thermal/core: Encapsulate the set_cur_state function
PATCH 1/3 thermal/core: Encapsulate the set_cur_state function
Concentrate the actions in a single place when a cooling device state is changed. Provide a function to do that instead of calling the underlying ops.
--- drivers/thermal/thermal_core.h | 1 + drivers/thermal/thermal_helpers.c | 32 ++++++++++++++++++++++++------- drivers/thermal/thermal_sysfs.c | 7 ++----- 3 files changed, 28 insertions(+), 12 deletions(-)
drivers/thermal/thermal_core.h +1 -0
@@ -65,6 +65,7 @@ static inline bool cdev_is_power_actor(struct thermal_cooling_device *cdev) cdev->ops->power2state; } +int thermal_cdev_set_state (struct thermal_cooling_device *cdev,int state);void thermal_cdev_update (struct thermal_cooling_device *);void __thermal_cdev_update (struct thermal_cooling_device *cdev);
drivers/thermal/thermal_helpers.c +25 -7
@@ -182,14 +182,32 @@ void thermal_set_delay_jiffies(unsigned long *delay_jiffies, int delay_ms) *delay_jiffies =round_jiffies (*delay_jiffies); } -static void thermal_cdev_set_cur_state (struct thermal_cooling_device *cdev, -int target) +/** +* thermal_cdev_set_state - set the cooling device state +* @cdev: a pointer to a thermal_cooling_device +* @state: the target state +* +* Set the state of the cooling device passed as parameter. The +* cooling device lock must be held when calling this function. +* +* Return: 0 in case of success, otherwise the return value is the one +* returned by the backend for the ops +*/ +int thermal_cdev_set_state (struct thermal_cooling_device *cdev,int state) { -if (cdev->ops->set_cur_state (cdev, target)) -return ; +int ret; -thermal_notify_cdev_state_update (cdev->id, target); -thermal_cooling_device_stats_update (cdev, target); +/* +* No check is needed for the ops->set_cur_state as the +* registering function checked the ops are correctly set +*/ + ret = cdev->ops->set_cur_state (cdev, state); +if (!ret) { +thermal_notify_cdev_state_update (cdev->id, state); +thermal_cooling_device_stats_update (cdev, state); + } + +return ret; }void __thermal_cdev_update (struct thermal_cooling_device *cdev)
@@ -207,7 +225,7 @@ void __thermal_cdev_update(struct thermal_cooling_device *cdev) target = instance->target; } -thermal_cdev_set_cur_state (cdev, target); +thermal_cdev_set_state (cdev, target);trace_cdev_update (cdev, target);dev_dbg (&cdev->device,"set to state %lu\n" , target);
drivers/thermal/thermal_sysfs.c +2 -5
@@ -617,12 +617,9 @@ cur_state_store(struct device *dev, struct device_attribute *attr,return -EINVAL ;mutex_lock (&cdev->lock); - - result = cdev->ops->set_cur_state (cdev, state); -if (!result) -thermal_cooling_device_stats_update (cdev, state); - + result =thermal_cdev_set_state (cdev, state);mutex_unlock (&cdev->lock); +return result ? result : count; }
Signature · 1 line
2.25.1
PATCH 2/3 thermal/debugfs: Add debugfs information
The thermal framework does not have any debug information except a sysfs stat which is a bit controversial. This one allocates big chunks of memory for every cooling devices with a high number of states and could represent on some systems in production several megabytes of memory for just a portion of it. As the syfs is limited to a page size, the output is not exploitable with large data array and gets truncated.
The patch provides the same information than sysfs except the transitions are dynamically allocated, thus they won't represent more events than the ones which actually occured. There is no longer a size limitation and it opens the field for more debugging information where the debugfs is designed for, not sysfs.
--- drivers/thermal/Kconfig | 7 + drivers/thermal/Makefile | 3 + drivers/thermal/thermal_core.c | 10 + drivers/thermal/thermal_core.h | 1 + drivers/thermal/thermal_debugfs.c | 362 ++++++++++++++++++++++++++++++ drivers/thermal/thermal_debugfs.h | 17 ++ drivers/thermal/thermal_helpers.c | 1 + include/linux/thermal.h | 7 + 8 files changed, 408 insertions(+) create mode 100644 drivers/thermal/thermal_debugfs.c create mode 100644 drivers/thermal/thermal_debugfs.h
drivers/thermal/Kconfig +7 -0
@@ -33,6 +33,13 @@ config THERMAL_STATISTICS If in doubt, say N. +config THERMAL_DEBUGFS + bool "Thermal debugging file system" + depends on DEBUG_FS + help + This option provides a debugfs entry giving useful + information about the thermal framework internals. + config THERMAL_EMERGENCY_POWEROFF_DELAY_MS int "Emergency poweroff delay in milli-seconds" default 0
drivers/thermal/Makefile +3 -0
@@ -10,6 +10,9 @@ thermal_sys-y += thermal_core.o thermal_sysfs.o \ # netlink interface to manage the thermal framework thermal_sys-$(CONFIG_THERMAL_NETLINK) += thermal_netlink.o +# debugfs interface to investigate the behavior and statistics +thermal_sys-$(CONFIG_THERMAL_DEBUGFS) += thermal_debugfs.o + # interface to/from other layers providing sensors thermal_sys-$(CONFIG_THERMAL_HWMON) += thermal_hwmon.o thermal_sys-$(CONFIG_THERMAL_OF) += thermal_of.o
drivers/thermal/thermal_core.c +10 -0
@@ -944,6 +944,8 @@ __thermal_cooling_device_register(struct device_node *np, THERMAL_EVENT_UNSPECIFIED);mutex_unlock (&thermal_list_lock); +thermal_debugfs_cdev_register (cdev); +return cdev; out_kfree_type:
@@ -1079,6 +1081,8 @@ void thermal_cooling_device_unregister(struct thermal_cooling_device *cdev)if (!cdev)return ; +thermal_debugfs_cdev_unregister (cdev); +mutex_lock (&thermal_list_lock);list_for_each_entry (pos, &thermal_cdev_list, node)if (pos == cdev)
@@ -1311,6 +1315,8 @@ thermal_zone_device_register(const char *type, int trips, int mask,thermal_notify_tz_create (tz->id, tz->type); +thermal_debugfs_tz_register (tz); +return tz; unregister:
@@ -1340,6 +1346,8 @@ void thermal_zone_device_unregister(struct thermal_zone_device *tz)if (!tz)return ; +thermal_debugfs_tz_unregister (tz); + tzp = tz->tzp; tz_id = tz->id;
@@ -1485,6 +1493,8 @@ static int __init thermal_init(void)pr_warn ("Thermal: Can not register suspend notifier, return %d\n" , result); +thermal_debugfs_init (); +return 0 ; unregister_class:
drivers/thermal/thermal_core.h +1 -0
@@ -13,6 +13,7 @@#include <linux/thermal.h> #include "thermal_netlink.h" +#include "thermal_debugfs.h" /* Default Thermal Governor */ #if defined(CONFIG_THERMAL_DEFAULT_GOV_STEP_WISE )
drivers/thermal/thermal_debugfs.c new file +362 -0
362 lines not on this page · raw message
drivers/thermal/thermal_debugfs.h new file +17 -0
@@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#ifdef CONFIG_THERMAL_DEBUGFS +void thermal_debugfs_init (void ); +void thermal_debugfs_cdev_register (struct thermal_cooling_device *cdev); +void thermal_debugfs_cdev_unregister (struct thermal_cooling_device *cdev); +void thermal_debugfs_tz_register (struct thermal_zone_device *tz); +void thermal_debugfs_tz_unregister (struct thermal_zone_device *tz); +void thermal_debugfs_cdev_transition (struct thermal_cooling_device *cdev,int state); +#else +static inline void thermal_debugfs_init (void ); +static inline void thermal_debugfs_cdev_register (struct thermal_cooling_device *cdev) {} +static inline void thermal_debugfs_cdev_unregister (struct thermal_cooling_device *cdev) {} +static inline void thermal_debugfs_tz_register (struct thermal_zone_device *tz) {} +static inline void thermal_debugfs_tz_unregister (struct thermal_zone_device *tz) {} +static inline void thermal_debugfs_cdev_transition (struct thermal_cooling_device *cdev,int state) {} +#endif /* CONFIG_THERMAL_DEBUGFS */
drivers/thermal/thermal_helpers.c +1 -0
@@ -205,6 +205,7 @@ int thermal_cdev_set_state(struct thermal_cooling_device *cdev, int state)if (!ret) {thermal_notify_cdev_state_update (cdev->id, state);thermal_cooling_device_stats_update (cdev, state); +thermal_debugfs_cdev_transition (cdev, state); }return ret;
include/linux/thermal.h +7 -0
@@ -34,6 +34,7 @@struct thermal_zone_device ;struct thermal_cooling_device ;struct thermal_instance ; +struct thermal_debugfs ;struct thermal_attr ;enum thermal_trend {
@@ -101,6 +102,9 @@ struct thermal_cooling_device {struct mutex lock;/* protect thermal_instances list */ struct list_head thermal_instances;struct list_head node; +#ifdef CONFIG_THERMAL_DEBUGFS +struct thermal_debugfs *debugfs; +#endif };/**
@@ -174,6 +178,9 @@ struct thermal_zone_device {struct list_head node;struct delayed_work poll_queue;enum thermal_notify_event notify_event; +#ifdef CONFIG_THERMAL_DEBUGFS +struct thermal_debugfs *debugfs; +#endif };/**
Signature · 1 line
2.25.1
kernel test robot
build botRe: [PATCH 2/3] thermal/debugfs: Add debugfs information
I love your patch! Yet something to improve:
Hi Daniel,
I love your patch! Yet something to improve:
[auto build test ERROR on rafael-pm/thermal]
[also build test ERROR on v5.18 next-20220601]
[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/intel-lab-lkp/linux/commits/Daniel-Lezcano/thermal-core-Encapsulate-the-set_cur_state-function/20220601-231733 base: https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git thermal config: x86_64-randconfig-a011 (https://download.01.org/0day-ci/archive/20220602/[email protected]/config) compiler: gcc-11 (Debian 11.3.0-1) 11.3.0 reproduce (this is a W=1 build): # https://github.com/intel-lab-lkp/linux/commit/04c295f7e8b49af742179609949736f6f056b49c git remote add linux-review https://github.com/intel-lab-lkp/linux git fetch --no-tags linux-review Daniel-Lezcano/thermal-core-Encapsulate-the-set_cur_state-function/20220601-231733 git checkout 04c295f7e8b49af742179609949736f6f056b49c # save the config file mkdir build_dir && cp config build_dir/.config make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash
If you fix the issue, kindly add following tag where applicable
All errors (new ones prefixed by >>):
ld: drivers/thermal/thermal_core.o: in function `thermal_init':
>> drivers/thermal/thermal_core.c:1496: undefined reference to `thermal_debugfs_init'
vim +1496 drivers/thermal/thermal_core.c
1470
1471 static int __init thermal_init(void)
1472 {
1473 int result;
1474
1475 result = thermal_netlink_init();
1476 if (result)
1477 goto error;
1478
1479 result = thermal_register_governors();
1480 if (result)
1481 goto error;
1482
1483 result = class_register(&thermal_class);
1484 if (result)
1485 goto unregister_governors;
1486
1487 result = of_parse_thermal_zones();
1488 if (result)
1489 goto unregister_class;
1490
1491 result = register_pm_notifier(&thermal_pm_nb);
1492 if (result)
1493 pr_warn("Thermal: Can not register suspend notifier, return %d\n",
1494 result);
1495
> 1496 thermal_debugfs_init();
Signature · 2 lines
0-DAY CI Kernel Test Service https://01.org/lkp
kernel test robot
build botRe: [PATCH 2/3] thermal/debugfs: Add debugfs information
In reply to Daniel LezcanoI love your patch! Perhaps something to improve:
Hi Daniel,
I love your patch! Perhaps something to improve:
[auto build test WARNING on rafael-pm/thermal]
[also build test WARNING on v5.18 next-20220601]
[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/intel-lab-lkp/linux/commits/Daniel-Lezcano/thermal-core-Encapsulate-the-set_cur_state-function/20220601-231733 base: https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git thermal config: x86_64-randconfig-a016 (https://download.01.org/0day-ci/archive/20220602/[email protected]/config) compiler: clang version 15.0.0 (https://github.com/llvm/llvm-project c825abd6b0198fb088d9752f556a70705bc99dfd)
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# https://github.com/intel-lab-lkp/linux/commit/04c295f7e8b49af742179609949736f6f056b49c
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Daniel-Lezcano/thermal-core-Encapsulate-the-set_cur_state-function/20220601-231733 git checkout 04c295f7e8b49af742179609949736f6f056b49c
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash drivers/
If you fix the issue, kindly add following tag where applicable
All warnings (new ones prefixed by >>):
In file included from drivers/thermal/thermal_core.c:27:
In file included from drivers/thermal/thermal_core.h:16:
>> drivers/thermal/thermal_debugfs.h:11:20: warning: function 'thermal_debugfs_init' has internal linkage but is not defined [-Wundefined-internal]
static inline void thermal_debugfs_init(void);
^
drivers/thermal/thermal_core.c:1496:2: note: used here
thermal_debugfs_init();
^1 warning generated.
vim +/thermal_debugfs_init +11 drivers/thermal/thermal_debugfs.h
2
3 #ifdef CONFIG_THERMAL_DEBUGFS
4 void thermal_debugfs_init(void);
5 void thermal_debugfs_cdev_register(struct thermal_cooling_device *cdev);
6 void thermal_debugfs_cdev_unregister(struct thermal_cooling_device *cdev);
7 void thermal_debugfs_tz_register(struct thermal_zone_device *tz);
8 void thermal_debugfs_tz_unregister(struct thermal_zone_device *tz);
9 void thermal_debugfs_cdev_transition(struct thermal_cooling_device *cdev, int state);
10 #else
> 11 static inline void thermal_debugfs_init(void);Signature · 2 lines
0-DAY CI Kernel Test Service https://01.org/lkp
kernel test robot
build botRe: [PATCH 2/3] thermal/debugfs: Add debugfs information
In reply to Daniel LezcanoI love your patch! Perhaps something to improve:
Hi Daniel,
I love your patch! Perhaps something to improve:
[auto build test WARNING on rafael-pm/thermal]
[also build test WARNING on v5.18 next-20220602]
[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/intel-lab-lkp/linux/commits/Daniel-Lezcano/thermal-core-Encapsulate-the-set_cur_state-function/20220601-231733 base: https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git thermal config: i386-randconfig-s001 (https://download.01.org/0day-ci/archive/20220603/[email protected]/config) compiler: gcc-11 (Debian 11.3.0-1) 11.3.0 reproduce: # apt-get install sparse # sparse version: v0.6.4-14-g5a0004b5-dirty # https://github.com/intel-lab-lkp/linux/commit/04c295f7e8b49af742179609949736f6f056b49c git remote add linux-review https://github.com/intel-lab-lkp/linux git fetch --no-tags linux-review Daniel-Lezcano/thermal-core-Encapsulate-the-set_cur_state-function/20220601-231733 git checkout 04c295f7e8b49af742179609949736f6f056b49c # save the config file mkdir build_dir && cp config build_dir/.config make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' O=build_dir ARCH=i386 SHELL=/bin/bash
If you fix the issue, kindly add following tag where applicable
sparse warnings: (new ones prefixed by >>)
drivers/thermal/thermal_core.c: note: in included file (through drivers/thermal/thermal_core.h):
>> drivers/thermal/thermal_debugfs.h:11:40: sparse: sparse: marked inline, but without a definition
vim +11 drivers/thermal/thermal_debugfs.h
2
3 #ifdef CONFIG_THERMAL_DEBUGFS
4 void thermal_debugfs_init(void);
5 void thermal_debugfs_cdev_register(struct thermal_cooling_device *cdev);
6 void thermal_debugfs_cdev_unregister(struct thermal_cooling_device *cdev);
7 void thermal_debugfs_tz_register(struct thermal_zone_device *tz);
8 void thermal_debugfs_tz_unregister(struct thermal_zone_device *tz);
9 void thermal_debugfs_cdev_transition(struct thermal_cooling_device *cdev, int state);
10 #else
> 11 static inline void thermal_debugfs_init(void);Signature · 2 lines
0-DAY CI Kernel Test Service https://01.org/lkp
PATCH 3/3 thermal/sysfs: Remove cooling device sysfs statistics
In reply to Daniel LezcanoThe statistics are for debugging purpose and belong to debugfs rather than sysfs. As the previous changes introduced the same statistics in debugfs, those in sysfs are no longer needed and can be removed.
That solves a couple of problems reported by different SoC vendors:
- Extra memory used in low profile boards where the memory is limited resource
- Truncated information and memory allocation failure with cooling devices having a large number of states
--- drivers/thermal/Kconfig | 7 - drivers/thermal/thermal_core.c | 2 - drivers/thermal/thermal_core.h | 10 -- drivers/thermal/thermal_helpers.c | 1 - drivers/thermal/thermal_sysfs.c | 217 ------------------------------ include/linux/thermal.h | 1 - 6 files changed, 238 deletions(-)
drivers/thermal/Kconfig +0 -7
@@ -26,13 +26,6 @@ config THERMAL_NETLINK trip point crossed, cooling device update or governor change. It is recommended to enable the feature. -config THERMAL_STATISTICS - bool "Thermal state transition statistics" - help - Export thermal state transition statistics information through sysfs. - - If in doubt, say N. - config THERMAL_DEBUGFS bool "Thermal debugging file system" depends on DEBUG_FS
drivers/thermal/thermal_core.c +0 -2
@@ -949,7 +949,6 @@ __thermal_cooling_device_register(struct device_node *np,return cdev; out_kfree_type: -thermal_cooling_device_destroy_sysfs (cdev);kfree (cdev->type);put_device (&cdev->device); cdev =NULL ;
@@ -1117,7 +1116,6 @@ void thermal_cooling_device_unregister(struct thermal_cooling_device *cdev)ida_simple_remove (&thermal_cdev_ida, cdev->id);device_del (&cdev->device); -thermal_cooling_device_destroy_sysfs (cdev);kfree (cdev->type);put_device (&cdev->device); }
drivers/thermal/thermal_core.h +0 -10
@@ -134,22 +134,12 @@ void thermal_set_delay_jiffies(unsigned long *delay_jiffies, int delay_ms);int thermal_zone_create_device_groups (struct thermal_zone_device *,int );void thermal_zone_destroy_device_groups (struct thermal_zone_device *);void thermal_cooling_device_setup_sysfs (struct thermal_cooling_device *); -void thermal_cooling_device_destroy_sysfs (struct thermal_cooling_device *cdev);/* used only at binding time */ ssize_t trip_point_show (struct device *,struct device_attribute *,char *);ssize_t weight_show (struct device *,struct device_attribute *,char *);ssize_t weight_store (struct device *,struct device_attribute *,const char *,size_t ); -#ifdef CONFIG_THERMAL_STATISTICS -void thermal_cooling_device_stats_update (struct thermal_cooling_device *cdev, -unsigned long new_state); -#else -static inline void -thermal_cooling_device_stats_update (struct thermal_cooling_device *cdev, -unsigned long new_state) {} -#endif /* CONFIG_THERMAL_STATISTICS */ -/* device tree support */ #ifdef CONFIG_THERMAL_OF int of_parse_thermal_zones (void );
drivers/thermal/thermal_helpers.c +0 -1
@@ -204,7 +204,6 @@ int thermal_cdev_set_state(struct thermal_cooling_device *cdev, int state) ret = cdev->ops->set_cur_state (cdev, state);if (!ret) {thermal_notify_cdev_state_update (cdev->id, state); -thermal_cooling_device_stats_update (cdev, state);thermal_debugfs_cdev_transition (cdev, state); }
drivers/thermal/thermal_sysfs.c +0 -217
231 lines not on this page · raw message
include/linux/thermal.h +0 -1
@@ -96,7 +96,6 @@ struct thermal_cooling_device {struct device device;struct device_node *np;void *devdata; -void *stats;const struct thermal_cooling_device_ops *ops;bool updated;/* true if the cooling device does not need update */ struct mutex lock;/* protect thermal_instances list */
Signature · 1 line
2.25.1
Lukasz Luba
reviewerRe: [PATCH 3/3] thermal/sysfs: Remove cooling device sysfs statistics
Hi Daniel,
Quoted 3 lines · Daniel Lezcano
The statistics are for debugging purpose and belong to debugfs rather than sysfs. As the previous changes introduced the same statistics in debugfs, those in sysfs are no longer needed and can be removed.
I just want to let you know that in current Android kernels we cannot even compile the kernel with CONFIG_DEBUG_FS. I have this pain with Energy Model there... Some vendors might see useful info via this sysfs interface in bring-up of the SoC.
I don't know if there are user-space tools tracking this information via sysfs. We probably should check that.
I agree that these statistics look more like debug info, rather than something useful for control.
Furthermore, we have trace events for the cooling state changes, which should be good enough for bring-up and experiments.
I don't have strong preferences here. I tend to agree to remove this interface if there are no user-space tools using it.
Regards,
Lukasz
Re: [PATCH 3/3] thermal/sysfs: Remove cooling device sysfs statistics
Hi Lukasz,
[Adding Todd]
Quoted 10 lines · Lukasz Luba
Hi Daniel,
Quoted 3 lines · Daniel Lezcano
The statistics are for debugging purpose and belong to debugfs rather than sysfs. As the previous changes introduced the same statistics in debugfs, those in sysfs are no longer needed and can be removed.
I just want to let you know that in current Android kernels we cannot even compile the kernel with
CONFIG_DEBUG_FS.
Right, it makes sense. Precisely, with the sysfs stats they are always compiled in for the Android kernel and is a problem for low memory systems. While debugfs can fulfill its purpose in the developement and will be removed in production systems.
Quoted 3 lines
I have this pain with
Energy Model there... Some vendors might see useful info via this sysfs interface in bring-up of the SoC.
Well alternatively, information can be extracted from procfs in the device-tree description.
What prevents to add energy information in sysfs now that the energy model is per device ?
Quoted 11 lines
I don't know if there are user-space tools tracking this information via sysfs. We probably should check that.
I agree that these statistics look more like debug info, rather than something useful for control.
Furthermore, we have trace events for the cooling state changes, which should be good enough for bring-up and experiments.
I don't have strong preferences here. I tend to agree to remove this interface if there are no user-space tools using it.
I agree userspace can also get information about the transition but the goal of the debugfs is also add information about thermal internals like average temperature at mitigation time, min and max, timings, etc ...
Signature · 5 lines
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | <http://twitter.com/#!/linaroorg> Twitter | <http://www.linaro.org/linaro-blog/> Blog
Lukasz Luba
reviewerRe: [PATCH 3/3] thermal/sysfs: Remove cooling device sysfs statistics
Quoted 10 lines · Daniel Lezcano
Hi Lukasz,
[Adding Todd]
Quoted 4 lines · Lukasz Luba
Hi Daniel,
On 6/1/22 16:14, Daniel Lezcano wrote:
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
The statistics are for debugging purpose and belong to debugfs rather than sysfs. As the previous changes introduced the same statistics in debugfs, those in sysfs are no longer needed and can be removed.
I just want to let you know that in current Android kernels we cannot even compile the kernel with
CONFIG_DEBUG_FS.Right, it makes sense. Precisely, with the sysfs stats they are always compiled in for the Android kernel and is a problem for low memory systems. While debugfs can fulfill its purpose in the developement and will be removed in production systems.
True.
Quoted 8 lines
Quoted 4 lines
I have this pain with
Energy Model there... Some vendors might see useful info via this sysfs interface in bring-up of the SoC.Well alternatively, information can be extracted from procfs in the device-tree description.
What prevents to add energy information in sysfs now that the energy model is per device ?
Probably nothing, but we need strong need. I have proposed this a few times internally, but this must have a requirement. If a user-space tool would ask for it, then I could send a patch exposing the sysfs. So far we have only one user-space tool, which suffers the missing debugfs EM dir: LISA (but we are working on a workaround for it).
If you have a tool or plan to have such, which uses EM, please let me know. I'm gathering the requirements.
Quoted 7 lines
Quoted 7 lines
I don't know if there are user-space tools tracking this information via sysfs. We probably should check that.
I agree that these statistics look more like debug info, rather than something useful for control.
Quoted 9 lines
Quoted 5 lines
Furthermore, we have trace events for the cooling state changes, which should be good enough for bring-up and experiments.
I don't have strong preferences here. I tend to agree to remove this interface if there are no user-space tools using it.
I agree userspace can also get information about the transition but the goal of the debugfs is also add information about thermal internals like average temperature at mitigation time, min and max, timings, etc ...
I see, it makes sense. Let's see if Todd and Android folks don't use this thermal sysfs stats, so we could remove them.
Re: [PATCH 3/3] thermal/sysfs: Remove cooling device sysfs statistics
Quoted 61 lines · Lukasz Luba
Quoted 14 lines · Daniel Lezcano
Hi Lukasz,
[Adding Todd]
Quoted 8 lines · Lukasz Luba
Hi Daniel,
Quoted 3 lines · Daniel Lezcano
The statistics are for debugging purpose and belong to debugfs rather than sysfs. As the previous changes introduced the same statistics in debugfs, those in sysfs are no longer needed and can be removed.
Quoted 7 lines
Quoted 2 lines
I just want to let you know that in current Android kernels we cannot even compile the kernel with
CONFIG_DEBUG_FS.Right, it makes sense. Precisely, with the sysfs stats they are always compiled in for the Android kernel and is a problem for low memory systems. While debugfs can fulfill its purpose in the developement and will be removed in production systems.
True.
Quoted 10 lines
Quoted 4 lines
I have this pain with
Energy Model there... Some vendors might see useful info via this sysfs interface in bring-up of the SoC.Well alternatively, information can be extracted from procfs in the device-tree description.
What prevents to add energy information in sysfs now that the energy model is per device ?
Probably nothing, but we need strong need. I have proposed this a few times internally, but this must have a requirement. If a user-space tool would ask for it, then I could send a patch exposing the sysfs. So far we have only one user-space tool, which suffers the missing debugfs EM dir: LISA (but we are working on a workaround for it).
If you have a tool or plan to have such, which uses EM, please let me know. I'm gathering the requirements.Quoted 13 lines
Quoted 12 lines
I don't know if there are user-space tools tracking this information via sysfs. We probably should check that.
I agree that these statistics look more like debug info, rather than something useful for control.
Furthermore, we have trace events for the cooling state changes, which should be good enough for bring-up and experiments.
I don't have strong preferences here. I tend to agree to remove this interface if there are no user-space tools using it.
Quoted 8 lines
Quoted 5 lines
I agree userspace can also get information about the transition but the goal of the debugfs is also add information about thermal internals like average temperature at mitigation time, min and max, timings, etc ...
I see, it makes sense. Let's see if Todd and Android folks don't use this thermal sysfs stats, so we could remove them.
Android HALs do use the thermal sysfs stats. debugfs isn't a viable replacement since debugfs must not be mounted during normal operation.
Re: [PATCH 3/3] thermal/sysfs: Remove cooling device sysfs statistics
Hi Todd,
[adding Wei]
Quoted 1 line · Todd Kjos
On Thu, Jun 2, 2022 at 2:16 AM Lukasz Luba <[email protected]> wrote:
[ ... ]
Quoted 5 lines
Quoted 2 lines
I see, it makes sense. Let's see if Todd and Android folks don't use this thermal sysfs stats, so we could remove them.
Android HALs do use the thermal sysfs stats. debugfs isn't a viable replacement since debugfs must not be mounted during normal operation.
Thanks for your answer.
I'm curious, what is the purpose of getting the statistics, especially the transitions stats from normal operation?
There were some complains about systems having a high number of cooling devices with a lot of states. The state transitions are represented as a matrix and result in up to hundred of megabytes of memory wasted.
Moreover, sysfs being limited a page size, the output is often truncated.
As it is automatically enabled for GKI, this waste of memory which is not negligible for system with low memory can not be avoided.
I went through the thermal HAL but did not find an usage of these statistics, do you have a pointer to the code using them ?
Thanks
-- Daniel
Signature · 5 lines
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | <http://twitter.com/#!/linaroorg> Twitter | <http://www.linaro.org/linaro-blog/> Blog
Re: [PATCH 3/3] thermal/sysfs: Remove cooling device sysfs statistics
Quoted 31 lines · Daniel Lezcano
Hi Todd,
[adding Wei]
Quoted 1 line · Todd Kjos
On Thu, Jun 2, 2022 at 2:16 AM Lukasz Luba <[email protected]> wrote:
[ ... ]
Quoted 5 lines
Quoted 2 lines
I see, it makes sense. Let's see if Todd and Android folks don't use this thermal sysfs stats, so we could remove them.
Android HALs do use the thermal sysfs stats. debugfs isn't a viable replacement since debugfs must not be mounted during normal operation.
Thanks for your answer.
I'm curious, what is the purpose of getting the statistics, especially the transitions stats from normal operation?
There were some complains about systems having a high number of cooling devices with a lot of states. The state transitions are represented as a matrix and result in up to hundred of megabytes of memory wasted.
Moreover, sysfs being limited a page size, the output is often truncated.
As it is automatically enabled for GKI, this waste of memory which is not negligible for system with low memory can not be avoided.
Quoted 8 lines
I went through the thermal HAL but did not find an usage of these statistics, do you have a pointer to the code using them ?
Thanks
-- Daniel
Sorry for the late reply, trying to catch up on emails after sick recovery. We use it for stats collection to understand thermal residency, and it is not in the HAL code, we don't use the transition table heavily though. Are some of the devices having too many cooling devices? Can we have a config to enable stats for a given cooling device?
Thanks!
-Wei
Quoted 7 lines
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCsFollow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
Re: [PATCH 3/3] thermal/sysfs: Remove cooling device sysfs statistics
Quoted 41 lines · Wei Wang
Quoted 31 lines · Daniel Lezcano
Hi Todd,
[adding Wei]
Quoted 1 line · Todd Kjos
On Thu, Jun 2, 2022 at 2:16 AM Lukasz Luba <[email protected]> wrote:
[ ... ]
Quoted 5 lines
Quoted 2 lines
I see, it makes sense. Let's see if Todd and Android folks don't use this thermal sysfs stats, so we could remove them.
Android HALs do use the thermal sysfs stats. debugfs isn't a viable replacement since debugfs must not be mounted during normal operation.
Thanks for your answer.
I'm curious, what is the purpose of getting the statistics, especially the transitions stats from normal operation?
There were some complains about systems having a high number of cooling devices with a lot of states. The state transitions are represented as a matrix and result in up to hundred of megabytes of memory wasted.
Moreover, sysfs being limited a page size, the output is often truncated.
As it is automatically enabled for GKI, this waste of memory which is not negligible for system with low memory can not be avoided.
Quoted 8 lines
I went through the thermal HAL but did not find an usage of these statistics, do you have a pointer to the code using them ?
Thanks
-- Daniel
Quoted 6 lines
Sorry for the late reply, trying to catch up on emails after sick recovery. We use it for stats collection to understand thermal residency, and it is not in the HAL code, we don't use the transition table heavily though. Are some of the devices having too many cooling devices? Can we have a config to enable stats for a given cooling device?
The stats table is bogus. As soon as the combination of the states leads to a size greater than a page size, then the result is truncated.
As the cooling devices is also abused to mimic power capping capable device, we are ending up to 1024 states sometimes.
Moreover, having the option set also create tables which take MB of memory for nothing.
The option only enables the stats for all the cooling device stats.
As you mentioned, it seems to you are using the stats for debugging purpose. The debugfs provides the same information, except it does only show the transitions would actually happened and we are no longer limited to a page size.
Would it be acceptable to remove the sysfs stats ?
Signature · 5 lines
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | <http://twitter.com/#!/linaroorg> Twitter | <http://www.linaro.org/linaro-blog/> Blog