summaryrefslogtreecommitdiffstats
path: root/sys/dev/kbd
diff options
context:
space:
mode:
authormurray <murray@FreeBSD.org>2002-03-27 13:16:11 +0000
committermurray <murray@FreeBSD.org>2002-03-27 13:16:11 +0000
commit20985fad164b94239e7f7e98119b1e39f822db12 (patch)
tree213077e764baed8486c5d3ac268e03f98dc0ca2f /sys/dev/kbd
parent589fdafe98afdf9e7d6fa2c4b4ab6ee3fe05132d (diff)
downloadFreeBSD-src-20985fad164b94239e7f7e98119b1e39f822db12.zip
FreeBSD-src-20985fad164b94239e7f7e98119b1e39f822db12.tar.gz
Fix spelling and grammar bogons in a comment.
PR: kern/30540 Submitted by: Tony Finch <dot@dotat.at> MFC after: 3 days
Diffstat (limited to 'sys/dev/kbd')
-rw-r--r--sys/dev/kbd/atkbdc.c14
1 files changed, 7 insertions, 7 deletions
diff --git a/sys/dev/kbd/atkbdc.c b/sys/dev/kbd/atkbdc.c
index c0616b3..f583f8f 100644
--- a/sys/dev/kbd/atkbdc.c
+++ b/sys/dev/kbd/atkbdc.c
@@ -219,10 +219,10 @@ atkbdc_open(int unit)
* I/O access to the controller/keyboard/mouse. The convention requires
* close cooperation of the calling device driver.
*
- * The device driver which utilizes the `kbdio' module are assumed to
+ * The device drivers which utilize the `kbdio' module are assumed to
* have the following set of routines.
* a. An interrupt handler (the bottom half of the driver).
- * b. Timeout routines which may briefly polls the keyboard controller.
+ * b. Timeout routines which may briefly poll the keyboard controller.
* c. Routines outside interrupt context (the top half of the driver).
* They should follow the rules below:
* 1. The interrupt handler may assume that it always has full access
@@ -241,12 +241,12 @@ atkbdc_open(int unit)
* Therefore, `spltty()' must be strategically placed in the device
* driver code. Also note that the timeout routine may interrupt
* `kbdc_lock()' called by the top half of the driver, but this
- * interruption is OK so long as the timeout routine observes the
- * the rule 4 below.
+ * interruption is OK so long as the timeout routine observes
+ * rule 4 below.
* 4. The interrupt and timeout routines should not extend I/O operation
- * across more than one interrupt or timeout; they must complete
- * necessary I/O operation within one invokation of the routine.
- * This measns that if the timeout routine acquires the lock flag,
+ * across more than one interrupt or timeout; they must complete any
+ * necessary I/O operation within one invocation of the routine.
+ * This means that if the timeout routine acquires the lock flag,
* it must reset the flag to FALSE before it returns.
*/
OpenPOWER on IntegriCloud