On 5/31/21 1:15 PM, David Hildenbrand wrote:
> On 31.05.21 12:50, Janosch Frank wrote:
>> To make our TAP parser (and me) happy we don't want to have to reports
>> with exactly the same wording.
>>
>> Signed-off-by: Janosch Frank <frankja@xxxxxxxxxxxxx>
>> ---
>> s390x/selftest.c | 18 +++++++++++++-----
>> 1 file changed, 13 insertions(+), 5 deletions(-)
>>
>> diff --git a/s390x/selftest.c b/s390x/selftest.c
>> index b2fe2e7b..c2ca9896 100644
>> --- a/s390x/selftest.c
>> +++ b/s390x/selftest.c
>> @@ -47,12 +47,19 @@ static void test_malloc(void)
>> *tmp2 = 123456789;
>> mb();
>>
>> - report((uintptr_t)tmp & 0xf000000000000000ul, "malloc: got vaddr");
>> - report(*tmp == 123456789, "malloc: access works");
>> + report_prefix_push("malloc");
>> + report_prefix_push("ptr_0");
>
> instead of this "ptr_0" vs. "ptr_1" I'd just use
>
> "allocated 1st page"
> "wrote to 1st page"
> "allocated 2nd page"
> "wrote to 2nd page"
> "1st and 2nd page differ"
>
> Avoids one hierarchy of prefix_push ...
I'd like to keep them since I'll also move the allocation and writes
into a prefix section for the v2 which will provide me with better error
reports if we trigger asserts.
Also from the allocation alone these could be on the same page since we
allocate ints.
>
>> + report((uintptr_t)tmp & 0xf000000000000000ul, "allocated memory");
>> + report(*tmp == 123456789, "wrote allocated memory");
>> + report_prefix_pop();
>> +
>> + report_prefix_push("ptr_1");
>> report((uintptr_t)tmp2 & 0xf000000000000000ul,
>> - "malloc: got 2nd vaddr");
>> - report((*tmp2 == 123456789), "malloc: access works");
>> - report(tmp != tmp2, "malloc: addresses differ");
>> + "allocated memory");
>> + report((*tmp2 == 123456789), "wrote allocated memory");
>> + report_prefix_pop();
>> +
>> + report(tmp != tmp2, "allocated memory addresses differ");
>>
>> expect_pgm_int();
>> configure_dat(0);
>> @@ -62,6 +69,7 @@ static void test_malloc(void)
>>
>> free(tmp);
>> free(tmp2);
>> + report_prefix_pop();
>> }
>>
>> int main(int argc, char**argv)
>>
>
>
