drivers: sensor: mcp9808: fix various problems and improve test

[pabigot]

While attempting to port to the new GPIO API I found the sample to be highly unreliable: The Adafruit breakout does not in fact have pull-ups on the alert signal, only on the I2C signals. A variety of other problems were identified, and resolving those issues resulted in a near complete rework:

• Correct handling of device encoded temperature values, which combine a 12-bit 2s complement signed value with a separate sign bit. Rework conversion between device and sensor temperature representations to support negative temperatures in both domains.
• Use a much simpler trigger configuration where the alert is driven by comparator output, rather than as an interrupt that requires a pair of I2C transactions to read and clear the flag.
• Move to storing devicetree configuration data in flash, which removes ultra-long identifiers making the code more readable.
• Refactor the trigger infrastructure to use the setup/handle/process idiom, which reduces duplicated code and to correctly detect alerts present when the triggers are set.
• Completely replace the sample with something that demonstrates updating upper and lower threshold values to track moving temperatures.

Tested with all three trigger options, and with temperatures below zero Celsius.

[zephyrbot]

All checks passed.

checkpatch (informational only, not a failure)

-:527: WARNING:LONG_LINE: line over 80 characters
#527: FILE: drivers/sensor/mcp9808/mcp9808_trigger.c:201:
+					GPIO_INT_ACTIVE_LOW | GPIO_INT_DEBOUNCE);

-:535: WARNING:LONG_LINE: line over 80 characters
#535: FILE: drivers/sensor/mcp9808/mcp9808_trigger.c:205:
+		gpio_init_callback(&data->alert_cb, alert_cb, BIT(cfg->alert_pin));

- total: 0 errors, 2 warnings, 776 lines checked

NOTE: For some of the reported defects, checkpatch may be able to
      mechanically convert to the typical style using --fix or --fix-inplace.

Your patch has style problems, please review.

NOTE: Ignored message types: AVOID_EXTERNS BRACES CONFIG_EXPERIMENTAL CONST_STRUCT DATE_TIME FILE_PATH_CHANGES MINMAX NETWORKING_BLOCK_COMMENT_STYLE PRINTK_WITHOUT_KERN_LEVEL SPLIT_STRING VOLATILE

NOTE: If any of the errors are false positives, please report
      them to the maintainers.

Tip: The bot edits this comment instead of posting a new one, so you can check the comment's history to see earlier messages.

[dbkinder]

doc changes in #21427 should win

[dbkinder]

The doc changes in #21427 are preferred.

[pabigot]

Updates made (plus-minus, explanation of interleaved output). Please continue documentation review here; #21427 will be rebased on top of this change when it goes in.

[pabigot]

@dbkinder Please revisit this, all documentation change requests here and from #21427 should be addressed. Thanks.

stale

[nashif]

nit: C instead of Cel, all other samples use C.

[pabigot]

All other samples are "wrong" because the unit is degrees Celsius, not coulombs, but I'll change this.

[nashif]

:)
some sample actually have °C, but those have some kind of display.

[MaureenHelm]

Remove this line

[MaureenHelm]

Typo?

[pabigot]

No, interleaved output from unsynchronized printf from thread and ISR.

[nashif]

tested with sensor breakout on disco_l475_iot1, works for me.

[pabigot]

Changes complete.