[Kittyhawk] digging deeper

Eric Van Hensbergen ericvh at gmail.com
Sat Sep 17 19:24:53 EDT 2011


my bad, its not the printk, its thr flush (of course):

0xc01d83f4 is in bgtree_inject_interrupt (drivers/bluegene/tree.c:840).
835	    for (chn = 0; chn < BGP_MAX_CHANNEL; chn++) {
836		if (pend_irq & tree->inj_wm_mask & tree->chn[chn].irq_inj_pending_mask) {
837		    struct bglink_proto *proto;
838		    list_for_each_entry_rcu(proto, &linkproto_list, list)
839			if (proto->tree_flush)
840			    if ((rc = proto->tree_flush(chn)) != 0)
841				break;
842	
843		    /* If we successfully flushed all packets, clear the high
844		     * watermark.  We are still holding tree->lock and thus no new

will dig a bit deeper after dinner.


More information about the Kittyhawk mailing list