On 10/4/24 11:08, Jeffrey Hugo wrote:
On 9/11/2024 12:06 PM, Lizhi Hou wrote:+static int amdxdna_rpmops_suspend(struct device *dev) +{ + struct amdxdna_dev *xdna = pci_get_drvdata(to_pci_dev(dev)); + int ret; + + mutex_lock(&xdna->dev_lock); + WARN_ON(!list_empty(&xdna->client_list));This feels weird. Can you explain?
The driver uses auto suspend. When there is opened client, auto suspend is held. So the client list should not be empty here. The WARN_ON is kind of assert and can be removed.
Thanks, Lizhi
+ ret = amdxdna_dev_suspend_nolock(xdna); + mutex_unlock(&xdna->dev_lock); + + XDNA_DBG(xdna, "Runtime suspend done ret: %d", ret); + return ret; +} +
