kernel/linux.git/drivers/base/power/opp/cpu.c, branch v4.4.151

PM / OPP: Protect updates to list_dev with mutex

2015-11-06T22:59:41+00:00

dev_opp_list_lock is used everywhere to protect device and OPP lists, but dev_pm_opp_set_sharing_cpus() is missed somehow. And instead we used rcu-lock, which wouldn't help here as we are adding a new list_dev. This also fixes a problem where we have called kzalloc(..., GFP_KERNEL) from within rcu-lock, which isn't allowed as kzalloc can sleep when called with GFP_KERNEL. With CONFIG_DEBUG_ATOMIC_SLEEP set, we get following lockdep-splat: include/linux/rcupdate.h:578 Illegal context switch in RCU read-side critical section! other info that might help us debug this: rcu_scheduler_active = 1, debug_locks = 0 5 locks held by swapper/0/1: #0: (&dev->mutex){......}, at: [] __driver_attach+0x48/0x98 #1: (&dev->mutex){......}, at: [] __driver_attach+0x58/0x98 #2: (cpu_hotplug.lock){++++++}, at: [] get_online_cpus+0x40/0xb0 #3: (subsys mutex#5){+.+.+.}, at: [] subsys_interface_register+0x44/0xdc #4: (rcu_read_lock){......}, at: [] dev_pm_opp_set_sharing_cpus+0x0/0x1e4 stack backtrace: CPU: 1 PID: 1 Comm: swapper/0 Tainted: G W 4.3.0-rc7-00047-g81f5932958a8 #59 Hardware name: SAMSUNG EXYNOS (Flattened Device Tree) [] (unwind_backtrace) from [] (show_stack+0x10/0x14) [] (show_stack) from [] (dump_stack+0x94/0xbc) [] (dump_stack) from [] (___might_sleep+0x24c/0x298) [] (___might_sleep) from [] (kmem_cache_alloc+0xe8/0x164) [] (kmem_cache_alloc) from [] (_add_list_dev+0x30/0x58) [] (_add_list_dev) from [] (dev_pm_opp_set_sharing_cpus+0xd0/0x1e4) [] (dev_pm_opp_set_sharing_cpus) from [] (cpufreq_init+0x4cc/0x62c) [] (cpufreq_init) from [] (cpufreq_online+0xbc/0x73c) [] (cpufreq_online) from [] (subsys_interface_register+0x98/0xdc) [] (subsys_interface_register) from [] (cpufreq_register_driver+0x110/0x17c) [] (cpufreq_register_driver) from [] (dt_cpufreq_probe+0x60/0x8c) [] (dt_cpufreq_probe) from [] (platform_drv_probe+0x44/0xa4) [] (platform_drv_probe) from [] (driver_probe_device+0x208/0x2f4) [] (driver_probe_device) from [] (__driver_attach+0x94/0x98) [] (__driver_attach) from [] (bus_for_each_dev+0x68/0x9c) Reported-by: Michael Turquette Reviewed-by: Stephen Boyd Signed-off-by: Viresh Kumar Cc: 4.3 # 4.3 Signed-off-by: Rafael J. Wysocki

PM / OPP: Propagate error properly from dev_pm_opp_set_sharing_cpus()

2015-11-06T22:59:41+00:00

We are returning 0 even in case of errors, fix it. Fixes: 8d4d4e98acd6 ("PM / OPP: Add helpers for initializing CPU OPPs") Reported-by: Dan Carpenter Reviewed-by: Stephen Boyd Signed-off-by: Viresh Kumar Cc: 4.3 # 4.3 Signed-off-by: Rafael J. Wysocki

PM / OPP: Improve print messages with pr_fmt

2015-11-02T00:47:43+00:00

To identify OPP core's print messages easily, prefix them with KBUILD_MODNAME. Signed-off-by: Viresh Kumar Reviewed-by: Stephen Boyd Signed-off-by: Rafael J. Wysocki

PM / OPP: passing NULL to PTR_ERR()

2015-10-09T20:17:34+00:00

The code was using PTR_ERR(NULL) which causes a static checker warning. I have fixed up the printks and changed the return to -ENOENT. Signed-off-by: Dan Carpenter Acked-by: Viresh Kumar Signed-off-by: Rafael J. Wysocki

PM / OPP: Move cpu specific code to opp/cpu.c

2015-09-15T00:03:16+00:00

Move cpu device specific code out of generic opp library, and add it to cpu.c. Along with that, create a core-internal opp.h header, which will be used to share structures and function prototypes within opp core. Reviewed-by: Stephen Boyd Signed-off-by: Viresh Kumar Signed-off-by: Rafael J. Wysocki

PM / OPP: Move opp core to its own directory

2015-09-15T00:03:16+00:00

OPP code is expanding and is already present in multiple directories (cpufreq and power). Lets move it to its own directory, to manage it better. This also moves/renames the cpufreq_opp file to cpu.c, as it will contain helpers for cpu device. Its not just about cpufreq, other frameworks can use OPPs as well. Reviewed-by: Stephen Boyd Signed-off-by: Viresh Kumar Signed-off-by: Rafael J. Wysocki