EDAC: Robustify workqueues destruction
authorBorislav Petkov <bp@suse.de>
Fri, 27 Nov 2015 09:38:38 +0000 (10:38 +0100)
committerJiri Slaby <jslaby@suse.cz>
Thu, 3 Mar 2016 11:46:03 +0000 (12:46 +0100)
commit fcd5c4dd8201595d4c598c9cca5e54760277d687 upstream.

EDAC workqueue destruction is really fragile. We cancel delayed work
but if it is still running and requeues itself, we still go ahead and
destroy the workqueue and the queued work explodes when workqueue core
attempts to run it.

Make the destruction more robust by switching op_state to offline so
that requeuing stops. Cancel any pending work *synchronously* too.

  EDAC i7core: Driver loaded.
  general protection fault: 0000 [#1] SMP
  CPU 12
  Modules linked in:
  Supported: Yes
  Pid: 0, comm: kworker/0:1 Tainted: G          IE   3.0.101-0-default #1 HP ProLiant DL380 G7
  RIP: 0010:[<ffffffff8107dcd7>]  [<ffffffff8107dcd7>] __queue_work+0x17/0x3f0
  < ... regs ...>
  Process kworker/0:1 (pid: 0, threadinfo ffff88019def6000, task ffff88019def4600)
  Stack:
   ...
  Call Trace:
   call_timer_fn
   run_timer_softirq
   __do_softirq
   call_softirq
   do_softirq
   irq_exit
   smp_apic_timer_interrupt
   apic_timer_interrupt
   intel_idle
   cpuidle_idle_call
   cpu_idle
  Code: ...
  RIP  __queue_work
   RSP <...>

Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
drivers/edac/edac_device.c
drivers/edac/edac_mc.c
drivers/edac/edac_pci.c

index 211021dfec734a5e5466e2155eaf108ee4ec997c..46ef63d055847dcf623765400dddd1cd07b8482b 100644 (file)
@@ -435,13 +435,10 @@ void edac_device_workq_setup(struct edac_device_ctl_info *edac_dev,
  */
 void edac_device_workq_teardown(struct edac_device_ctl_info *edac_dev)
 {
-       int status;
+       edac_dev->op_state = OP_OFFLINE;
 
-       status = cancel_delayed_work(&edac_dev->work);
-       if (status == 0) {
-               /* workq instance might be running, wait for it */
-               flush_workqueue(edac_workqueue);
-       }
+       cancel_delayed_work_sync(&edac_dev->work);
+       flush_workqueue(edac_workqueue);
 }
 
 /*
index a9d98cdd11f4b1cfe07e4be94e8e29dafe8fa778..f1f298b3ff16e3621675b79579ccd66b5a5c3253 100644 (file)
@@ -584,18 +584,10 @@ static void edac_mc_workq_setup(struct mem_ctl_info *mci, unsigned msec,
  */
 static void edac_mc_workq_teardown(struct mem_ctl_info *mci)
 {
-       int status;
-
-       if (mci->op_state != OP_RUNNING_POLL)
-               return;
-
-       status = cancel_delayed_work(&mci->work);
-       if (status == 0) {
-               edac_dbg(0, "not canceled, flush the queue\n");
+       mci->op_state = OP_OFFLINE;
 
-               /* workq instance might be running, wait for it */
-               flush_workqueue(edac_workqueue);
-       }
+       cancel_delayed_work_sync(&mci->work);
+       flush_workqueue(edac_workqueue);
 }
 
 /*
index dd370f92ace3402321bbc6f94cfd6814cbb3fb4f..e1e6d3653e03240f26d5985c5e1719fcd4741dd6 100644 (file)
@@ -274,13 +274,12 @@ static void edac_pci_workq_setup(struct edac_pci_ctl_info *pci,
  */
 static void edac_pci_workq_teardown(struct edac_pci_ctl_info *pci)
 {
-       int status;
-
        edac_dbg(0, "\n");
 
-       status = cancel_delayed_work(&pci->work);
-       if (status == 0)
-               flush_workqueue(edac_workqueue);
+       pci->op_state = OP_OFFLINE;
+
+       cancel_delayed_work_sync(&pci->work);
+       flush_workqueue(edac_workqueue);
 }
 
 /*