Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Totally revamp device resets using clue from the reference driver.

Browse Source
This commit is contained in:
dyoung 2004-07-15 07:25:40 +00:00
parent d6014cf055
commit 9ad89f8c4a
1 changed files with 39 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
sys/dev/ic/atw.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: atw.c,v 1.62 2004/07/15 07:25:06 dyoung Exp $ */
/* $NetBSD: atw.c,v 1.63 2004/07/15 07:25:40 dyoung Exp $ */
/*-
* Copyright (c) 1998, 1999, 2000, 2002, 2003, 2004 The NetBSD Foundation, Inc.
@ -41,7 +41,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: atw.c,v 1.62 2004/07/15 07:25:06 dyoung Exp $");
__KERNEL_RCSID(0, "$NetBSD: atw.c,v 1.63 2004/07/15 07:25:40 dyoung Exp $");
#include "bpfilter.h"
@ -881,34 +881,65 @@ void
atw_reset(struct atw_softc *sc)
{
int i;
uint32_t lpc;
ATW_WRITE(sc, ATW_NAR, 0x0);
DELAY(20 * 1000);
/* Reference driver has a cryptic remark indicating that this might
* power-on the chip. I know that it turns off power-saving....
*/
ATW_WRITE(sc, ATW_FRCTL, 0x0);
ATW_WRITE(sc, ATW_PAR, ATW_PAR_SWR);
for (i = 0; i < 10000; i++) {
if (ATW_ISSET(sc, ATW_PAR, ATW_PAR_SWR) == 0)
for (i = 0; i < 50; i++) {
if (ATW_READ(sc, ATW_PAR) == 0)
break;
DELAY(1);
DELAY(1000);
}
/* ... and then pause 100ms longer for good measure. */
DELAY(100 * 1000);
DPRINTF2(sc, ("%s: atw_reset %d iterations\n", sc->sc_dev.dv_xname, i));
if (ATW_ISSET(sc, ATW_PAR, ATW_PAR_SWR))
printf("%s: reset failed to complete\n", sc->sc_dev.dv_xname);
/* Turn off maximum power saving. */
ATW_CLR(sc, ATW_FRCTL, ATW_FRCTL_MAXPSP);
/*
* Initialize the PCI Access Register.
*/
sc->sc_busmode = ATW_PAR_PBL_8DW;
ATW_WRITE(sc, ATW_PAR, sc->sc_busmode);
DPRINTF(sc, ("%s: ATW_PAR %08x busmode %08x\n", sc->sc_dev.dv_xname,
ATW_READ(sc, ATW_PAR), sc->sc_busmode));
/* Turn off maximum power saving, etc.
*
* XXX Following example of reference driver, should I set
* an AID of 1? It didn't seem to help....
*/
ATW_WRITE(sc, ATW_FRCTL, 0x0);
DELAY(100 * 1000);
/* Recall EEPROM. */
ATW_SET(sc, ATW_TEST0, ATW_TEST0_EPRLD);
DELAY(10 * 1000);
lpc = ATW_READ(sc, ATW_LPC);
DPRINTF(sc, ("%s: ATW_LPC %#08x\n", __func__, lpc));
/* A reset seems to affect the SRAM contents, so put them into
* a known state.
*/
atw_clear_sram(sc);
memset(sc->sc_bssid, 0, sizeof(sc->sc_bssid));
memset(sc->sc_bssid, 0xff, sizeof(sc->sc_bssid));
}
static void