Re: [PATCH v6 11/11] ext4: run mballoc test with different layouts setting

[Kemeng Shi]

on 9/1/2023 10:36 PM, Ritesh Harjani wrote:
> Kemeng Shi <shikemeng@xxxxxxxxxxxxxxx> writes:
> 
>> Use KUNIT_CASE_PARAM to run mbalaloc test with different layouts setting.
>                               ^^^ mballoc
> small nit below
>   
>>
>> Signed-off-by: Kemeng Shi <shikemeng@xxxxxxxxxxxxxxx>
>> ---
>>  fs/ext4/mballoc-test.c | 52 ++++++++++++++++++++++++++++++------------
>>  1 file changed, 38 insertions(+), 14 deletions(-)
>>
>> diff --git a/fs/ext4/mballoc-test.c b/fs/ext4/mballoc-test.c
>> index d643c56ac003..af48a39c8ba2 100644
>> --- a/fs/ext4/mballoc-test.c
>> +++ b/fs/ext4/mballoc-test.c
>> @@ -196,21 +196,11 @@ static int ext4_mb_mark_context_stub(struct ext4_mark_context *mc,
>>  	return 0;
>>  }
>>  
>> -#define TEST_BLOCKSIZE_BITS 10
>> -#define TEST_CLUSTER_BITS 3
>> -#define TEST_BLOCKS_PER_GROUP 8192
>> -#define TEST_GROUP_COUNT 4
>> -#define TEST_DESC_SIZE 64
>>  #define TEST_GOAL_GROUP 1
>>  static int mbt_kunit_init(struct kunit *test)
>>  {
>> -	struct mbt_ext4_block_layout layout = {
>> -		.blocksize_bits = TEST_BLOCKSIZE_BITS,
>> -		.cluster_bits = TEST_CLUSTER_BITS,
>> -		.blocks_per_group = TEST_BLOCKS_PER_GROUP,
>> -		.group_count = TEST_GROUP_COUNT,
>> -		.desc_size = TEST_DESC_SIZE,
>> -	};
>> +	struct mbt_ext4_block_layout *layout =
>> +		(struct mbt_ext4_block_layout *)(test->param_value);
>>  	struct super_block *sb;
>>  	int ret;
>>  
>> @@ -218,7 +208,7 @@ static int mbt_kunit_init(struct kunit *test)
>>  	if (sb == NULL)
>>  		return -ENOMEM;
>>  
>> -	mbt_init_sb_layout(sb, &layout);
>> +	mbt_init_sb_layout(sb, layout);
>>  
>>  	ret = mbt_ctx_init(sb);
>>  	if (ret != 0) {
>> @@ -304,9 +294,43 @@ static void test_new_blocks_simple(struct kunit *test)
>>  		"unexpectedly get block when no block is available");
>>  }
>>  
>> +static const struct mbt_ext4_block_layout mbt_test_layouts[] = {
>> +	{
>> +		.blocksize_bits = 10,
>> +		.cluster_bits = 3,
>> +		.blocks_per_group = 8192,
>> +		.group_count = 4,
>> +		.desc_size = 64,
>> +	},
>> +	{
>> +		.blocksize_bits = 12,
>> +		.cluster_bits = 3,
>> +		.blocks_per_group = 8192,
>> +		.group_count = 4,
>> +		.desc_size = 64,
>> +	},
>> +	{
>> +		.blocksize_bits = 18,
> 
> 64k blocksize is more common due to platforms with 64k pagesize like
> Power and sometimes arm64. I would rather make it 16 here.
> 
> I tested it on Power - 
Sure, I will make it 16 in next version. Thanks!

> 
> [    2.546687][    T1] KTAP version 1
> [    2.547123][    T1] 1..2
> [    2.547447][    T1]     KTAP version 1
> [    2.547927][    T1]     # Subtest: ext4_mballoc_test
> [    2.548562][    T1]     1..1
> [    2.548933][    T1]         KTAP version 1
> [    2.549457][    T1]         # Subtest: test_new_blocks_simple
> [    2.549550][  T108] kunit_try_catch (108) used greatest stack depth: 14512 bytes left
> [    2.549644][    T1]         ok 1 block_bits=10 cluster_bits=3 blocks_per_group=8192 group_count=4 desc_size=64
> [    2.552780][  T110] kunit_try_catch (110) used greatest stack depth: 14464 bytes left
> [    2.552882][    T1]         ok 2 block_bits=12 cluster_bits=3 blocks_per_group=8192 group_count=4 desc_size=64
> [    2.555909][    T1]         ok 3 block_bits=18 cluster_bits=3 blocks_per_group=8192 group_count=4 desc_size=64
> [    2.557184][    T1]     # test_new_blocks_simple: pass:3 fail:0 skip:0 total:3
> [    2.557186][    T1]     ok 1 test_new_blocks_simple
> [    2.558083][    T1] # Totals: pass:3 fail:0 skip:0 total:3
> [    2.558688][    T1] ok 1 ext4_mballoc_test
> 
> Looks good to me. Feel free to add -
> 
> Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@xxxxxxxxx>
>