All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stefan Richter <stefanr@s5r6.in-berlin.de>
To: linux1394-devel@lists.sourceforge.net
Cc: linux-kernel@vger.kernel.org
Subject: [PATCH 1/2 rebased] firewire: core: use non-reentrant workqueue with rescuer
Date: Sun, 1 May 2011 20:48:53 +0200	[thread overview]
Message-ID: <20110501204853.2060ab25@stein> (raw)

Date: Wed Oct 13 13:39:46 CEST 2010
From: Stefan Richter <stefanr@s5r6.in-berlin.de>

firewire-core manages the following types of work items:

fw_card.br_work:
  - resets the bus on a card and possibly sends a PHY packet before that
  - does not sleep for long or not at all
  - is scheduled via fw_schedule_bus_reset() by
      - firewire-ohci's pci_probe method
      - firewire-ohci's set_config_rom method, called by kernelspace
        protocol drivers and userspace drivers which add/remove
	Configuration ROM descriptors
      - userspace drivers which use the bus reset ioctl
      - itself if the last reset happened less than 2 seconds ago

fw_card.bm_work:
  - performs bus management duties
  - usually does not (but may in corner cases) sleep for long
  - is scheduled via fw_schedule_bm_work() by
      - firewire-ohci's self-ID-complete IRQ handler tasklet
      - firewire-core's fw_device.work instances whenever the root node
        device was (successfully or unsuccessfully) discovered,
	refreshed, or rediscovered
      - itself in case of resource allocation failures or in order to
        obey the 125ms bus manager arbitration interval

fw_device.work:
  - performs node probe, update, shutdown, revival, removal; including
    kernel driver probe, update, shutdown and bus reset notification to
    userspace drivers
  - usually sleeps moderately long, in corner cases very long
  - is scheduled by
      - firewire-ohci's self-ID-complete IRQ handler tasklet via the
        core's fw_node_event
      - firewire-ohci's pci_remove method via core's fw_destroy_nodes/
        fw_node_event
      - itself during retries, e.g. while a node is powering up

iso_resource.work:
  - accesses registers at the Isochronous Resource Manager node
  - usually does not (but may in corner cases) sleep for long
  - is scheduled via schedule_iso_resource() by
      - the owning userspace driver at addition and removal of the
        resource
      - firewire-core's fw_device.work instances after bus reset
      - itself in case of resource allocation if necessary to obey the
        1000ms reallocation period after bus reset

fw_card.br_work instances should not, and instances of the others must
not, be executed in parallel by multiple CPUs -- but were not protected
against that.  Hence allocate a non-reentrant workqueue for them.

fw_device.work may be used in the memory reclaim path in case of SBP-2
device updates.  Hence we need a workqueue with rescuer and cannot use
system_nrt_wq.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Reviewed-by: Tejun Heo <tj@kernel.org>
---
 drivers/firewire/core-card.c        |    6 +++---
 drivers/firewire/core-cdev.c        |    2 +-
 drivers/firewire/core-device.c      |   30 ++++++++++++++++++----------
 drivers/firewire/core-transaction.c |   12 ++++++++++-
 drivers/firewire/core.h             |    2 ++
 5 files changed, 36 insertions(+), 16 deletions(-)

Index: b/drivers/firewire/core-card.c
===================================================================
--- a/drivers/firewire/core-card.c
+++ b/drivers/firewire/core-card.c
@@ -228,8 +228,8 @@ void fw_schedule_bus_reset(struct fw_car
 
 	/* Use an arbitrary short delay to combine multiple reset requests. */
 	fw_card_get(card);
-	if (!schedule_delayed_work(&card->br_work,
-				   delayed ? DIV_ROUND_UP(HZ, 100) : 0))
+	if (!queue_delayed_work(fw_wq, &card->br_work,
+				delayed ? DIV_ROUND_UP(HZ, 100) : 0))
 		fw_card_put(card);
 }
 EXPORT_SYMBOL(fw_schedule_bus_reset);
@@ -241,7 +241,7 @@ static void br_work(struct work_struct *
 	/* Delay for 2s after last reset per IEEE 1394 clause 8.2.1. */
 	if (card->reset_jiffies != 0 &&
 	    time_before64(get_jiffies_64(), card->reset_jiffies + 2 * HZ)) {
-		if (!schedule_delayed_work(&card->br_work, 2 * HZ))
+		if (!queue_delayed_work(fw_wq, &card->br_work, 2 * HZ))
 			fw_card_put(card);
 		return;
 	}
Index: b/drivers/firewire/core-cdev.c
===================================================================
--- a/drivers/firewire/core-cdev.c
+++ b/drivers/firewire/core-cdev.c
@@ -149,7 +149,7 @@ static void release_iso_resource(struct 
 static void schedule_iso_resource(struct iso_resource *r, unsigned long delay)
 {
 	client_get(r->client);
-	if (!schedule_delayed_work(&r->work, delay))
+	if (!queue_delayed_work(fw_wq, &r->work, delay))
 		client_put(r->client);
 }
 
Index: b/drivers/firewire/core-device.c
===================================================================
--- a/drivers/firewire/core-device.c
+++ b/drivers/firewire/core-device.c
@@ -725,6 +725,14 @@ struct fw_device *fw_device_get_by_devt(
 	return device;
 }
 
+struct workqueue_struct *fw_wq;
+
+static void fw_schedule_device_work(struct fw_device *device,
+				    unsigned long delay)
+{
+	queue_delayed_work(fw_wq, &device->work, delay);
+}
+
 /*
  * These defines control the retry behavior for reading the config
  * rom.  It shouldn't be necessary to tweak these; if the device
@@ -750,7 +758,7 @@ static void fw_device_shutdown(struct wo
 	if (time_before64(get_jiffies_64(),
 			  device->card->reset_jiffies + SHUTDOWN_DELAY)
 	    && !list_empty(&device->card->link)) {
-		schedule_delayed_work(&device->work, SHUTDOWN_DELAY);
+		fw_schedule_device_work(device, SHUTDOWN_DELAY);
 		return;
 	}
 
@@ -862,7 +870,7 @@ static int lookup_existing_device(struct
 		fw_notify("rediscovered device %s\n", dev_name(dev));
 
 		PREPARE_DELAYED_WORK(&old->work, fw_device_update);
-		schedule_delayed_work(&old->work, 0);
+		fw_schedule_device_work(old, 0);
 
 		if (current_node == card->root_node)
 			fw_schedule_bm_work(card, 0);
@@ -953,7 +961,7 @@ static void fw_device_init(struct work_s
 		if (device->config_rom_retries < MAX_RETRIES &&
 		    atomic_read(&device->state) == FW_DEVICE_INITIALIZING) {
 			device->config_rom_retries++;
-			schedule_delayed_work(&device->work, RETRY_DELAY);
+			fw_schedule_device_work(device, RETRY_DELAY);
 		} else {
 			if (device->node->link_on)
 				fw_notify("giving up on config rom for node id %x\n",
@@ -1019,7 +1027,7 @@ static void fw_device_init(struct work_s
 			   FW_DEVICE_INITIALIZING,
 			   FW_DEVICE_RUNNING) == FW_DEVICE_GONE) {
 		PREPARE_DELAYED_WORK(&device->work, fw_device_shutdown);
-		schedule_delayed_work(&device->work, SHUTDOWN_DELAY);
+		fw_schedule_device_work(device, SHUTDOWN_DELAY);
 	} else {
 		if (device->config_rom_retries)
 			fw_notify("created device %s: GUID %08x%08x, S%d00, "
@@ -1098,7 +1106,7 @@ static void fw_device_refresh(struct wor
 		if (device->config_rom_retries < MAX_RETRIES / 2 &&
 		    atomic_read(&device->state) == FW_DEVICE_INITIALIZING) {
 			device->config_rom_retries++;
-			schedule_delayed_work(&device->work, RETRY_DELAY / 2);
+			fw_schedule_device_work(device, RETRY_DELAY / 2);
 
 			return;
 		}
@@ -1131,7 +1139,7 @@ static void fw_device_refresh(struct wor
 		if (device->config_rom_retries < MAX_RETRIES &&
 		    atomic_read(&device->state) == FW_DEVICE_INITIALIZING) {
 			device->config_rom_retries++;
-			schedule_delayed_work(&device->work, RETRY_DELAY);
+			fw_schedule_device_work(device, RETRY_DELAY);
 
 			return;
 		}
@@ -1158,7 +1166,7 @@ static void fw_device_refresh(struct wor
  gone:
 	atomic_set(&device->state, FW_DEVICE_GONE);
 	PREPARE_DELAYED_WORK(&device->work, fw_device_shutdown);
-	schedule_delayed_work(&device->work, SHUTDOWN_DELAY);
+	fw_schedule_device_work(device, SHUTDOWN_DELAY);
  out:
 	if (node_id == card->root_node->node_id)
 		fw_schedule_bm_work(card, 0);
@@ -1214,7 +1222,7 @@ void fw_node_event(struct fw_card *card,
 		 * first config rom scan half a second after bus reset.
 		 */
 		INIT_DELAYED_WORK(&device->work, fw_device_init);
-		schedule_delayed_work(&device->work, INITIAL_DELAY);
+		fw_schedule_device_work(device, INITIAL_DELAY);
 		break;
 
 	case FW_NODE_INITIATED_RESET:
@@ -1230,7 +1238,7 @@ void fw_node_event(struct fw_card *card,
 			    FW_DEVICE_RUNNING,
 			    FW_DEVICE_INITIALIZING) == FW_DEVICE_RUNNING) {
 			PREPARE_DELAYED_WORK(&device->work, fw_device_refresh);
-			schedule_delayed_work(&device->work,
+			fw_schedule_device_work(device,
 				device->is_local ? 0 : INITIAL_DELAY);
 		}
 		break;
@@ -1245,7 +1253,7 @@ void fw_node_event(struct fw_card *card,
 		device->generation = card->generation;
 		if (atomic_read(&device->state) == FW_DEVICE_RUNNING) {
 			PREPARE_DELAYED_WORK(&device->work, fw_device_update);
-			schedule_delayed_work(&device->work, 0);
+			fw_schedule_device_work(device, 0);
 		}
 		break;
 
@@ -1270,7 +1278,7 @@ void fw_node_event(struct fw_card *card,
 		if (atomic_xchg(&device->state,
 				FW_DEVICE_GONE) == FW_DEVICE_RUNNING) {
 			PREPARE_DELAYED_WORK(&device->work, fw_device_shutdown);
-			schedule_delayed_work(&device->work,
+			fw_schedule_device_work(device,
 				list_empty(&card->link) ? 0 : SHUTDOWN_DELAY);
 		}
 		break;
Index: b/drivers/firewire/core-transaction.c
===================================================================
--- a/drivers/firewire/core-transaction.c
+++ b/drivers/firewire/core-transaction.c
@@ -36,6 +36,7 @@
 #include <linux/string.h>
 #include <linux/timer.h>
 #include <linux/types.h>
+#include <linux/workqueue.h>
 
 #include <asm/byteorder.h>
 
@@ -1213,13 +1214,21 @@ static int __init fw_core_init(void)
 {
 	int ret;
 
+	fw_wq = alloc_workqueue(KBUILD_MODNAME,
+				WQ_NON_REENTRANT | WQ_MEM_RECLAIM, 0);
+	if (!fw_wq)
+		return -ENOMEM;
+
 	ret = bus_register(&fw_bus_type);
-	if (ret < 0)
+	if (ret < 0) {
+		destroy_workqueue(fw_wq);
 		return ret;
+	}
 
 	fw_cdev_major = register_chrdev(0, "firewire", &fw_device_ops);
 	if (fw_cdev_major < 0) {
 		bus_unregister(&fw_bus_type);
+		destroy_workqueue(fw_wq);
 		return fw_cdev_major;
 	}
 
@@ -1235,6 +1244,7 @@ static void __exit fw_core_cleanup(void)
 {
 	unregister_chrdev(fw_cdev_major, "firewire");
 	bus_unregister(&fw_bus_type);
+	destroy_workqueue(fw_wq);
 	idr_destroy(&fw_device_idr);
 }
 
Index: b/drivers/firewire/core.h
===================================================================
--- a/drivers/firewire/core.h
+++ b/drivers/firewire/core.h
@@ -138,6 +138,8 @@ void fw_cdev_handle_phy_packet(struct fw
 extern struct rw_semaphore fw_device_rwsem;
 extern struct idr fw_device_idr;
 extern int fw_cdev_major;
+struct workqueue_struct;
+extern struct workqueue_struct *fw_wq;
 
 struct fw_device *fw_device_get_by_devt(dev_t devt);
 int fw_device_set_broadcast_channel(struct device *dev, void *gen);


-- 
Stefan Richter
-=====-==-== -=-= ----=
http://arcgraph.de/sr/

             reply	other threads:[~2011-05-01 18:49 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-05-01 18:48 Stefan Richter [this message]
2011-05-01 18:50 ` [PATCH 2/2] firewire: sbp2: parallelize login, reconnect, logout Stefan Richter

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=20110501204853.2060ab25@stein \
    --to=stefanr@s5r6.in-berlin.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux1394-devel@lists.sourceforge.net \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.