BMC/Intel-BMC/linux.git/drivers/base/power/domain.c, branch dev-5.1

PM / Domains: Avoid a potential deadlock

2019-03-19T21:58:04+00:00

Lockdep warns that prepare_lock and genpd->mlock can cause a deadlock the deadlock scenario is like following: First thread is probing cs2000 cs2000_probe() clk_register() __clk_core_init() clk_prepare_lock() ----> acquires prepare_lock cs2000_recalc_rate() i2c_smbus_read_byte_data() rcar_i2c_master_xfer() dma_request_chan() rcar_dmac_of_xlate() rcar_dmac_alloc_chan_resources() pm_runtime_get_sync() __pm_runtime_resume() rpm_resume() rpm_callback() genpd_runtime_resume() ----> acquires genpd->mlock Second thread is attaching any device to the same PM domain genpd_add_device() genpd_lock() ----> acquires genpd->mlock cpg_mssr_attach_dev() of_clk_get_from_provider() __of_clk_get_from_provider() __clk_create_clk() clk_prepare_lock() ----> acquires prepare_lock Since currently no PM provider access genpd's critical section in .attach_dev, and .detach_dev callbacks, so there is no need to protect these two callbacks with genpd->mlock. This patch avoids a potential deadlock by moving out .attach_dev and .detach_dev from genpd->mlock, so that genpd->mlock won't be held when prepare_lock is acquired in .attach_dev and .detach_dev Signed-off-by: Jiada Wang Reviewed-by: Ulf Hansson Tested-by: Geert Uytterhoeven Reviewed-by: Geert Uytterhoeven Signed-off-by: Rafael J. Wysocki

Merge branch 'pm-domains'

2019-03-14T09:53:08+00:00

* pm-domains: PM / domains: Remove one unnecessary blank line PM / Domains: Return early for all errors in _genpd_power_off() PM / Domains: Improve warn for multiple states but no governor

PM / Domains: Return early for all errors in _genpd_power_off()

2019-03-11T10:26:45+00:00

It is strange to only return early for -EBUSY state and left other errors to be still measured execution time. As for error cases, the elapsed_ns computed actually is not quite accurate and meaningful for governor to use. So let's simply return for all error cases. Signed-off-by: Dong Aisheng Reviewed-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki

PM / Domains: Improve warn for multiple states but no governor

2019-03-11T10:26:45+00:00

It's possible a PM domain defines only one state and it does not need a governor to work. For such case, a warning actually is not necessary. Signed-off-by: Dong Aisheng Reviewed-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki

PM: Add and use pr_fmt()

2019-03-07T10:02:02+00:00

Prefix all printk/pr_ messages with "PM: " to make the logging a bit more consistent. Miscellanea: o Convert a few printks to pr_ o Whitespace to align to open parentheses o Remove embedded "PM: " from pr_debugs as pr_fmt adds it Signed-off-by: Joe Perches Reviewed-by: Kees Cook Reviewed-by: Kevin Hilman Signed-off-by: Rafael J. Wysocki

PM / Domains: Mark "name" const in genpd_dev_pm_attach_by_name()

2019-02-15T10:36:15+00:00

The genpd_dev_pm_attach_by_name() simply takes the name and passes it to of_property_match_string() where the argument is "const char *". Adding a const here allows a later patch to add a const to dev_pm_domain_attach_by_name() which allows drivers to pass in a name that was declared "const" in a driver. Fixes: 5d6be70add65 ("PM / Domains: Introduce option to attach a device by name to genpd") Signed-off-by: Douglas Anderson Reviewed-by: Stephen Boyd Acked-by: Viresh Kumar Reviewed-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki

PM: domains: no need to check return value of debugfs_create functions

2019-01-23T21:59:57+00:00

When calling debugfs functions, there is no need to ever check the return value. The function can work or not, but the code logic should never do something different based on this. Signed-off-by: Greg Kroah-Hartman Reviewed-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki

Merge branches 'pm-core', 'pm-qos', 'pm-domains' and 'pm-sleep'

2018-12-21T09:06:44+00:00

* pm-core: PM-runtime: Switch autosuspend over to using hrtimers * pm-qos: PM / QoS: Change to use DEFINE_SHOW_ATTRIBUTE macro * pm-domains: PM / Domains: remove define_genpd_open_function() and define_genpd_debugfs_fops() * pm-sleep: PM / sleep: convert to DEFINE_SHOW_ATTRIBUTE

PM / Domains: remove define_genpd_open_function() and define_genpd_debugfs_fops()

2018-12-19T09:37:06+00:00

We already have the DEFINE_SHOW_ATTRIBUTE, There is no need to define such a macro, so remove define_genpd_open_function and define_genpd_debugfs_fops. Convert them to DEFINE_SHOW_ATTRIBUTE. Signed-off-by: Yangtao Li Acked-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki

PM / Domains: Propagate performance state updates

2018-12-14T10:58:18+00:00

Currently a genpd only handles the performance state requirements from the devices under its control. This commit extends that to also handle the performance state requirement(s) put on the master genpd by its sub-domains. There is a separate value required for each master that the genpd has and so a new field is added to the struct gpd_link (link->performance_state), which represents the link between a genpd and its master. The struct gpd_link also got another field prev_performance_state, which is used by genpd core as a temporary variable during transitions. On a call to dev_pm_genpd_set_performance_state(), the genpd core first updates the performance state of the masters of the device's genpd and then updates the performance state of the genpd. The masters do the same and propagate performance state updates to their masters before updating their own. The performance state transition from genpd to its master is done with the help of dev_pm_opp_xlate_performance_state(), which looks at the OPP tables of both the domains to translate the state. Tested-by: Rajendra Nayak Reviewed-by: Ulf Hansson Signed-off-by: Viresh Kumar