xfs
[Top] [All Lists]

Re: [PATCH v7 2/4] xfs: Introduce a helper routine to probe data or hole

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: [PATCH v7 2/4] xfs: Introduce a helper routine to probe data or hole offset from page cache
From: Jie Liu <jeff.liu@xxxxxxxxxx>
Date: Tue, 21 Aug 2012 13:04:50 +0800
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20120820231309.GJ19235@dastard>
References: <5028FC2E.2010802@xxxxxxxxxx> <20120820231309.GJ19235@dastard>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:14.0) Gecko/20120714 Thunderbird/14.0
On 08/21/12 07:13, Dave Chinner wrote:
> On Mon, Aug 13, 2012 at 09:07:58PM +0800, Jeff Liu wrote:
>> helper routine to lookup data or hole offset from page cache for unwritten 
>> extents.
>>
>> Signed-off-by: Jie Liu <jeff.liu@xxxxxxxxxx>
> ....
>> +            unsigned int    i;
>> +
>> +            want = min_t(pgoff_t, end - index, (pgoff_t)PAGEVEC_SIZE);
> No need for the cast to (pgoff_t) here. That's what the min_t()
> macro does for you.
Thanks for pointing this out.
>
>> +            for (i = 0; i < nr_pages; i++) {
>> +                    struct page     *page = pvec.pages[i];
>> +                    loff_t          b_offset;
>> +
>> +                    /*
>> +                     * Page index is out of range, searching done.
>> +                     * If the current offset is not reaches the end
>> +                     * of the specified search range, there should
>> +                     * be a hole between them.
>> +                     */
>> +                    if (page->index > end) {
>> +                            if (type == HOLE_OFF && lastoff < endoff) {
>> +                                    *offset = lastoff;
>> +                                    found = true;
>> +                            }
>> +                            goto out;
>> +                    }
>> +
>> +                    lock_page(page);
>> +                    /*
>> +                     * Page truncated or invalidated(page->mapping == NULL).
>> +                     * We can freely skip it and proceed to check the next
>> +                     * page.
>> +                     */
>> +                    if (unlikely(page->mapping != inode->i_mapping)) {
>> +                            unlock_page(page);
>> +                            continue;
>> +                    }
> Still need to check page->index again after locking the page. As
> this isn't performance critical, I don't think you need the check
> before the page is locked, anyway....
I understand that we have to check page->mapping with page locked
because race against inode truncation.
But could you teaching me why page->index will be changed in either
memory reclaim or truncates even if
the page returned with refcount > 0?

Thanks,
-Jeff
>
> Cheers,
>
> Dave.

<Prev in Thread] Current Thread [Next in Thread>