Re: [PATCH v2 10/18] mm: teach truncate_inode_pages_range() to handle non page aligned ranges
From: Lukáš Czerner <hidden>
Date: 2013-02-21 08:33:56
Also in:
linux-fsdevel, linux-mm, lkml
From: Lukáš Czerner <hidden>
Date: 2013-02-21 08:33:56
Also in:
linux-fsdevel, linux-mm, lkml
On Fri, 8 Feb 2013, Lukáš Czerner wrote:
Date: Fri, 8 Feb 2013 10:08:05 +0100 (CET)
From: Lukáš Czerner <redacted>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Lukas Czerner <redacted>, linux-mm@kvack.org,
linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-ext4@vger.kernel.org, Hugh Dickins [off-list ref]
Subject: Re: [PATCH v2 10/18] mm: teach truncate_inode_pages_range() to handle
non page aligned ranges..snip..
quoted
quoted
+ /* + * 'start' and 'end' always covers the range of pages to be fully + * truncated. Partial pages are covered with 'partial_start' at the + * start of the range and 'partial_end' at the end of the range. + * Note that 'end' is exclusive while 'lend' is inclusive. + */That helped ;) So the bytes to be truncated are (start*PAGE_SIZE + partial_start) -> (end*PAGE_SIZE + partial_end) - 1 yes?The start of the range is not right, because 'start' and 'end' covers pages to be _fully_ truncated. See the while cycle and then 'if (partial_start)' condition where we search for the page (start - 1) and do_invalidate within that page. So it should be like this: (start*PAGE_SIZE - partial_start*(PAGE_SIZE - partial_start) -> (end*PAGE_END + partial_end) - 1 assuming that you want the range to be inclusive on both sides. -Lukas
Hi Andrew, what's the status of the patch set ? Do you have any more comments or questions ? Can we get this in in this merge window ? Thanks! -Lukas