svn commit: r226529 - user/adrian/if_ath_tx/sys/dev/ath
Adrian Chadd
adrian at FreeBSD.org
Wed Oct 19 08:52:15 UTC 2011
Author: adrian
Date: Wed Oct 19 08:52:14 2011
New Revision: 226529
URL: http://svn.freebsd.org/changeset/base/226529
Log:
Remove unneeded code and comment. The race has been fixed
by placing the relevant bits behind ATH_LOCK.
Modified:
user/adrian/if_ath_tx/sys/dev/ath/if_ath.c
Modified: user/adrian/if_ath_tx/sys/dev/ath/if_ath.c
==============================================================================
--- user/adrian/if_ath_tx/sys/dev/ath/if_ath.c Wed Oct 19 08:39:44 2011 (r226528)
+++ user/adrian/if_ath_tx/sys/dev/ath/if_ath.c Wed Oct 19 08:52:14 2011 (r226529)
@@ -4130,38 +4130,11 @@ rx_next:
* need to be handled, kick the PCU if there's
* been an RXEOL condition.
*/
- /*
- * XXX TODO!
- * It is very likely that we're unfortunately
- * racing with other places where ath_hal_intrset()
- * may be called. It may be that we do need to
- * add some more locking (eg the pcu lock from ath9k/
- * reference), or introduce some other way to cope
- * with this.
- */
if (resched && sc->sc_kickpcu) {
CTR0(ATH_KTR_ERR, "ath_rx_proc: kickpcu");
device_printf(sc->sc_dev, "%s: kickpcu; handled %d packets\n",
__func__, npkts);
-#if 0
- /*
- * This re-links all of the descriptors together.
- * (Is it possible that somehow, some state/issue
- * is leaving us with badly linked descriptors?)
- * Is it possible that we're receiving another RXEOL
- * _during_ this function?
- */
- if (ath_startrecv(sc) != 0) {
- if_printf(ifp,
- "%s: couldn't restart RX after RXEOL; resetting\n",
- __func__);
- ath_reset(ifp);
- sc->sc_kickpcu = 0;
- return;
- }
-#endif
-
/* XXX rxslink? */
bf = TAILQ_FIRST(&sc->sc_rxbuf);
ath_hal_putrxbuf(ah, bf->bf_daddr);
More information about the svn-src-user
mailing list