Re: [kvm-unit-tests PATCH v2 1/5] s390x: Add specification exception test

[Janis Schoetterl-Glausch <scgl@linux.vnet.ibm.com>]

On 10/5/21 4:51 PM, Thomas Huth wrote:
> On 05/10/2021 11.09, Janis Schoetterl-Glausch wrote:
>> Generate specification exceptions and check that they occur.
>> With the iterations argument one can check if specification
>> exception interpretation occurs, e.g. by using a high value and
>> checking that the debugfs counters are substantially lower.
>> The argument is also useful for estimating the performance benefit
>> of interpretation.
>>
>> Signed-off-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com>
>> ---
>>   s390x/Makefile      |   1 +
>>   s390x/spec_ex.c     | 182 ++++++++++++++++++++++++++++++++++++++++++++
>>   s390x/unittests.cfg |   3 +
>>   3 files changed, 186 insertions(+)
>>   create mode 100644 s390x/spec_ex.c
>>
>> diff --git a/s390x/Makefile b/s390x/Makefile
>> index ef8041a..57d7c9e 100644
>> --- a/s390x/Makefile
>> +++ b/s390x/Makefile
>> @@ -24,6 +24,7 @@ tests += $(TEST_DIR)/mvpg.elf
>>   tests += $(TEST_DIR)/uv-host.elf
>>   tests += $(TEST_DIR)/edat.elf
>>   tests += $(TEST_DIR)/mvpg-sie.elf
>> +tests += $(TEST_DIR)/spec_ex.elf
>>     tests_binary = $(patsubst %.elf,%.bin,$(tests))
>>   ifneq ($(HOST_KEY_DOCUMENT),)
>> diff --git a/s390x/spec_ex.c b/s390x/spec_ex.c
>> new file mode 100644
>> index 0000000..dd0ee53
>> --- /dev/null
>> +++ b/s390x/spec_ex.c
>> @@ -0,0 +1,182 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * © Copyright IBM Corp. 2021
> 
> Could we please avoid non-ASCII characters in source code if possible? ... it's maybe best if you do the Copyright line similar to the other *.c files from IBM that are already in the repository.

Yes, I'll remove it. I thought it would be fine since it's in a comment,
didn't consider that it might cause trouble with some mail clients.
So that's grounds for removal by itself.
> 
>> + * Specification exception test.
>> + * Tests that specification exceptions occur when expected.
>> + */
>> +#include <stdlib.h>
>> +#include <libcflat.h>
>> +#include <asm/interrupt.h>
>> +#include <asm/facility.h>
>> +
>> +static struct lowcore *lc = (struct lowcore *) 0;
>> +
>> +static bool expect_invalid_psw;
>> +static struct psw expected_psw;
>> +static struct psw fixup_psw;
>> +
>> +/* The standard program exception handler cannot deal with invalid old PSWs,
>> + * especially not invalid instruction addresses, as in that case one cannot
>> + * find the instruction following the faulting one from the old PSW.
>> + * The PSW to return to is set by load_psw.
>> + */
>> +static void fixup_invalid_psw(void)
>> +{
>> +    if (expect_invalid_psw) {
>> +        report(expected_psw.mask == lc->pgm_old_psw.mask
>> +               && expected_psw.addr == lc->pgm_old_psw.addr,
>> +               "Invalid program new PSW as expected");
>> +        expect_invalid_psw = false;
>> +    }
>> +    lc->pgm_old_psw = fixup_psw;
>> +}
>> +
>> +static void load_psw(struct psw psw)
>> +{
>> +    uint64_t r0 = 0, r1 = 0;
>> +
>> +    asm volatile (
>> +        "    epsw    %0,%1\n"
>> +        "    st    %0,%[mask]\n"
>> +        "    st    %1,4+%[mask]\n"
>> +        "    larl    %0,nop%=\n"
>> +        "    stg    %0,%[addr]\n"
>> +        "    lpswe    %[psw]\n"
>> +        "nop%=:    nop\n"
>> +        : "+&r"(r0), "+&a"(r1), [mask] "=&R"(fixup_psw.mask),
>> +          [addr] "=&R"(fixup_psw.addr)
> 
> stg uses long displacement, so maybe the constraint should rather be "T" instead?

Good catch.
> 
>> +        : [psw] "Q"(psw)
>> +        : "cc", "memory"
>> +    );
>> +}
>> +

[...]

>> +static struct args parse_args(int argc, char **argv)
>> +{
>> +    struct args args = {
>> +        .iterations = 1,
>> +    };
>> +    unsigned int i;
>> +    long arg;
>> +    bool no_arg;
>> +    char *end;
>> +
>> +    for (i = 1; i < argc; i++) {
>> +        no_arg = true;
>> +        if (i < argc - 1) {
>> +            no_arg = *argv[i+1] == '\0';
>> +            arg = strtol(argv[i+1], &end, 10);
> 
> Nit: It's more common to use spaces around the "+" (i.e. "i + 1")

Ok
> 
>> +            no_arg |= *end != '\0';
>> +            no_arg |= arg < 0;
>> +        }
>> +
>> +        if (!strcmp("--iterations", argv[i])) {
>> +            if (no_arg)
>> +                report_abort("--iterations needs a positive parameter");
>> +            args.iterations = arg;
>> +            ++i;
>> +        } else {
>> +            report_abort("Unsupported parameter '%s'",
>> +                     argv[i]);
>> +        }
>> +    }
>> +    return args;
>> +}
>> +
>> +int main(int argc, char **argv)
>> +{
>> +    unsigned int i;
>> +
>> +    struct args args = parse_args(argc, argv);
>> +
>> +    report_prefix_push("specification exception");
>> +    for (i = 0; spec_ex_triggers[i].name; i++) {
>> +        report_prefix_push(spec_ex_triggers[i].name);
>> +        test_spec_ex(&args, &spec_ex_triggers[i]);
>> +        report_prefix_pop();
>> +    }
>> +    report_prefix_pop();
>> +
>> +    return report_summary();
>> +}
> 
> Apart from the nits, this looks fine to me.

Thanks for the review.
> 
>  Thomas
>