Re: [PATCH v2 09/18] btrfs: subpage: introduce helper for subpage uptodate status

[Nikolay Borisov <nborisov@suse.com>]

On 11.12.20 г. 12:48 ч., Qu Wenruo wrote:
> 
> 
> On 2020/12/11 下午6:10, Nikolay Borisov wrote:
>>
>>
>> On 10.12.20 г. 8:38 ч., Qu Wenruo wrote:
>>> This patch introduce the following functions to handle btrfs subpage
>>> uptodate status:
>>> - btrfs_subpage_set_uptodate()
>>> - btrfs_subpage_clear_uptodate()
>>> - btrfs_subpage_test_uptodate()
>>>   Those helpers can only be called when the range is ensured to be
>>>   inside the page.
>>>
>>> - btrfs_page_set_uptodate()
>>> - btrfs_page_clear_uptodate()
>>> - btrfs_page_test_uptodate()
>>>   Those helpers can handle both regular sector size and subpage without
>>>   problem.
>>>
>>> Signed-off-by: Qu Wenruo <wqu@suse.com>
>>> ---
>>>  fs/btrfs/subpage.h | 98 ++++++++++++++++++++++++++++++++++++++++++++++
>>>  1 file changed, 98 insertions(+)
>>>
>>> diff --git a/fs/btrfs/subpage.h b/fs/btrfs/subpage.h
>>> index 87b4e028ae18..b3cf9171ec98 100644
>>> --- a/fs/btrfs/subpage.h
>>> +++ b/fs/btrfs/subpage.h
>>> @@ -23,6 +23,7 @@
>>>  struct btrfs_subpage {
>>>  	/* Common members for both data and metadata pages */
>>>  	spinlock_t lock;
>>> +	u16 uptodate_bitmap;
>>>  	union {
>>>  		/* Structures only used by metadata */
>>>  		struct {
>>> @@ -35,6 +36,17 @@ struct btrfs_subpage {
>>>  int btrfs_attach_subpage(struct btrfs_fs_info *fs_info, struct page *page);
>>>  void btrfs_detach_subpage(struct btrfs_fs_info *fs_info, struct page *page);
>>>
>>> +static inline void btrfs_subpage_clamp_range(struct page *page,
>>> +					     u64 *start, u32 *len)
>>> +{
>>> +	u64 orig_start = *start;
>>> +	u32 orig_len = *len;
>>> +
>>> +	*start = max_t(u64, page_offset(page), orig_start);
>>> +	*len = min_t(u64, page_offset(page) + PAGE_SIZE,
>>> +		     orig_start + orig_len) - *start;
>>> +}
>>
>> This handles EB's which span pages, right? If so - a comment is in order
>> since there is no design document specifying whether eb can or cannot
>> span multiple pages.
> 
> Didn't I have already stated that in the subpage eb accessors patch?
> 
> No subpage eb can across page bounday.
> 

As just discussed during the whiteboard session this function is really
dead code for eb's because they are guaranteed to not span pages. Even
for RW support it seems there is only btrfs_dirty_pages which changes
page flags without having clamped the data i.e. there's only 1
exception. In light of this I think it would be better to replace this
function with ASSERTS and handle the only exception at the call site.


<snip>