led: add function naming option from linux
PATCH v3 1/1 led: add function naming option from linux
in linux we have the option to create the name of a led optionally through the following properties:
- function
- color
- function-enumerator
This patch adds support for parsing this properties if there is no label property.
To be as close as possible to linux import the following files from it:
include/dt-bindings/leds/common.hinclude/linux/uapi/linux/uleds.h
base commit was:
commit: b8f52214c61a ("Merge tag 'audit-pr-20241205' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit")
The led name is created in led_post_bind() and we need some storage place for it. Currently this patch prevents to use malloc() instead it stores the name in new member :
char name[LED_MAX_NAME_SIZE];
of struct led_uc_plat. While at it append led tests for the new feature.
---
as we introduce a new element in struct led_uc_priv, may we should add this functionality through a new Kconfig option? (Which also would make this optional and save some bytes for boards who do not need this...)
Azure build:
https://dev.azure.com/hs0298/hs/_build/results?buildId=165&view=results
Changes in v3:
- work in comment from Tom:
remove new file include/dt-bindings/leds/common.h in this patch as it is already in dts/upstream/include/dt-bindings/leds/common.h and so the checkpatch warnings from this file are gone! Thanks! Good catch!
remove file include/linux/uapi/linux/uleds.h and
as we only need one define from it, add this define
in include/led.h
Changes in v2:
- remove unused name variable in function led_get_label()
- rebased patch to current HEAD
82d262ae162 ("Merge patch series "MediaTek MT7629 OF_UPSTREAM migration (v2)"")
arch/sandbox/dts/test.dts | 37 ++++++++++++- doc/device-tree-bindings/leds/common.txt | 28 ++++++++++ drivers/led/led-uclass.c | 70 ++++++++++++++++++++++-- include/led.h | 6 ++ test/dm/led.c | 22 +++++++- test/dm/ofnode.c | 2 +- 6 files changed, 158 insertions(+), 7 deletions(-)
arch/sandbox/dts/test.dts +36 -1
@@ -13,6 +13,7 @@ #include <dt-bindings/gpio/gpio.h> #include <dt-bindings/gpio/sandbox-gpio.h> #include <dt-bindings/input/input.h> +#include <dt-bindings/leds/common.h> #include <dt-bindings/pinctrl/sandbox-pinmux.h> #include <dt-bindings/mux/mux.h>
@@ -820,7 +821,7 @@ gpio-controller; #gpio-cells = <1>; gpio-bank-name = "a"; - sandbox,gpio-count = <20>; + sandbox,gpio-count = <25>; hog_input_active_low { gpio-hog; input;
@@ -1010,6 +1011,40 @@ /* label intentionally omitted */ default-state = "off"; }; + + led-20 { + gpios = <&gpio_a 20 0>; + /* label intentionally omitted */ + function = LED_FUNCTION_STATUS; + color = <LED_COLOR_ID_RED>; + function-enumerator = <20>; + }; + + led-21 { + gpios = <&gpio_a 21 0>; + /* label intentionally omitted */ + function = LED_FUNCTION_STATUS; + color = <LED_COLOR_ID_GREEN>; + }; + + led-22 { + gpios = <&gpio_a 22 0>; + /* label intentionally omitted */ + function = LED_FUNCTION_STATUS; + }; + + led-23 { + gpios = <&gpio_a 23 0>; + /* label intentionally omitted */ + color = <LED_COLOR_ID_GREEN>; + }; + + led-24 { + gpios = <&gpio_a 24 0>; + label = "sandbox:function"; + function = LED_FUNCTION_STATUS; + color = <LED_COLOR_ID_GREEN>; + }; }; wdt-gpio-toggle {
doc/device-tree-bindings/leds/common.txt +28 -0
@@ -4,6 +4,21 @@ Optional properties for child nodes: - label : The label for this LED. If omitted, the label is taken from the node name (excluding the unit address). +- function: + LED function. Use one of the LED_FUNCTION_* prefixed definitions + from the header include/dt-bindings/leds/common.h. If there is no + matching LED_FUNCTION available, add a new one. + +- color: + Color of the LED. Use one of the LED_COLOR_ID_* prefixed definitions from + the header include/dt-bindings/leds/common.h. If there is no matching + LED_COLOR_ID available, add a new one. + +- function-enumerator: + optional + Integer to be used when more than one instance of the same function is + needed, differing only with an ordinal number. + - linux,default-trigger : This parameter, if present, is a string defining the trigger assigned to the LED. Current triggers are: "backlight" - LED will act as a back-light, controlled by the framebuffer
@@ -21,3 +36,16 @@ system-status { linux,default-trigger = "heartbeat"; ... }; + +led-0 { + function = LED_FUNCTION_STATUS; + color = <LED_COLOR_ID_RED>; + function-enumerator = <20>; + ... +}; + +led-1 { + function = LED_FUNCTION_STATUS; + color = <LED_COLOR_ID_YELLOW>; + ... +};
drivers/led/led-uclass.c +66 -4
104 lines not on this page · raw message
include/led.h +6 -0
@@ -53,6 +53,11 @@struct udevice ; +/* +* value imported from linux:include/linux/uapi/linux/uleds.h +*/ +#define LED_MAX_NAME_SIZE 64 +enum led_state_t {LEDST_OFF =0 ,LEDST_ON =1 ,
@@ -86,6 +91,7 @@ struct led_sw_blink {struct led_uc_plat {const char *label;enum led_state_t default_state; +char name[LED_MAX_NAME_SIZE ];#ifdef CONFIG_LED_SW_BLINK struct led_sw_blink *sw_blink;#endif
test/dm/led.c +21 -1
@@ -20,7 +20,12 @@ static int dm_test_led_base(struct unit_test_state *uts)ut_assertok (uclass_get_device (UCLASS_LED ,1 , &dev));ut_assertok (uclass_get_device (UCLASS_LED ,2 , &dev));ut_assertok (uclass_get_device (UCLASS_LED ,3 , &dev)); -ut_asserteq (-ENODEV ,uclass_get_device (UCLASS_LED ,4 , &dev)); +ut_assertok (uclass_get_device (UCLASS_LED ,4 , &dev)); +ut_assertok (uclass_get_device (UCLASS_LED ,5 , &dev)); +ut_assertok (uclass_get_device (UCLASS_LED ,6 , &dev)); +ut_assertok (uclass_get_device (UCLASS_LED ,7 , &dev)); +ut_assertok (uclass_get_device (UCLASS_LED ,8 , &dev)); +ut_asserteq (-ENODEV ,uclass_get_device (UCLASS_LED ,9 , &dev));return 0 ; }
@@ -110,6 +115,21 @@ static int dm_test_led_label(struct unit_test_state *uts)ut_asserteq (-ENODEV ,led_get_by_label ("sandbox:blue" , &dev)); +/* Test if function, color and function-enumerator naming works */ +ut_assertok (led_get_by_label ("red:status-20" , &dev)); + +/* Test if function, color naming works */ +ut_assertok (led_get_by_label ("green:status" , &dev)); + +/* Test if function, without color naming works */ +ut_assertok (led_get_by_label (":status" , &dev)); + +/* Test if color without function naming works */ +ut_assertok (led_get_by_label ("green:" , &dev)); + +/* Test if function, color naming is ignored if label is found */ +ut_assertok (led_get_by_label ("sandbox:function" , &dev)); +return 0 ; }DM_TEST (dm_test_led_label,UTF_SCAN_PDATA |UTF_SCAN_FDT );
test/dm/ofnode.c +1 -1
@@ -1580,7 +1580,7 @@ static int dm_test_ofnode_delete(struct unit_test_state *uts)ut_assert (!ofnode_valid (node));ut_assert (!ofnode_valid (ofnode_path ("/leds/default_on" ))); -ut_asserteq (2 ,ofnode_get_child_count (ofnode_path ("/leds" ))); +ut_asserteq (7 ,ofnode_get_child_count (ofnode_path ("/leds" )));return 0 ; }
Signature · 1 line
2.20.1
Quoted 15 lines · Heiko Schocher
in linux we have the option to create the name of a led optionally through the following properties:
- function
- color
- function-enumeratorThis patch adds support for parsing this properties if there is no label property.
To be as close as possible to linux import the following files from it:
include/dt-bindings/leds/common.hinclude/linux/uapi/linux/uleds.h
[snip]
Quoted 4 lines
---
as we introduce a new element in structled_uc_priv, may we should add this functionality through a newKconfigoption? (Which also would make this optional and save some bytes for boards who do not need this...)
[snip]
Quoted 7 lines
arch/sandbox/dts/test.dts | 37 ++++++++++++- doc/device-tree-bindings/leds/common.txt | 28 ++++++++++ drivers/led/led-uclass.c | 70 ++++++++++++++++++++++-- include/led.h | 6 ++ test/dm/led.c | 22 +++++++- test/dm/ofnode.c | 2 +- 6 files changed, 158 insertions(+), 7 deletions(-)
Getting things out of order slightly, sorry. The commit message should be updated a bit more still since we aren't copying files in. And we should delete doc/device-tree-bindings/leds/common.txt in a pre-patch because we have that from upstream now. As for size, how much is the growth? You can use tools/qconfig.py -b, and then -f LED to find some platforms to size compare before/after, and if you've not used buildman for size checking before I've put my script up over at
https://source.denx.de/u-boot/u-boot-extras/-/raw/master/contrib/trini/u-boot-size-test.sh?ref_type=heads
now for easier sharing.
Signature · 1 line
Tom
Hello Tom,
Quoted 31 lines · Tom Rini
Quoted 15 lines · Heiko Schocher
in linux we have the option to create the name of a led optionally through the following properties:
- function
- color
- function-enumeratorThis patch adds support for parsing this properties if there is no label property.
To be as close as possible to linux import the following files from it:
include/dt-bindings/leds/common.hinclude/linux/uapi/linux/uleds.h[snip]
Quoted 4 lines
---
as we introduce a new element in structled_uc_priv, may we should add this functionality through a newKconfigoption? (Which also would make this optional and save some bytes for boards who do not need this...)[snip]
Quoted 7 lines
arch/sandbox/dts/test.dts | 37 ++++++++++++- doc/device-tree-bindings/leds/common.txt | 28 ++++++++++ drivers/led/led-uclass.c | 70 ++++++++++++++++++++++-- include/led.h | 6 ++ test/dm/led.c | 22 +++++++- test/dm/ofnode.c | 2 +- 6 files changed, 158 insertions(+), 7 deletions(-)
Quoted 4 lines
Getting things out of order slightly, sorry. The commit message should be updated a bit more still since we aren't copying files in. And we should delete
doc/device-tree-bindings/leds/common.txtin a pre-patch because we have that from upstream now. As for size, how much is the
removed! And updated commit message
Quoted 5 lines
growth? You can use
tools/qconfig.py-b, and then -f LED to find some platforms to size compare before/after, and if you've not used buildman for size checking before I've put my script up over at
https://source.denx.de/u-boot/u-boot-extras/-/raw/master/contrib/trini/u-boot-size-test.sh?ref_type=heads
now for easier sharing.
Thanks for the script!
Here 2 examples:
$ ./u-boot-size-test.sh capricorn_cxg3 --branch led-mainline
[...]
aarch64: (for 1/1 boards) all +886.0 rodata +230.0 text +656.0
capricorn_cxg3 : all +886 rodata +230 text +656
u-boot: add: 1/0, grow: 1/0 bytes: 416/0 (416)
function old new delta
led_post_bind 220 516 +296
led_colors - 120 +120$ ./u-boot-size-test.sh aristainetos2c --branch led-mainline
[...]
01: Merge patch series "add the support of sha256_hmac and sha256_hkdf"
40: led: add function naming option from linux
arm: (for 2/2 boards) all +557.0 bss +20.0 rodata +161.0 text +376.0
aristainetos2c : all +557 bss +20 rodata +161 text +376
u-boot: add: 1/0, grow: 1/0 bytes: 264/0 (264)
function old new delta
led_post_bind 128 332 +204
led_colors - 60 +60
aristainetos2ccslb: all +557 bss +20 rodata +161 text +376
u-boot: add: 1/0, grow: 1/0 bytes: 264/0 (264)
function old new delta
led_post_bind 128 332 +204
led_colors - 60 +60So, may not worth to add a Kconfig option here... or?
bye,
Heiko
Signature · 3 lines
DENX Software Engineering GmbH, Managing Director: Erika Unter HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-52 Fax: +49-8142-66989-80 Email: [email protected]
Quoted 75 lines · Heiko Schocher
Hello Tom,
Quoted 31 lines · Tom Rini
Quoted 15 lines · Heiko Schocher
in linux we have the option to create the name of a led optionally through the following properties:
- function
- color
- function-enumeratorThis patch adds support for parsing this properties if there is no label property.
To be as close as possible to linux import the following files from it:
include/dt-bindings/leds/common.hinclude/linux/uapi/linux/uleds.h[snip]
Quoted 4 lines
---
as we introduce a new element in structled_uc_priv, may we should add this functionality through a newKconfigoption? (Which also would make this optional and save some bytes for boards who do not need this...)[snip]
Quoted 7 lines
arch/sandbox/dts/test.dts | 37 ++++++++++++- doc/device-tree-bindings/leds/common.txt | 28 ++++++++++ drivers/led/led-uclass.c | 70 ++++++++++++++++++++++-- include/led.h | 6 ++ test/dm/led.c | 22 +++++++- test/dm/ofnode.c | 2 +- 6 files changed, 158 insertions(+), 7 deletions(-)Quoted 4 lines
Getting things out of order slightly, sorry. The commit message should be updated a bit more still since we aren't copying files in. And we should delete
doc/device-tree-bindings/leds/common.txtin a pre-patch because we have that from upstream now. As for size, how much is theremoved! And updated commit message
Quoted 5 lines
growth? You can use
tools/qconfig.py-b, and then -f LED to find some platforms to size compare before/after, and if you've not used buildman for size checking before I've put my script up over at
https://source.denx.de/u-boot/u-boot-extras/-/raw/master/contrib/trini/u-boot-size-test.sh?ref_type=heads
now for easier sharing.Thanks for the script!
Here 2 examples:
$ ./u-boot-size-test.sh capricorn_cxg3 --branch led-mainline [...] aarch64: (for 1/1 boards) all +886.0 rodata +230.0 text +656.0 capricorn_cxg3 : all +886 rodata +230 text +656 u-boot: add: 1/0, grow: 1/0 bytes: 416/0 (416) function old new delta led_post_bind 220 516 +296 led_colors - 120 +120$ ./u-boot-size-test.sh aristainetos2c --branch led-mainline [...] 01: Merge patch series "add the support of sha256_hmac and sha256_hkdf" 40: led: add function naming option from linux arm: (for 2/2 boards) all +557.0 bss +20.0 rodata +161.0 text +376.0 aristainetos2c : all +557 bss +20 rodata +161 text +376 u-boot: add: 1/0, grow: 1/0 bytes: 264/0 (264) function old new delta led_post_bind 128 332 +204 led_colors - 60 +60 aristainetos2ccslb: all +557 bss +20 rodata +161 text +376 u-boot: add: 1/0, grow: 1/0 bytes: 264/0 (264) function old new delta led_post_bind 128 332 +204 led_colors - 60 +60
Quoted 1 line
So, may not worth to add a
Kconfigoption here... or?
That doesn't seem too bad for non-SPL growth. Thanks for checking.
Signature · 1 line
Tom