d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

This seems to fix my problem that after resume/suspend, sometimes

Browse Source 
pccard claims that the driver is already allocated.
It works around a race when pccardd gets woken up too late after a resume.

This is a 2.2.6 candidate.
Reviewed by:	nate@freebsd.org
This commit is contained in:
Guido van Rooij 1998-02-04 20:19:39 +00:00
parent 1540674007
commit 3d08e8c2cc
1 changed files with 15 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
usr.sbin/pccard/pccardd/cardd.c
View File

@ -26,7 +26,7 @@
#ifndef lint
static const char rcsid[] =
"$Id: cardd.c,v 1.22 1997/11/25 19:15:59 nate Exp $";
"$Id: cardd.c,v 1.23 1997/12/08 06:35:07 nate Exp $";
#endif /* not lint */
#include <fcntl.h>
@ -220,19 +220,31 @@ slot_change(struct slot *sp)
if (state.state == sp->state)
logmsg("State same as before, continuing anyway");
#endif
sp->state = state.state;
switch (sp->state) {
switch (state.state) {
case empty:
case noslot:
card_removed(sp);
break;
case filled:
/*
* If state was already filled, fake a removal first to get
* our state in sync with the kernel. This happens when the
* systems resumes and we only get to process the state
* change from suspend to empty after inserted() has run.
* In that case the kernel state is perfectly normal.
*
* The reason for not doing nothing is that the kernel
* has to be informed again about IRQ and IO window.
*/
if (state.state == sp->state)
card_removed(sp);
card_inserted(sp);
break;
case suspend:
/* ignored */
break;
}
sp->state = state.state;
}
/*