IB/ipath: Add locking for interrupt use of ipath_pd contexts vs free

Fixes timing race resulting in panic.  Not a performance sensitive path.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

1 files changed, 30 insertions, 19 deletions

diff --git a/drivers/infiniband/hw/ipath/ipath_driver.c b/drivers/infiniband/hw/ipath/ipath_driver.c
index ad0aab60b051..69c0ce321b4e 100644
--- a/drivers/infiniband/hw/ipath/ipath_driver.c
+++ b/drivers/infiniband/hw/ipath/ipath_driver.c
@@ -661,6 +661,8 @@ bail:
 static void __devexit cleanup_device(struct ipath_devdata *dd)
 {
 	int port;
+	struct ipath_portdata **tmp;
+	unsigned long flags;
 
 	if (*dd->ipath_statusp & IPATH_STATUS_CHIP_PRESENT) {
 		/* can't do anything more with chip; needs re-init */
@@ -742,20 +744,21 @@ static void __devexit cleanup_device(struct ipath_devdata *dd)
 
 	/*
 	 * free any resources still in use (usually just kernel ports)
-	 * at unload; we do for portcnt, not cfgports, because cfgports
-	 * could have changed while we were loaded.
+	 * at unload; we do for portcnt, because that's what we allocate.
+	 * We acquire lock to be really paranoid that ipath_pd isn't being
+	 * accessed from some interrupt-related code (that should not happen,
+	 * but best to be sure).
 	 */
+	spin_lock_irqsave(&dd->ipath_uctxt_lock, flags);
+	tmp = dd->ipath_pd;
+	dd->ipath_pd = NULL;
+	spin_unlock_irqrestore(&dd->ipath_uctxt_lock, flags);
 	for (port = 0; port < dd->ipath_portcnt; port++) {
-		struct ipath_portdata *pd = dd->ipath_pd[port];
-		dd->ipath_pd[port] = NULL;
+		struct ipath_portdata *pd = tmp[port];
+		tmp[port] = NULL; /* debugging paranoia */
 		ipath_free_pddata(dd, pd);
 	}
-	kfree(dd->ipath_pd);
-	/*
-	 * debuggability, in case some cleanup path tries to use it
-	 * after this
-	 */
-	dd->ipath_pd = NULL;
+	kfree(tmp);
 }
 
 static void __devexit ipath_remove_one(struct pci_dev *pdev)
@@ -2586,6 +2589,7 @@ int ipath_reset_device(int unit)
 {
 	int ret, i;
 	struct ipath_devdata *dd = ipath_lookup(unit);
+	unsigned long flags;
 
 	if (!dd) {
 		ret = -ENODEV;
@@ -2611,18 +2615,21 @@ int ipath_reset_device(int unit)
 		goto bail;
 	}
 
+	spin_lock_irqsave(&dd->ipath_uctxt_lock, flags);
 	if (dd->ipath_pd)
 		for (i = 1; i < dd->ipath_cfgports; i++) {
-			if (dd->ipath_pd[i] && dd->ipath_pd[i]->port_cnt) {
-				ipath_dbg("unit %u port %d is in use "
-					  "(PID %u cmd %s), can't reset\n",
-					  unit, i,
-					  pid_nr(dd->ipath_pd[i]->port_pid),
-					  dd->ipath_pd[i]->port_comm);
-				ret = -EBUSY;
-				goto bail;
-			}
+			if (!dd->ipath_pd[i] || !dd->ipath_pd[i]->port_cnt)
+				continue;
+			spin_unlock_irqrestore(&dd->ipath_uctxt_lock, flags);
+			ipath_dbg("unit %u port %d is in use "
+				  "(PID %u cmd %s), can't reset\n",
+				  unit, i,
+				  pid_nr(dd->ipath_pd[i]->port_pid),
+				  dd->ipath_pd[i]->port_comm);
+			ret = -EBUSY;
+			goto bail;
 		}
+	spin_unlock_irqrestore(&dd->ipath_uctxt_lock, flags);
 
 	if (dd->ipath_flags & IPATH_HAS_SEND_DMA)
 		teardown_sdma(dd);
@@ -2656,9 +2663,12 @@ static int ipath_signal_procs(struct ipath_devdata *dd, int sig)
 {
 	int i, sub, any = 0;
 	struct pid *pid;
+	unsigned long flags;
 
 	if (!dd->ipath_pd)
 		return 0;
+
+	spin_lock_irqsave(&dd->ipath_uctxt_lock, flags);
 	for (i = 1; i < dd->ipath_cfgports; i++) {
 		if (!dd->ipath_pd[i] || !dd->ipath_pd[i]->port_cnt)
 			continue;
@@ -2682,6 +2692,7 @@ static int ipath_signal_procs(struct ipath_devdata *dd, int sig)
 			any++;
 		}
 	}
+	spin_unlock_irqrestore(&dd->ipath_uctxt_lock, flags);
 	return any;
 }