From: Nicholas Marriott (nicm__ntlworld.com)
Date: Mon Aug 04 2008 - 15:15:16 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

This works for me too. Without patch:

$ apm
Battery state: high, 100% remaining, unknown life estimate
A/C adapter state: connected
Performance adjustment mode: cool running (1000 MHz)
$ for i in `jot 100`; do apm|head -1; sleep 0.$RANDOM; done|grep -v high
Battery state: absent, 0% remaining, unknown life estimate
Battery state: absent, 0% remaining, unknown life estimate
Battery state: absent, 0% remaining, unknown life estimate
Battery state: CRITICAL, 0% remaining, unknown life estimate
Battery state: absent, 0% remaining, unknown life estimate
Battery state: CRITICAL, 0% remaining, unknown life estimate
$

With patch:

$ apm
Battery state: high, 100% remaining, unknown life estimate
A/C adapter state: connected
Performance adjustment mode: cool running (1000 MHz)
$ for i in `jot 100`; do apm|head -1; sleep 0.$RANDOM; done|grep -v high
$

-- Nicholas

On Mon, Aug 04, 2008 at 01:01:44PM -0500, Marco Peereboom wrote:
> Stefan, my apologies, you were right and so was your patch.
>
> Slightly modified patch.
>
> Index: acpibat.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/acpi/acpibat.c,v
> retrieving revision 1.50
> diff -u -p -u -p -r1.50 acpibat.c
> --- acpibat.c 13 Jun 2008 05:50:21 -0000 1.50
> +++ acpibat.c 4 Aug 2008 18:01:33 -0000
> -292,9 +292,10 acpibat_getbif(struct acpibat_softc *sc)
> struct aml_value res;
> int rv = EINVAL;
>
> - memset(&sc->sc_bif, 0, sizeof(sc->sc_bif));
> - if (!sc->sc_bat_present)
> - return 0;
> + if (!sc->sc_bat_present) {
> + memset(&sc->sc_bif, 0, sizeof(sc->sc_bif));
> + return (0);
> + }
>
> if (aml_evalname(sc->sc_acpi, sc->sc_devnode, "_BIF", 0, NULL, &res)) {
> dnprintf(10, "%s: no _BIF\n", DEVNAME(sc));
> -355,9 +356,10 acpibat_getbst(struct acpibat_softc *sc)
> struct aml_value res;
> int rv = EINVAL;
>
> - memset(&sc->sc_bst, 0, sizeof(sc->sc_bst));
> - if (!sc->sc_bat_present)
> - return 0;
> + if (!sc->sc_bat_present) {
> + memset(&sc->sc_bst, 0, sizeof(sc->sc_bst));
> + return (0);
> + }
>
> if (aml_evalname(sc->sc_acpi, sc->sc_devnode, "_BST", 0, NULL, &res)) {
> dnprintf(10, "%s: no _BST\n", DEVNAME(sc));
> -400,7 +402,8 int
> acpibat_notify(struct aml_node *node, int notify_type, void *arg)
> {
> struct acpibat_softc *sc = arg;
> - struct aml_value res;
> + struct aml_value res;
> + int present
>
> dnprintf(10, "acpibat_notify: %.2x %s\n", notify_type,
> sc->sc_devnode->name);
> -408,8 +411,7 acpibat_notify(struct aml_node *node, in
> /* Check if installed state of battery has changed */
> memset(&res, 0, sizeof(res));
> if (aml_evalname(sc->sc_acpi, node, "_STA", 0, NULL, &res) == 0) {
> - int present = (res.v_integer & STA_BATTERY);
> -
> + present = res.v_integer & STA_BATTERY;
> if (!sc->sc_bat_present && present) {
> printf("%s: %s inserted\n", DEVNAME(sc),
> sc->sc_devnode->name);
>
> On Mon, Aug 04, 2008 at 12:22:27PM -0500, Marco Peereboom wrote:
> > Ok I misdiagnosed the issue.
> >
> > My bad let me restart from scratch.
> >
> > On Mon, Aug 04, 2008 at 06:59:45PM +0200, Marco Pfatschbacher wrote:
> > > On Mon, Aug 04, 2008 at 10:57:34AM -0500, Marco Peereboom wrote:
> > > > Since we have always used NOPOLL and put POLL in there to work around
> > > > stupid systems I'll keep the current code as is. Release is getting
> > > > closer and this is not important enough to risk. I will revisit this
> > > > after release.
> > >
> > > Why not just backout acpibat.c -r 1.50 ?
> > > Or maybe reconsider Stefan's first patch.
> > > No ACPIDEV_POLL changes involved and it's enough to fix things.
> > >
> > >
> > > Marco

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]