]

According to the architecture, SET PREFIX must try to access the new
prefix area and recognize an addressing exception if the area is not
accessible.
Test that the exception occurs when we try to set a prefix higher
than the available memory.

Signed-off-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com>
---
 s390x/intercept.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/s390x/intercept.c b/s390x/intercept.c
index 86e57e11..0b90e588 100644
--- a/s390x/intercept.c
+++ b/s390x/intercept.c
@@ -74,6 +74,20 @@ static void test_spx(void)
 	expect_pgm_int();
 	asm volatile(" spx 0(%0) " : : "r"(-8L));
 	check_pgm_int_code(PGM_INT_CODE_ADDRESSING);
+
+	new_prefix = get_ram_size() & 0x7fffe000;
+	if (get_ram_size() - new_prefix < 2 * PAGE_SIZE) {
+		expect_pgm_int();
+		asm volatile("spx	%0 " : : "Q"(new_prefix));
+		check_pgm_int_code(PGM_INT_CODE_ADDRESSING);
+
+		/*
+		 * Cannot test inaccessibility of the second page the same way.
+		 * If we try to use the last page as first half of the prefix
+		 * area and our ram size is a multiple of 8k, after SPX aligns
+		 * the address to 8k we have a completely accessible area.
+		 */
+	}
 }
 
 /* Test the STORE CPU ADDRESS instruction */

base-commit: 110c69492b53f0070e1bbce986fb635e72a423b4
-- 
2.36.1

Re: [kvm-unit-tests PATCH] s390x/intercept: Test invalid prefix argument to SET PREFIX

[Claudio Imbrenda]

On Mon, 27 Jun 2022 14:43:56 +0200
Janis Schoetterl-Glausch <scgl@linux.ibm.com> wrote:

> According to the architecture, SET PREFIX must try to access the new
> prefix area and recognize an addressing exception if the area is not
> accessible.
> Test that the exception occurs when we try to set a prefix higher
> than the available memory.
> 
> Signed-off-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com>
> ---
>  s390x/intercept.c | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
> 
> diff --git a/s390x/intercept.c b/s390x/intercept.c
> index 86e57e11..0b90e588 100644
> --- a/s390x/intercept.c
> +++ b/s390x/intercept.c
> @@ -74,6 +74,20 @@ static void test_spx(void)
>  	expect_pgm_int();
>  	asm volatile(" spx 0(%0) " : : "r"(-8L));
>  	check_pgm_int_code(PGM_INT_CODE_ADDRESSING);
> +
> +	new_prefix = get_ram_size() & 0x7fffe000;
> +	if (get_ram_size() - new_prefix < 2 * PAGE_SIZE) {
> +		expect_pgm_int();
> +		asm volatile("spx	%0 " : : "Q"(new_prefix));
> +		check_pgm_int_code(PGM_INT_CODE_ADDRESSING);
> +
> +		/*
> +		 * Cannot test inaccessibility of the second page the same way.
> +		 * If we try to use the last page as first half of the prefix
> +		 * area and our ram size is a multiple of 8k, after SPX aligns
> +		 * the address to 8k we have a completely accessible area.
> +		 */
> +	}

please add something like:

else {
	report_skip("Inaccessible prefix");
}

>  }
>  
>  /* Test the STORE CPU ADDRESS instruction */
> 
> base-commit: 110c69492b53f0070e1bbce986fb635e72a423b4

Re: [kvm-unit-tests PATCH] s390x/intercept: Test invalid prefix argument to SET PREFIX

[Janis Schoetterl-Glausch]

On 6/27/22 14:52, Claudio Imbrenda wrote:
> On Mon, 27 Jun 2022 14:43:56 +0200
> Janis Schoetterl-Glausch <scgl@linux.ibm.com> wrote:
> 
>> According to the architecture, SET PREFIX must try to access the new
>> prefix area and recognize an addressing exception if the area is not
>> accessible.
>> Test that the exception occurs when we try to set a prefix higher
>> than the available memory.
>>
>> Signed-off-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com>
>> ---
>>  s390x/intercept.c | 14 ++++++++++++++
>>  1 file changed, 14 insertions(+)
>>
>> diff --git a/s390x/intercept.c b/s390x/intercept.c
>> index 86e57e11..0b90e588 100644
>> --- a/s390x/intercept.c
>> +++ b/s390x/intercept.c
>> @@ -74,6 +74,20 @@ static void test_spx(void)
>>  	expect_pgm_int();
>>  	asm volatile(" spx 0(%0) " : : "r"(-8L));
>>  	check_pgm_int_code(PGM_INT_CODE_ADDRESSING);
>> +
>> +	new_prefix = get_ram_size() & 0x7fffe000;
>> +	if (get_ram_size() - new_prefix < 2 * PAGE_SIZE) {
>> +		expect_pgm_int();
>> +		asm volatile("spx	%0 " : : "Q"(new_prefix));
>> +		check_pgm_int_code(PGM_INT_CODE_ADDRESSING);
>> +
>> +		/*
>> +		 * Cannot test inaccessibility of the second page the same way.
>> +		 * If we try to use the last page as first half of the prefix
>> +		 * area and our ram size is a multiple of 8k, after SPX aligns
>> +		 * the address to 8k we have a completely accessible area.
>> +		 */
>> +	}
> 
> please add something like:
> 
> else {
> 	report_skip("Inaccessible prefix");
> }
> 

Yeah, good idea.

Is there any low effort way of generating an invalid prefix other than > ram_size ?
That might then allow for testing the second page also.

>>  }
>>  
>>  /* Test the STORE CPU ADDRESS instruction */
>>
>> base-commit: 110c69492b53f0070e1bbce986fb635e72a423b4
>