[src/trunk]: src/sys/dev/acpi acpiec(4): Take a lock around acpiec_cold updates.

[riastradh <riastradh%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/63b321f0e23a
branches:  trunk
changeset: 377599:63b321f0e23a
user:      riastradh <riastradh%NetBSD.org@localhost>
date:      Tue Jul 18 10:17:02 2023 +0000

description:
acpiec(4): Take a lock around acpiec_cold updates.

Otherwise we race with readers -- probably harmlessly, but let's
avoid the appearance of problems.

XXX Maybe acpiec_suspend and acpiec_shutdown should interrupt
transactions and force them to fail promptly?

XXX This looks bad because acpiec_cold is global and sc->sc_mtx
doesn't look like it's global, but we expect to have only one
acpiec(4) device anyway from what I understand.  Maybe we should move
acpiec_cold into the softc?

diffstat:

 sys/dev/acpi/acpi_ec.c |  26 ++++++++++++++++++++++++--
 1 files changed, 24 insertions(+), 2 deletions(-)

diffs (67 lines):

diff -r 841a1d248b59 -r 63b321f0e23a sys/dev/acpi/acpi_ec.c
--- a/sys/dev/acpi/acpi_ec.c    Tue Jul 18 10:10:49 2023 +0000
+++ b/sys/dev/acpi/acpi_ec.c    Tue Jul 18 10:17:02 2023 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: acpi_ec.c,v 1.106 2023/07/18 10:10:49 riastradh Exp $  */
+/*     $NetBSD: acpi_ec.c,v 1.107 2023/07/18 10:17:02 riastradh Exp $  */
 
 /*-
  * Copyright (c) 2007 Joerg Sonnenberger <joerg%NetBSD.org@localhost>.
@@ -57,7 +57,7 @@
  */
 
 #include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: acpi_ec.c,v 1.106 2023/07/18 10:10:49 riastradh Exp $");
+__KERNEL_RCSID(0, "$NetBSD: acpi_ec.c,v 1.107 2023/07/18 10:17:02 riastradh Exp $");
 
 #ifdef _KERNEL_OPT
 #include "opt_acpi_ec.h"
@@ -477,8 +477,23 @@ post_data_map:
 static bool
 acpiec_suspend(device_t dv, const pmf_qual_t *qual)
 {
+       struct acpiec_softc *sc = device_private(dv);
 
+       /*
+        * XXX This looks bad because acpiec_cold is global and
+        * sc->sc_mtx doesn't look like it's global, but we can have
+        * only one acpiec(4) device anyway.  Maybe acpiec_cold should
+        * live in the softc to make this look less bad?
+        *
+        * XXX Should this block read/write/query transactions until
+        * resume?
+        *
+        * XXX Should this interrupt existing transactions to make them
+        * fail promptly or restart on resume?
+        */
+       mutex_enter(&sc->sc_mtx);
        acpiec_cold = true;
+       mutex_exit(&sc->sc_mtx);
 
        return true;
 }
@@ -486,8 +501,11 @@ acpiec_suspend(device_t dv, const pmf_qu
 static bool
 acpiec_resume(device_t dv, const pmf_qual_t *qual)
 {
+       struct acpiec_softc *sc = device_private(dv);
 
+       mutex_enter(&sc->sc_mtx);
        acpiec_cold = false;
+       mutex_exit(&sc->sc_mtx);
 
        return true;
 }
@@ -495,8 +513,12 @@ acpiec_resume(device_t dv, const pmf_qua
 static bool
 acpiec_shutdown(device_t dv, int how)
 {
+       struct acpiec_softc *sc = device_private(dv);
 
+       mutex_enter(&sc->sc_mtx);
        acpiec_cold = true;
+       mutex_exit(&sc->sc_mtx);
+
        return true;
 }