Re: [PATCH] btrfs: remove btrfs_fs_info::stripesize

在 2026/6/25 01:03, David Sterba 写道:
> On Wed, Jun 24, 2026 at 01:53:28PM +0930, Qu Wenruo wrote:
>> Btrfs does not support variable stripe length yet, all RAID0/5/6/10
>> chunks have the fixed stripe length 64K for now.
> 
> The question is if the 'for now' will turn into configurable stripe size
> in the future. This was the plan, but long time ago. I think we can
> remove it anyway, the code is no-op and adding proper support would need
> more changes that just that.

I'm already working on the per-chunk stripe length.

But even with that feature, we're going to respect per-chunk stripe 
length, not the global one.

So the existing naming "stripesize" doesn't even make any sense.

> 
>> Furthermore, btrfs_fs_info::stripesize is not the real chunk stripe
>> length, it's always the same value as sectorsize.
>>
>> Remove btrfs_fs_info::stripesize, and for the only callsite utilizing
>> that member, replace it with fs_info->sectorsize instead.
>>
>> Signed-off-by: Qu Wenruo <[email protected]>
> 
> Reviewed-by: David Sterba <[email protected]>
> 
>> --- a/fs/btrfs/extent-tree.c
>> +++ b/fs/btrfs/extent-tree.c
>> @@ -4757,7 +4757,7 @@ static noinline int find_free_extent(struct btrfs_root *root,
>>   
>>   		/* Checks */
>>   		ffe_ctl->search_start = round_up(ffe_ctl->found_offset,
>> -						 fs_info->stripesize);
>> +						 fs_info->sectorsize);
> 
> This seems to be the only relevant use in case the stripe size would not
> be equal to sector size.
> 

I think we can even remove this one. As the extent tree should always 
have everything block aligned.

Thanks,
Qu