From 5d52423d74994d4f32a58d73fed1dcd9d63ea706 Mon Sep 17 00:00:00 2001
From: Matthew Dillon <dillon@dragonflybsd.org>
Date: Sun, 4 Sep 2005 05:16:59 +0000
Subject: [PATCH] 1.101   Update code comments.

1.100   Update code comments.

1.99    DEFERRED
1.98    DEFERRED

1.97    Timeout while polling is in ms not ticks.  [This should be
        computationally equivalent to the code that was there before, but
        is far simpler].

1.96    DEFERRED - Abort race.  It is unclear whether an xfer can be
                   aborted multiple times.

1.95    DEFERRED (interrupt fixes that may still be buggy)
1.94    DEFERRED (interrupt fixes that may still be buggy)
1.93    DEFERRED (interrupt fixes that may still be buggy)
1.92    DEFERRED (interrupt fixes that may still be buggy)

1.91    NOT APPLICABLE (whitespace cleanup)

1.90    WILL NOT BE APPLIED - This rev tries to setup some sort of
        memory reserve but doesn't really solve any problems.  A
        reserve is useless if it is not operated as a circular
        resource-limited pipeline.

1.89    Don't error out if the number of companion controllers is wrong, just
        complain and go ahead.
---
 sys/bus/usb/ehci.c | 26 ++++++++++++--------------
 1 file changed, 12 insertions(+), 14 deletions(-)

diff --git a/sys/bus/usb/ehci.c b/sys/bus/usb/ehci.c
index 47e6f3fa2e..870b514d1b 100644
--- a/sys/bus/usb/ehci.c
+++ b/sys/bus/usb/ehci.c
@@ -1,7 +1,7 @@
 /*
  * $NetBSD: ehci.c,v 1.67 2004/07/06 04:18:05 mycroft Exp $
  * $FreeBSD: src/sys/dev/usb/ehci.c,v 1.5 2003/11/10 00:20:52 joe Exp $
- * $DragonFly: src/sys/bus/usb/ehci.c,v 1.14 2005/09/04 04:46:12 dillon Exp $
+ * $DragonFly: src/sys/bus/usb/ehci.c,v 1.15 2005/09/04 05:16:59 dillon Exp $
  */
 
 /*
@@ -54,15 +54,11 @@
  * TODO:
  * 1) hold off explorations by companion controllers until ehci has started.
  *
- * 2) The EHCI driver lacks support for interrupt isochronous transfers, so
+ * 2) The EHCI driver lacks support for isochronous transfers, so
  *    devices using them don't work.
- *    Interrupt transfers are not difficult, it's just not done.
  *
- * 3) The meaty part to implement is the support for USB 2.0 hubs.
- *    They are quite compolicated since the need to be able to do
- *    "transaction translation", i.e., converting to/from USB 2 and USB 1.
- *    So the hub driver needs to handle and schedule these things, to
- *    assign place in frame where different devices get to go. See chapter
+ * 3) The hub driver needs to handle and schedule the transaction translator,
+ *    to assign place in frame where different devices get to go. See chapter
  *    on hubs in USB 2.0 for details.
  *
  * 4) command failures are not recovered correctly
@@ -337,6 +333,7 @@ ehci_init(ehci_softc_t *sc)
 	u_int i;
 	usbd_status err;
 	ehci_soft_qh_t *sqh;
+	u_int ncomp;
 
 	DPRINTF(("ehci_init: start\n"));
 #ifdef USB_DEBUG
@@ -352,11 +349,13 @@ ehci_init(ehci_softc_t *sc)
 	sparams = EREAD4(sc, EHCI_HCSPARAMS);
 	DPRINTF(("ehci_init: sparams=0x%x\n", sparams));
 	sc->sc_npcomp = EHCI_HCS_N_PCC(sparams);
-	if (EHCI_HCS_N_CC(sparams) != sc->sc_ncomp) {
+	ncomp = EHCI_HCS_N_CC(sparams);
+	if (ncomp != sc->sc_ncomp) {
 		printf("%s: wrong number of companions (%d != %d)\n",
 		       USBDEVNAME(sc->sc_bus.bdev),
-		       EHCI_HCS_N_CC(sparams), sc->sc_ncomp);
-		return (USBD_IOERROR);
+		       ncomp, sc->sc_ncomp);
+		if (ncomp < sc->sc_ncomp)
+			sc->sc_ncomp = ncomp;
 	}
 	if (sc->sc_ncomp > 0) {
 		printf("%s: companion controller%s, %d port%s each:",
@@ -820,12 +819,11 @@ ehci_idone(struct ehci_xfer *ex)
 void
 ehci_waitintr(ehci_softc_t *sc, usbd_xfer_handle xfer)
 {
-	int timo = xfer->timeout;
-	int usecs;
+	int timo;
 	u_int32_t intrs;
 
 	xfer->status = USBD_IN_PROGRESS;
-	for (usecs = timo * 1000000 / hz; usecs > 0; usecs -= 1000) {
+	for (timo = xfer->timeout; timo >= 0; timo--) {
 		usb_delay_ms(&sc->sc_bus, 1);
 		if (sc->sc_dying)
 			break;
-- 
2.41.0