[PATCH v3 1/8] workqueue: Add resource managed version of delayed work init

[Matti Vaittinen <matti.vaittinen@fi.rohmeurope.com>]

A few drivers which need a delayed work-queue must cancel work at driver
detach. Some of those implement remove() solely for this purpose. Help
drivers to avoid unnecessary remove and error-branch implementation by
adding managed verision of delayed work initialization. This will also
help drivers to avoid mixing manual and devm based unwinding when other
resources are handled by devm.

Signed-off-by: Matti Vaittinen <matti.vaittinen@fi.rohmeurope.com>
---
Changelog from RFCv2:
 - RFC dropped. No functional changes.

 include/linux/devm-helpers.h | 53 ++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)
 create mode 100644 include/linux/devm-helpers.h

diff --git a/include/linux/devm-helpers.h b/include/linux/devm-helpers.h
new file mode 100644
index 000000000000..f64e0c9f3763
--- /dev/null
+++ b/include/linux/devm-helpers.h
@@ -0,0 +1,53 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+#ifndef __LINUX_DEVM_HELPERS_H
+#define __LINUX_DEVM_HELPERS_H
+
+/*
+ * Functions which do automatically cancel operations or release resources upon
+ * driver detach.
+ *
+ * These should be helpful to avoid mixing the manual and devm-based resource
+ * management which can be source of annoying, rarely occurring,
+ * hard-to-reproduce bugs.
+ *
+ * Please take into account that devm based cancellation may be performed some
+ * time after the remove() is ran.
+ *
+ * Thus mixing devm and manual resource management can easily cause problems
+ * when unwinding operations with dependencies. IRQ scheduling a work in a queue
+ * is typical example where IRQs are often devm-managed and WQs are manually
+ * cleaned at remove(). If IRQs are not manually freed at remove() (and this is
+ * often the case when we use devm for IRQs) we have a period of time after
+ * remove() - and before devm managed IRQs are freed - where new IRQ may fire
+ * and schedule a work item which won't be cancelled because remove() was
+ * already ran.
+ */
+
+#include <linux/device.h>
+#include <linux/workqueue.h>
+
+static inline void devm_delayed_work_drop(void *res)
+{
+	cancel_delayed_work_sync(res);
+}
+
+/**
+ * devm_delayed_work_autocancel - Resource-managed work allocation
+ * @dev: Device which lifetime work is bound to
+ * @pdata: work to be cancelled when driver is detached
+ *
+ * Initialize work which is automatically cancelled when driver is detached.
+ * A few drivers need delayed work which must be cancelled before driver
+ * is detached to avoid accessing removed resources.
+ * devm_delayed_work_autocancel() can be used to omit the explicit
+ * cancelleation when driver is detached.
+ */
+static inline int devm_delayed_work_autocancel(struct device *dev,
+					       struct delayed_work *w,
+					       work_func_t worker)
+{
+	INIT_DELAYED_WORK(w, worker);
+	return devm_add_action(dev, devm_delayed_work_drop, w);
+}
+
+#endif
-- 
2.25.4


-- 
Matti Vaittinen, Linux device drivers
ROHM Semiconductors, Finland SWDC
Kiviharjunlenkki 1E
90220 OULU
FINLAND

~~~ "I don't think so," said Rene Descartes. Just then he vanished ~~~
Simon says - in Latin please.
~~~ "non cogito me" dixit Rene Descarte, deinde evanescavit ~~~
Thanks to Simon Glass for the translation =]