Thread (17 messages) 17 messages, 4 authors, 2005-03-31

Re: cls_u32 compile failure in current 2.6.12-rc1+BK tree

From: Thomas Graf <tgraf@suug.ch>
Date: 2005-03-30 01:02:09

* jamal [ref] 2005-03-29 15:51
In general (from 1 mile away looks fine).
One issue: a->type could be used for in the future to carry
other things in addition to backward compatibility. So you cant assume
it being set means only backward compat is needed (and nothing else).
Not sure if I get you but maybe a mistake of mine while mapping
the logic in mind into code has confused you. The code really
should have been:

	/* compat_mode being true specifies a call that is supposed
	 * to add additional backward compatiblity statistic TLVs.
	 */
	if (compat_mode) {
		if (a->type == TCA_OLD_COMPAT)
			err = gnet_stats_start_copy_compat(skb, 0,
				TCA_STATS, TCA_XSTATS, h->stats_lock, &d);
	} else
		err = gnet_stats_start_copy(skb, TCA_ACT_STATS,
			h->stats_lock, &d);

So we always dump TCA_ACT_STATS when calling from tcf_exts_dump
and add the backward compatibility TLVs in the second call
comming from tcf_exts_dump_stats for actions that need it.

Patches tested, works for me but might be worth for someone
else to give it a try as well just to make sure. Meelis, could
you give the 3 attched patches a try?

Dave, you can pull from bk://kernel.bkbits.net/tgraf/net-2.6-tcf_exts
if everyone can agree on this.

Attachments

Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help