Re: [PATCH] ext4: Resolve the hang of direct i/o read in handling EXT4_IO_END_UNWRITTEN.

On Tue, Aug 16, 2011 at 02:37:53PM +0800, Tao Ma wrote:

From: Tao Ma <redacted>

EXT4_IO_END_UNWRITTEN flag set and the increase of i_aiodio_unwritten should
be done simultaneously since ext4_end_io_nolock always clear the flag and
decrease the counter in the same time.

We don't increase i_aiodio_unwritten when setting EXT4_IO_END_UNWRITTEN so
it will go nagative and causes some process to wait forever.

Part of the patch came from Eric in his e-mail, but it doesn't fix the problem
met by Michael actually.
http://marc.info/?l=linux-ext4&m=131316851417460&w=2

Reported-and-Tested-by: Michael Tokarev<redacted>
Signed-off-by: Eric Sandeen <redacted>
Signed-off-by: Tao Ma <redacted>

Thanks I've taken this into the ext4 tree.  I am a bit worried this
will trigger a GCC warning:

+	/*
+	 * It may be over-defensive here to check EXT4_IO_END_UNWRITTEN now,
+	 * but being more careful is always safe for the future change.
+	 */
 	inode = io_end->inode;
+	if (!(io_end->flag & EXT4_IO_END_UNWRITTEN)) {
+		io_end->flag |= EXT4_IO_END_UNWRITTEN;
+		atomic_inc(&EXT4_I(inode)->i_aiodio_unwritten);
+	}
 
 	/* Add the io_end to per-inode completed io list*/
 	spin_lock_irqsave(&EXT4_I(inode)->i_completed_io_lock, flags);

... since in Google we've been compiling with -Werror, but it's not
causing an error on gcc 4.4, which is what I still have on my laptop.
It may be that newer versions of GCC are smart enough to notice tha
the above is dead code, and then complain with a warning.

        	     	  	    	 - Ted