On Thu, Feb 25, 2021 at 2:32 AM elaine.zhang [off-list ref] wrote:

Hi, Rafael:

在 2021/2/25 上午1:53, Rafael J. Wysocki 写道:

quoted

From: Rafael J. Wysocki <redacted>

Because the PM-runtime status of the device is not updated in
__rpm_callback(), attempts to suspend the suppliers of the given
device triggered by rpm_put_suppliers() called by it may fail.

Fix this by making __rpm_callback() update the device's status to
RPM_SUSPENDED before calling rpm_put_suppliers() if the current
status of the device is RPM_SUSPENDING and the callback just invoked
by it has returned 0 (success).

While at it, modify the code in __rpm_callback() to always check
the device's PM-runtime status under its PM lock.

Link: https://lore.kernel.org/linux-pm/CAPDyKFqm06KDw_p8WXsM4dijDbho4bb6T4k50UqqvR1_COsp8g@mail.gmail.com/ (local)
Fixes: 21d5c57b3726 ("PM / runtime: Use device links")
Reported-by: elaine.zhang <redacted>
Diagnosed-by: Ulf Hansson [off-list ref]
Signed-off-by: Rafael J. Wysocki <redacted>
---

This is different from the previously posted tentative patch, please retest.

---
  drivers/base/power/runtime.c |   61 +++++++++++++++++++++++++------------------
  1 file changed, 36 insertions(+), 25 deletions(-)

Index: linux-pm/drivers/base/power/runtime.c
===================================================================

--- linux-pm.orig/drivers/base/power/runtime.c
+++ linux-pm/drivers/base/power/runtime.c

@@ -325,22 +325,21 @@ static void rpm_put_suppliers(struct dev
  static int __rpm_callback(int (*cb)(struct device *), struct device *dev)
      __releases(&dev->power.lock) __acquires(&dev->power.lock)
  {
-     int retval, idx;
      bool use_links = dev->power.links_count > 0;
+     int retval, idx;
+     bool get, put;

      if (dev->power.irq_safe) {
              spin_unlock(&dev->power.lock);
+     } else if (!use_links) {
+             spin_unlock_irq(&dev->power.lock);
      } else {
+             get = dev->power.runtime_status == RPM_RESUMING;
+
              spin_unlock_irq(&dev->power.lock);

-             /*
-              * Resume suppliers if necessary.
-              *
-              * The device's runtime PM status cannot change until this
-              * routine returns, so it is safe to read the status outside of
-              * the lock.
-              */
-             if (use_links && dev->power.runtime_status == RPM_RESUMING) {
+             /* Resume suppliers if necessary. */
+             if (get) {
                      idx = device_links_read_lock();

                      retval = rpm_get_suppliers(dev);

@@ -355,24 +354,36 @@ static int __rpm_callback(int (*cb)(stru

      if (dev->power.irq_safe) {
              spin_lock(&dev->power.lock);
-     } else {
-             /*
-              * If the device is suspending and the callback has returned
-              * success, drop the usage counters of the suppliers that have
-              * been reference counted on its resume.
-              *
-              * Do that if resume fails too.
-              */
-             if (use_links
-                 && ((dev->power.runtime_status == RPM_SUSPENDING && !retval)
-                 || (dev->power.runtime_status == RPM_RESUMING && retval))) {
-                     idx = device_links_read_lock();
+             return retval;
+     }

- fail:
-                     rpm_put_suppliers(dev);
+     spin_lock_irq(&dev->power.lock);

-                     device_links_read_unlock(idx);
-             }
+     if (!use_links)
+             return retval;
+
+     /*
+      * If the device is suspending and the callback has returned success,
+      * drop the usage counters of the suppliers that have been reference
+      * counted on its resume.
+      *
+      * Do that if the resume fails too.
+      */
+     put = dev->power.runtime_status == RPM_SUSPENDING && !retval;
+     if (put)
+             __update_runtime_status(dev, RPM_SUSPENDED);
+     else
+             put = get && retval;
+
+     if (put) {
+             spin_unlock_irq(&dev->power.lock);
+
+             idx = device_links_read_lock();
+
+fail:
+             rpm_put_suppliers(dev);
+
+             device_links_read_unlock(idx);

              spin_lock_irq(&dev->power.lock);
      }

drivers/base/power/runtime.c: In function '__rpm_callback':
drivers/base/power/runtime.c:355:13: warning: 'get' may be used
uninitialized in this function [-Wmaybe-uninitialized]
error, forbidden warning:runtime.c:355

This really is a false-positive, because in fact the "get" variable
cannot be uninitialized at this point.

    put = get && retval;

There is a compilation error. I change it as:

put = dev->power.runtime_status == RPM_SUSPENDING && retval;

I prefer to initialize "get" to false (even though that is not
strictly necessary).

And test works well.Please check it.

OK, thanks for testing!

I will send an updated patch later today.