Add /proc/gpu#/resubmit_runlist API

Resubmits the runlist in an identical configuration. Causes the
runlist scheduler to:
1. Reload and cache timeslice and scale values from TSGs.
2. Restart scheduling from the head of the runlist [may cause a
   preempt to be scheduled for the currently-running task (?)].
3. Address (?) an errata on Turing where re-enabled channels are
   not always detected.

Above behavior tested on GV100 and partially tested on TU102.

1 files changed, 12 insertions, 0 deletions

diff --git a/nvdebug_entry.c b/nvdebug_entry.c
index 24fcd32..eee7351 100644
--- a/nvdebug_entry.c
+++ b/nvdebug_entry.c
@@ -25,6 +25,7 @@ MODULE_DESCRIPTION("A scheduling debugging module for NVIDIA GPUs");
 // runlist_procfs.c
 extern struct file_operations runlist_file_ops;
 extern struct file_operations preempt_tsg_file_ops;
+extern struct file_operations resubmit_runlist_file_ops;
 extern struct file_operations disable_channel_file_ops;
 extern struct file_operations enable_channel_file_ops;
 extern struct file_operations switch_to_tsg_file_ops;
@@ -256,6 +257,17 @@ int __init nvdebug_init(void) {
                                 "preempt_tsg", 0222, dir, compat_ops(&preempt_tsg_file_ops),
                                 (void*)device_id))
                         goto out_nomem;
+                /* On the TU104, the context scheduler (contained in the Host, aka
+                 * PFIFO, unit) has been observed to sometimes to fail to schedule TSGs
+                 * containing re-enabled channels. Resubmitting the runlist
+                 * configuration appears to remediate this condition, and so this API
+                 * is exposed to help reset GPU scheduling as necessary.
+                 */
+                // Create file `/proc/gpu#/resubmit_runlist`, world writable
+                if (!proc_create_data(
+                                "resubmit_runlist", 0222, dir, compat_ops(&resubmit_runlist_file_ops),
+                                (void*)device_id))
+                        goto out_nomem;
                 // Create file `/proc/gpu#/disable_channel`, world writable
                 if (!proc_create_data(
                                 "disable_channel", 0222, dir, compat_ops(&disable_channel_file_ops),