Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

coretemp_refresh: run xc_unicast() regardless if sc->sc_ci is curcpu()

Browse Source 
or not, this fixes a deadlock seen by Greg Oster in a Dual Quad Core
machine with 8 coretemp instances.
This commit is contained in:
xtraeme 2008-01-28 22:49:49 +00:00
parent ab245e294e
commit 1cc7b13a55
1 changed files with 4 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
sys/arch/x86/x86/coretemp.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: coretemp.c,v 1.8 2008/01/28 20:19:06 xtraeme Exp $ */
/* $NetBSD: coretemp.c,v 1.9 2008/01/28 22:49:49 xtraeme Exp $ */
/*-
* Copyright (c) 2007 Juan Romero Pardines.
@ -36,7 +36,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: coretemp.c,v 1.8 2008/01/28 20:19:06 xtraeme Exp $");
__KERNEL_RCSID(0, "$NetBSD: coretemp.c,v 1.9 2008/01/28 22:49:49 xtraeme Exp $");
#include <sys/param.h>
#include <sys/kmem.h>
@ -161,15 +161,8 @@ coretemp_refresh(struct sysmon_envsys *sme, envsys_data_t *edata)
struct coretemp_softc *sc = sme->sme_cookie;
uint64_t where;
/*
* No need to cross-call if we are running on the same CPU.
*/
if (curcpu() != sc->sc_ci) {
where = xc_unicast(0, coretemp_refresh_xcall,
sc, edata, sc->sc_ci);
xc_wait(where);
} else
coretemp_refresh_xcall(sc, edata);
where = xc_unicast(0, coretemp_refresh_xcall, sc, edata, sc->sc_ci);
xc_wait(where);
}
static void