Re: [PATCH 3/3] lib/test-string_helpers.c: add string_get_size() tests

[Vitaly Kuznetsov <vkuznets@redhat.com>]

Rasmus Villemoes <linux@rasmusvillemoes.dk> writes:

> On Mon, Oct 26 2015, Vitaly Kuznetsov <vkuznets@redhat.com> wrote:
>
>> Add a couple of simple tests for string_get_size().
>>
>> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>> ---
>>  lib/test-string_helpers.c | 44 ++++++++++++++++++++++++++++++++++++++++++++
>>  1 file changed, 44 insertions(+)
>>
>> diff --git a/lib/test-string_helpers.c b/lib/test-string_helpers.c
>> index 8e376ef..a158cb3 100644
>> --- a/lib/test-string_helpers.c
>> +++ b/lib/test-string_helpers.c
>> @@ -326,6 +326,47 @@ out:
>>  	kfree(out_test);
>>  }
>>  
>> +#define string_get_size_maxbuf 16
>> +#define test_string_get_size_one(size, blk_size, units, exp_result)            \
>> +	do {                                                                   \
>> +		BUILD_BUG_ON(sizeof(exp_result) >= string_get_size_maxbuf);    \
>> +		__test_string_get_size((size), (blk_size), (units),            \
>> +				       (exp_result));                          \
>> +	} while (0)
>> +
>> +
>> +static __init void __test_string_get_size(const u64 size, const u32 blk_size,
>> +					  const enum string_size_units units,
>> +					  const char *exp_result)
>> +{
>> +	char buf[string_get_size_maxbuf];
>> +
>> +	string_get_size(size, blk_size, units, buf, sizeof(buf));
>> +	if (!memcmp(buf, exp_result, strlen(exp_result) + 1))
>> +		return;
>> +
>> +	buf[sizeof(buf) - 1] = '\0';
>> +	pr_warn("Test 'test_string_get_size_one' failed!\n");
>> +	pr_warn("string_get_size(size = %llu, blk_size = %u, units = %d\n",
>> +		size, blk_size, units);
>> +	pr_warn("expected: '%s', got '%s'\n", exp_result, buf);
>> +}
>> +
>> +static __init void test_string_get_size(void)
>> +{
>> +	test_string_get_size_one(16384, 512, STRING_UNITS_2, "8.00 MiB");
>> +	test_string_get_size_one(500118192, 512, STRING_UNITS_2, "238 GiB");
>> +	test_string_get_size_one(8192, 4096, STRING_UNITS_10, "33.5 MB");
>> +	test_string_get_size_one(1100, 1, STRING_UNITS_10, "1.10 kB");
>> +	test_string_get_size_one(3000, 1900, STRING_UNITS_10, "5.70 MB");
>> +	test_string_get_size_one(151234561234657, 3456789, STRING_UNITS_10,
>> +				 "522 EB");
>
> Since we're changing this anyway, can't we test every pair of
> (size,blk_size) with both units? That'll be twice the number of tests
> for less horizontal real estate. E.g.
>
>   test_string_get_size_one(8192, 4096, "32.0 MiB", "33.5 MB");
>

Nice idea, will do.

> Do we really care how and if string_get_size works for a non-power-of-2
> blk_size? 

Probably not but there is nothing in current algorithm which prevents it
from working correctly with a non-power-of-2 blk_sizes. The issue you
found is even more visible on (3000, 1900) test.

> I certainly assume that we're passed a non-zero value.

Oh crap, we don't have a check for blk_size = 0 and this leads to an
infinite loop now... will do something in v2.

>
> Rasmus

-- 
  Vitaly