Re: [PATCH 08/10] btrfs: simplify data stripe calculation helpers

[Qu Wenruo <quwenruo.btrfs@gmx.com>]

On 2021/7/27 下午4:39, David Sterba wrote:
> On Tue, Jul 27, 2021 at 06:23:03AM +0800, Qu Wenruo wrote:
>>>>> --- a/fs/btrfs/volumes.c
>>>>> +++ b/fs/btrfs/volumes.c
>>>>> @@ -3567,10 +3567,7 @@ static u64 calc_data_stripes(u64 type, int num_stripes)
>>>>>     	const int ncopies = btrfs_raid_array[index].ncopies;
>>>>>     	const int nparity = btrfs_raid_array[index].nparity;
>>>>>
>>>>> -	if (nparity)
>>>>> -		return num_stripes - nparity;
>>>>> -	else
>>>>> -		return num_stripes / ncopies;
>>>>
>>>> I would prefer an ASSERT() here to be extra sure.
>>>> But it's my personal taste (and love for tons of ASSERT()).
>>>
>>> Assert for what exactly? I had a thought about that too but was not sure
>>> what to put there.
>>>
>>
>> To ensure we have either non-zero nparity with 1 ncopy or zero nparity
>> with ncopies > 1.
>
> Yeah but that's statically defined in the raid table, it does not change
> so we don't have to verify it on each call. If anything, such
> constraints could be verified as _static_assert right after the table
> but IMO that's pointless and definitely not a runtime check.
>
OK, that makes sense.

Reviewed-by: Qu Wenruo <wqu@suse.com>

Thanks,
Qu