Hi Christoph

On 4 March 2012 13:32, Saugata Das [off-list ref] wrote:

On 3 March 2012 15:04, Christoph Hellwig [off-list ref] wrote:

quoted

On Fri, Mar 02, 2012 at 06:15:09PM +0530, Saugata Das wrote:

quoted

--- a/fs/buffer.c
+++ b/fs/buffer.c

@@ -2926,6 +2926,9 @@ int submit_bh(int rw, struct buffer_head * bh)

      if (test_set_buffer_req(bh) && (rw & WRITE))
              clear_buffer_write_io_error(bh);

+     if (buffer_meta(bh))
+             rw |= REQ_META;
+

submit_bh has a perfectly valid rw argument where you can pass REQ_META.
Please use that instead of abusing a buffer state bit for an I/O
submission flag.

Typically, during meta-data writes, the buffer is only marked dirty
(__ext4_handle_dirty_metadata). The submit_bh of the dirty meta-data
buffer happens later from writeback thread which in general flushes
out all the buffers. From this generic handling of dirty buffers from
writeback thread via generic writepage functions till submit_bh, it is
not possible to identify whether the buffer was corresponding to any
meta-data without having an extra information in the buffer itself.

Is the above explanation OK for you ? Or, do you some alternative
suggestion for me to explore ?

Based on your feedback, I shall resubmit the patch on top of 3.4.


regards
saugata
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html