linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jiang Liu <liuj97@gmail.com>
To: Yinghai Lu <yinghai@kernel.org>,
	Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>,
	Bjorn Helgaas <bhelgaas@google.com>,
	Don Dutile <ddutile@redhat.com>,
	Greg KH <gregkh@linuxfoundation.org>
Cc: Jiang Liu <jiang.liu@huawei.com>,
	Keping Chen <chenkeping@huawei.com>,
	linux-kernel@vger.kernel.org, linux-pci@vger.kernel.org,
	Jiang Liu <liuj97@gmail.com>
Subject: [PATCH v2 14/19] PCI: fix race windows when shutting down cpcihp controller
Date: Fri, 27 Apr 2012 23:16:55 +0800	[thread overview]
Message-ID: <1335539820-11232-15-git-send-email-jiang.liu@huawei.com> (raw)
In-Reply-To: <1335539820-11232-1-git-send-email-jiang.liu@huawei.com>

From: Jiang Liu <jiang.liu@huawei.com>

When cpci_hp_stop() is called to disabled cpcihp controller, it will disable
interrupt for that controller. But there's small window for event_thread()
to reenable the interrupt again. So stop the worker thread before disabling
the interrupt.

If check_slots() returns error, ther working thread (cpci_thread) will exit.
Later when cpci_stop_thread() or cpci_hp_intr() tries to access cpci_thread,
it may have already been destroyed. So hold a reference count to cpci_thread
to avoid invalid memory access.

Signed-off-by: Jiang Liu <liuj97@gmail.com>
---
 drivers/pci/hotplug/cpci_hotplug_core.c |   27 +++++++++++++++++----------
 1 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/drivers/pci/hotplug/cpci_hotplug_core.c b/drivers/pci/hotplug/cpci_hotplug_core.c
index 7898023..68e43c7 100644
--- a/drivers/pci/hotplug/cpci_hotplug_core.c
+++ b/drivers/pci/hotplug/cpci_hotplug_core.c
@@ -60,7 +60,6 @@ static atomic_t extracting;
 int cpci_debug;
 static struct cpci_hp_controller *controller;
 static struct task_struct *cpci_thread;
-static int thread_finished;
 
 static int enable_slot(struct hotplug_slot *slot);
 static int disable_slot(struct hotplug_slot *slot);
@@ -341,7 +340,8 @@ cpci_hp_intr(int irq, void *data)
 	controller->ops->disable_irq();
 
 	/* Trigger processing by the event thread */
-	wake_up_process(cpci_thread);
+	if (cpci_thread)
+		wake_up_process(cpci_thread);
 	return IRQ_HANDLED;
 }
 
@@ -508,7 +508,6 @@ event_thread(void *data)
 				msleep(500);
 			} else if (rc < 0) {
 				dbg("%s - error checking slots", __func__);
-				thread_finished = 1;
 				goto out;
 			}
 		} while (atomic_read(&extracting) && !kthread_should_stop());
@@ -540,7 +539,6 @@ poll_thread(void *data)
 					msleep(500);
 				} else if (rc < 0) {
 					dbg("%s - error checking slots", __func__);
-					thread_finished = 1;
 					goto out;
 				}
 			} while (atomic_read(&extracting) && !kthread_should_stop());
@@ -562,15 +560,24 @@ cpci_start_thread(void)
 		err("Can't start up our thread");
 		return PTR_ERR(cpci_thread);
 	}
-	thread_finished = 0;
+	get_task_struct(cpci_thread);
+
 	return 0;
 }
 
 static void
 cpci_stop_thread(void)
 {
-	kthread_stop(cpci_thread);
-	thread_finished = 1;
+	struct task_struct *tp;
+
+	if (cpci_thread) {
+		local_irq_disable();
+		tp = cpci_thread;
+		cpci_thread = NULL;
+		local_irq_enable();
+		kthread_stop(tp);
+		put_task_struct(tp);
+	}
 }
 
 int
@@ -627,8 +634,7 @@ cpci_hp_unregister_controller(struct cpci_hp_controller *old_controller)
 	int status = 0;
 
 	if (controller) {
-		if (!thread_finished)
-			cpci_stop_thread();
+		cpci_stop_thread();
 		if (controller->irq)
 			free_irq(controller->irq, controller->dev_id);
 		controller = NULL;
@@ -680,12 +686,13 @@ cpci_hp_stop(void)
 {
 	if (!controller)
 		return -ENODEV;
+	cpci_stop_thread();
 	if (controller->irq) {
 		/* Stop enum interrupt processing */
 		dbg("%s - disabling irq", __func__);
 		controller->ops->disable_irq();
+		synchronize_irq(controller->irq);
 	}
-	cpci_stop_thread();
 	return 0;
 }
 
-- 
1.7.5.4


  parent reply	other threads:[~2012-04-27 15:16 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-04-27 15:16 [PATCH v2 00/19] Introduce a global lock to serialize all PCI hotplug Jiang Liu
2012-04-27 15:16 ` [PATCH v2 01/19] PCI: introduce pci_bus_get()/pci_bus_put() to hide PCI implementation details Jiang Liu
2012-04-27 15:16 ` [PATCH v2 02/19] PCI: introduce recursive rwsem to serialize PCI hotplug operations Jiang Liu
2012-05-02  5:00   ` Greg KH
2012-05-02  7:25     ` Jiang Liu
2012-05-02 21:46       ` Greg KH
2012-04-27 15:16 ` [PATCH v2 03/19] PCI: replace pci_remove_rescan_mutex with the PCI hotplug lock Jiang Liu
2012-04-27 15:16 ` [PATCH v2 04/19] PCI: serialize hotplug operations triggered by PCI hotplug sysfs interfaces Jiang Liu
2012-05-02  5:06   ` Greg KH
2012-05-02  7:20     ` Jiang Liu
2012-05-02 21:48       ` Greg KH
2012-04-27 15:16 ` [PATCH v2 05/19] PCI: correctly flush workqueue when destroy pcie hotplug controller Jiang Liu
2012-05-02  5:08   ` Greg KH
2012-04-27 15:16 ` [PATCH v2 06/19] PCI: prepare for serializing hotplug operations triggered by pciehp driver Jiang Liu
2012-05-02  5:10   ` Greg KH
2012-04-27 15:16 ` [PATCH v2 07/19] PCI: serialize hotplug operaitons triggered by the " Jiang Liu
2012-04-27 15:16 ` [PATCH v2 08/19] PCI: fix two race windows when probing/removing SHPC controller Jiang Liu
2012-04-27 15:16 ` [PATCH v2 09/19] PCI: correctly flush workqueues and timer when destroy " Jiang Liu
2012-04-27 15:16 ` [PATCH v2 10/19] PCI: serialize hotplug operaitons triggered by the shpchp driver Jiang Liu
     [not found]   ` <4FA4F3E5.5040600@fold.natur.cuni.cz>
     [not found]     ` <4FA5EF97.8010306@gmail.com>
     [not found]       ` <20120506152036.GA31551@kroah.com>
     [not found]         ` <4FA7D882.3070903@gmail.com>
     [not found]           ` <4FA7D839.9090109@fold.natur.cuni.cz>
2012-05-07 14:13             ` Martin Mokrejs
2012-05-07 18:40               ` Greg Kroah-Hartman
2012-05-07 19:23                 ` Martin Mokrejs
2012-05-08  0:01         ` Jiang Liu
2012-05-08  0:18           ` Greg Kroah-Hartman
2012-05-08 14:13             ` Jiang Liu
2012-05-08 14:48               ` Greg Kroah-Hartman
2012-05-08 16:34                 ` Jiang Liu
2012-04-27 15:16 ` [PATCH v2 11/19] PCI: release IO resource in error handling path in cpcihp_generic_init() Jiang Liu
2012-04-27 15:16 ` [PATCH v2 12/19] PCI: clean up all resources in error handling path in zt5550_hc_init_one() Jiang Liu
2012-04-27 15:16 ` [PATCH v2 13/19] PCI: trivial code clean up in cpci_hotplug_core.c Jiang Liu
2012-04-27 15:16 ` Jiang Liu [this message]
2012-04-27 15:16 ` [PATCH v2 15/19] PCI: hold a reference count to the PCI bus used by cpcihp drivers Jiang Liu
2012-04-27 15:16 ` [PATCH v2 16/19] PCI: serialize PCI hotplug operations triggered " Jiang Liu
2012-04-27 15:16 ` [PATCH v2 17/19] PCI: serialize PCI hotplug operations triggered by fakephp drivers Jiang Liu
2012-04-27 15:16 ` [PATCH v2 18/19] PCI, sysfs: Use device_type and attr_groups with pci dev Jiang Liu
2012-04-27 15:17 ` [PATCH v2 19/19] PCI: hide sys interface 'remove' and 'rescan' for SR-IOV virtual devices Jiang Liu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1335539820-11232-15-git-send-email-jiang.liu@huawei.com \
    --to=liuj97@gmail.com \
    --cc=bhelgaas@google.com \
    --cc=chenkeping@huawei.com \
    --cc=ddutile@redhat.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=jiang.liu@huawei.com \
    --cc=kaneshige.kenji@jp.fujitsu.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=yinghai@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).