On Fri, 08 Jul 2016 at 18:51:17 -0500, joshua stein wrote:
> If the EC fails to go into burst mode for whatever reason, the Burst
> Acknowledge byte will not be there to read, which means the status
> won't have EC_STAT_OBF, which means acpiec_wait will spin forever,
> hanging the machine.
>
> This at least gets us moving again, ignoring the failure to enter
> burst mode.
That patch was put into the latest snapshot (but not yet in CVS) and
some people are seeing fallout from it like strange thermal and
battery status readings.
This is a different patch which just doesn't bother with burst mode
unless the transfer is big. This should match how Linux and FreeBSD
behave. I have at least one confirmation that this fixes things on
the ThinkPad X260.
Anyone else that is seeing strangeness on the latest snapshot, I ask
that you try this patch against -current and see if your issues go
away.
Index: sys/dev/acpi/acpiec.c
===================================================================
RCS file: /cvs/src/sys/dev/acpi/acpiec.c,v
retrieving revision 1.53
diff -u -p -u -r1.53 acpiec.c
--- sys/dev/acpi/acpiec.c 7 May 2016 18:03:36 -0000 1.53
+++ sys/dev/acpi/acpiec.c 11 Aug 2016 20:06:54 -0000
@@ -218,10 +218,12 @@ acpiec_read(struct acpiec_softc *sc, u_i
*/
dnprintf(20, "%s: read %d, %d\n", DEVNAME(sc), (int)addr, len);
sc->sc_ecbusy = 1;
- acpiec_burst_enable(sc);
+ if (len > 1)
+ acpiec_burst_enable(sc);
for (reg = 0; reg < len; reg++)
buffer[reg] = acpiec_read_1(sc, addr + reg);
- acpiec_burst_disable(sc);
+ if (len > 1)
+ acpiec_burst_disable(sc);
sc->sc_ecbusy = 0;
}
@@ -237,10 +239,12 @@ acpiec_write(struct acpiec_softc *sc, u_
*/
dnprintf(20, "%s: write %d, %d\n", DEVNAME(sc), (int)addr, len);
sc->sc_ecbusy = 1;
- acpiec_burst_enable(sc);
+ if (len > 1)
+ acpiec_burst_enable(sc);
for (reg = 0; reg < len; reg++)
acpiec_write_1(sc, addr + reg, buffer[reg]);
- acpiec_burst_disable(sc);
+ if (len > 1)
+ acpiec_burst_disable(sc);
sc->sc_ecbusy = 0;
}
