On 08/12/2010 10:25 AM, Joel Becker wrote: > On Thu, Jul 15, 2010 at 11:04:37AM +0800, Tao Ma wrote: >> On 07/15/2010 10:31 AM, Joel Becker wrote: >>> On Wed, Jul 14, 2010 at 03:52:26PM +0800, Tao Ma wrote: >>>> When we try to CoW some clusters for a reflinked file, we have to >>>> read the contents first, allocate some clusters and then map these >>>> pages to the new clusters. Currently, we use block_read_full_page to >>>> read it, but it is a little bit slower. >>>> So this patch set try to add readahead support for CoW. Before we >>>> start, we call readahead first so that the pages can be read at the >>>> very first. And during CoW, when we find a readahead page, we know >>>> that we need to move the readahead window, so a new asyncreadahead >>>> is called. >>> >>> I think this looks good. Can you create a branch in your git >>> tree for it? I'll pull when I build merge-window. >> done. please pull from >> git://oss.oracle.com/git/tma/linux-2.6.git cow_readahead. > > I tried pulling this, and it conflicts with the zero-tail fix. > Can you rebase it for me? I want to make sure we get the readahead > right. done. Please pull the same branch. It should work now.
btw, I still have the patch "[PATCH v5] ocfs2: Let ocfs2_setattr use new truncate sequence." waiting for being included. So what's your option? Regards, Tao _______________________________________________ Ocfs2-devel mailing list [email protected] http://oss.oracle.com/mailman/listinfo/ocfs2-devel
