summaryrefslogtreecommitdiffstats
path: root/sys/isa/atrtc.c
diff options
context:
space:
mode:
authorbde <bde@FreeBSD.org>2003-09-07 14:23:08 +0000
committerbde <bde@FreeBSD.org>2003-09-07 14:23:08 +0000
commitbc4334bd197dcc51c53c30a538178bc696984a1b (patch)
tree55d005beb572003839ca45eab12db7da3884ae7c /sys/isa/atrtc.c
parent34673ebfc37763889cbbdea9a6682a0ec1a559e1 (diff)
downloadFreeBSD-src-bc4334bd197dcc51c53c30a538178bc696984a1b.zip
FreeBSD-src-bc4334bd197dcc51c53c30a538178bc696984a1b.tar.gz
clock.c:
Quick fix for calling DELAY() for ddb input in some (atkbd-based) console drivers. ddb must not use any normal locks, but DELAY() normally calls getit() which needs clock_lock. One problem with using normal locks in ddb is that deadlock is possible, but deadlock on clock_lock is unlikely becaluse clock_lock is bogusly recursive, apparently just to hide the problem of ddb using it. The i8254 clock hardware has mostly write-only registers so it is important for it to use a lock that gives exclusive access. (atkbd hardware is also unfriendly to reentrant software but that problem is more local and already solved.) I mostly saw the symptoms of the bug caused by unlocking in getit() running cpu_unpend(). cpu_unpend() should not be called while in ddb and Debugger() calls for failing assertions about this caused a breakpoint within ddb. ddb must also not call getit() because ddb may be being used to step through clock initialization code that has stopped or otherwise mangled the clock. If the clock is stopped, then getit() always returns the same value and DELAY() takes forever if it trusts getit(). The quick fix is implement DELAY(n) as (n * timer_freq / 1000000) inb(0x84)'s if ddb is active. machdep.c: Don't permit recursion on clock_lock.
Diffstat (limited to 'sys/isa/atrtc.c')
-rw-r--r--sys/isa/atrtc.c20
1 files changed, 18 insertions, 2 deletions
diff --git a/sys/isa/atrtc.c b/sys/isa/atrtc.c
index 3d7d8c1..29cfefc 100644
--- a/sys/isa/atrtc.c
+++ b/sys/isa/atrtc.c
@@ -322,8 +322,18 @@ DELAY(int n)
* takes about 1.5 usec for each of the i/o's in getit(). The loop
* takes about 6 usec on a 486/33 and 13 usec on a 386/20. The
* multiplications and divisions to scale the count take a while).
+ *
+ * However, if ddb is active then use a fake counter since reading
+ * the i8254 counter involves acquiring a lock. ddb must not go
+ * locking for many reasons, but it calls here for at least atkbd
+ * input.
*/
- prev_tick = getit();
+#ifdef DDB
+ if (db_active)
+ prev_tick = 0;
+ else
+#endif
+ prev_tick = getit();
n -= 0; /* XXX actually guess no initial overhead */
/*
* Calculate (n * (timer_freq / 1e6)) without using floating point
@@ -350,7 +360,13 @@ DELAY(int n)
/ 1000000;
while (ticks_left > 0) {
- tick = getit();
+#ifdef DDB
+ if (db_active) {
+ inb(0x84);
+ tick = prev_tick + 1;
+ } else
+#endif
+ tick = getit();
#ifdef DELAYDEBUG
++getit_calls;
#endif
OpenPOWER on IntegriCloud