svn commit: r223219 - head/sys/dev/pccbb

[Warner Losh]

Author: imp
Date: Sat Jun 18 03:16:51 2011
New Revision: 223219
URL: http://svn.freebsd.org/changeset/base/223219

Log:
  More expeirmentation suggests that 10ms isn't as reliable as
  previously thought, but 100ms seems to be.  Likely there's a good
  middle ground, but for now be conservative.

Modified:
  head/sys/dev/pccbb/pccbb.c

Modified: head/sys/dev/pccbb/pccbb.c
==============================================================================
--- head/sys/dev/pccbb/pccbb.c	Sat Jun 18 02:25:08 2011	(r223218)
+++ head/sys/dev/pccbb/pccbb.c	Sat Jun 18 03:16:51 2011	(r223219)
@@ -811,14 +811,14 @@ cbb_power(device_t brdev, int volts)
 		mtx_unlock(&sc->mtx);
 
 		/*
-		 * Relax for 10ms.  Some bridges appear to assert this signal
+		 * Relax for 100ms.  Some bridges appear to assert this signal
 		 * right away, but before the card has stabilized.  Other
 		 * cards need need more time to cope up reliabily.
 		 * Experiments with troublesome setups show this to be a
 		 * "cheap" way to enhance reliabilty.  We need not do this for
 		 * "off" since we don't touch the card after we turn it off.
 		 */
-		pause("cbbPwr", min(hz / 100, 1));
+		pause("cbbPwr", min(hz / 10, 1));
 
 		/*
 		 * The TOPIC95B requires a little bit extra time to get its