Date: Sun, 7 Dec 2008 18:32:09 +0000 (UTC) From: Warner Losh <imp@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r185740 - head/sys/dev/pccbb Message-ID: <200812071832.mB7IW96f032499@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: imp Date: Sun Dec 7 18:32:09 2008 New Revision: 185740 URL: http://svn.freebsd.org/changeset/base/185740 Log: Use atomic_add_int rather than a simple ++ to ensure no cache races if the power interrupt and init code waiting for the interrupt are running on different CPUs. I haven't seen this make any real difference, but I've also had some reports of odd behavior I can't otherwise explain. It is an infrequent operation, and certainly wouldn't hurt. Modified: head/sys/dev/pccbb/pccbb_pci.c Modified: head/sys/dev/pccbb/pccbb_pci.c ============================================================================== --- head/sys/dev/pccbb/pccbb_pci.c Sun Dec 7 13:25:06 2008 (r185739) +++ head/sys/dev/pccbb/pccbb_pci.c Sun Dec 7 18:32:09 2008 (r185740) @@ -730,7 +730,7 @@ cbb_pci_filt(void *arg) if (sockevent & CBB_SOCKET_EVENT_POWER) { cbb_clrb(sc, CBB_SOCKET_MASK, CBB_SOCKET_EVENT_POWER); cbb_set(sc, CBB_SOCKET_EVENT, CBB_SOCKET_EVENT_POWER); - sc->powerintr++; + atomic_add_int(&sc->powerintr, 1); wakeup((void *)&sc->powerintr); } retval = FILTER_HANDLED;
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200812071832.mB7IW96f032499>