i386: i8254_initclocks: set IPL_MPSAFE for clock/rtc IRQs

Setting IPL_MPSAFE for the i8254/mc146818 IRQs appeases a KASSERT in
apic_intr_establish() and allows the system to boot via the i8254
path.  This makes testing changes to the i8254/mc146818 code much
easier on modern hardware without mucking with the GENERIC config.  We
already set IPL_MPSAFE for these IRQs in the equivalent amd64 code.

Now, setting IPL_MPSAFE is a lie: the i8254 and mc146818 IRQs are not
MP-safe.  However, the lie is harmless because we only reach
i8254_initclocks() if (a) there is no APIC at all, or (b) we fail to
calibrate the local APIC timer.

Thread: https://marc.info/?l=openbsd-tech&m=169258915227321&w=2

ok mlarkin@

diff --git a/sys/arch/i386/isa/clock.c b/sys/arch/i386/isa/clock.c
index 48d9675..563feaa 100644 (file)
--- a/sys/arch/i386/isa/clock.c
+++ b/sys/arch/i386/isa/clock.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: clock.c,v 1.65 2023/07/25 18:16:20 cheloha Exp $      */
+/*     $OpenBSD: clock.c,v 1.66 2023/08/22 17:13:22 cheloha Exp $      */
 /*     $NetBSD: clock.c,v 1.39 1996/05/12 23:11:54 mycroft Exp $       */
 
 /*-
@@ -430,10 +430,16 @@ i8254_initclocks(void)
 
        clockintr_cpu_init(NULL);
 
-       /* When using i8254 for clock, we also use the rtc for profclock */
-       (void)isa_intr_establish(NULL, 0, IST_PULSE, IPL_CLOCK,
+       /*
+        * When using i8254 for clock, we also use the rtc for profclock.
+        *
+        * These IRQs are not MP-safe, but it is harmless to lie about it
+        * because we cannot reach this point unless we are only booting
+        * a single CPU.
+        */
+       (void)isa_intr_establish(NULL, 0, IST_PULSE, IPL_CLOCK | IPL_MPSAFE,
            clockintr, 0, "clock");
-       (void)isa_intr_establish(NULL, 8, IST_PULSE, IPL_STATCLOCK,
+       (void)isa_intr_establish(NULL, 8, IST_PULSE, IPL_STATCLOCK | IPL_MPSAFE,
            rtcintr, 0, "rtc");
 
        rtcstart();                     /* start the mc146818 clock */