remoteproc: add rproc_report_crash function to notify rproc crashes

Allow low-level remoteproc drivers to report rproc crashes by exporting
a new rproc_report_crash() function (invoking this from non-rproc drivers
is probably wrong, and should be carefully scrutinized if ever needed).

rproc_report_crash() can be called from any context; it offloads the
tasks of handling the crash to a separate thread.

Handling the crash from a separate thread is helpful because:
- Ability to call invoke rproc_report_crash() from atomic context, due to
  the fact that many crashes trigger an interrupt, so this function can be
  called directly from ISR context.
- Avoiding deadlocks which could happen if rproc_report_crash() is called
  from a function which indirectly holds the rproc lock.

Handling the crash might involve:
- Remoteproc register dump
- Remoteproc stack dump
- Remoteproc core dump
- Saving Remoteproc traces so they can be read after the crash
- Reseting the remoteproc in order to make it functional again (hard recovery)

Right now, we only print the crash type which was detected, and only the
mmufault type is supported. Remoteproc low-level drivers can add more types
when needed.

Signed-off-by: Fernando Guzman Lugo <fernando.lugo@ti.com>
[ohad: some commentary, white space and commit log changes]
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>

1 files changed, 18 insertions, 0 deletions

diff --git a/include/linux/remoteproc.h b/include/linux/remoteproc.h
index 131b53957b9f..a46ed2723803 100644
--- a/include/linux/remoteproc.h
+++ b/include/linux/remoteproc.h
@@ -361,6 +361,19 @@ enum rproc_state {
 };
 
 /**
+ * enum rproc_crash_type - remote processor crash types
+ * @RPROC_MMUFAULT:     iommu fault
+ *
+ * Each element of the enum is used as an array index. So that, the value of
+ * the elements should be always something sane.
+ *
+ * Feel free to add more types when needed.
+ */
+enum rproc_crash_type {
+        RPROC_MMUFAULT,
+};
+
+/**
  * struct rproc - represents a physical remote processor device
  * @node: klist node of this rproc object
  * @domain: iommu domain
@@ -383,6 +396,8 @@ enum rproc_state {
  * @rvdevs: list of remote virtio devices
  * @notifyids: idr for dynamically assigning rproc-wide unique notify ids
  * @index: index of this rproc device
+ * @crash_handler: workqueue for handling a crash
+ * @crash_cnt: crash counter
  */
 struct rproc {
         struct klist_node node;
@@ -406,6 +421,8 @@ struct rproc {
         struct list_head rvdevs;
         struct idr notifyids;
         int index;
+        struct work_struct crash_handler;
+        unsigned crash_cnt;
 };
 
 /* we currently support only two vrings per rvdev */
@@ -460,6 +477,7 @@ int rproc_del(struct rproc *rproc);
 
 int rproc_boot(struct rproc *rproc);
 void rproc_shutdown(struct rproc *rproc);
+void rproc_report_crash(struct rproc *rproc, enum rproc_crash_type type);
 
 static inline struct rproc_vdev *vdev_to_rvdev(struct virtio_device *vdev)
 {