JFFS2 read Inode related issue

[Mouawad, Tony]

Hello,

I'm using Linux 2.6.10 on Coldfire architecture and I seem to be
experiencing very intermittent problems reading inodes.  It seems that
the following case is being hit during a call to jffs2_do_read_inode():

case INO_STATE_READING:
case INO_STATE_PRESENT:
	/* Eep. This should never happen. It can
	happen if Linux calls read_inode() again
	before clear_inode() has finished though. */
	printk(KERN_WARNING "Eep. Trying to read_inode #%u when it's
already 		in state %d!\n", ino, f->inocache->state);
	/* Fail. That's probably better than allowing it to succeed */
	f->inocache = NULL;
	break;

I notice that the comment states "This should never happen" but
apparently it is happening to me.  Does anyone have ideas on how I
should go about mitigating this problem?  

Cheers,
Tony