make: enable selection of periph drivers via USEMODULE by haukepetersen · Pull Request #3420 · RIOT-OS/RIOT

haukepetersen · 2015-07-15T17:55:19Z

So far, we are compiling all available peripheral drivers for a CPU, independent if they are really used or not. This PR enables us to treat peripheral drivers just as ordinary modules (e.g. select them via USEMODULE += periph_gpio, etc.).

I re-ordered some includes in the main Makefile.include, would be nice if somebody could re-check this change! @gebart, @Kijewski maybe?

This PR brings some advantages:

we save compile time
we can check automatically during compile time, if a board/cpu has all the peripheral drivers needed -> no need to explicitly state the required features in an application anymore...
device drivers now state explicitly, what kind of peripherals they need. This could be used for automatic extraction of a board-to-driver compatibility matrix or the like

Outlook: Next step is to look at the test application and scripts and remove the explicit FEATURES_REQUIRED definitions from their Makefiles...

jnohlgard · 2015-07-15T18:15:06Z

drivers/Makefile.dep

don't the cc*** use any periph stuff?

nope, these are old (deprecated) drivers. The SPI/GPIO stuff of this drivers is implemented in a board specific part of the driver for some of the boards... These drivers are anyway subject to cleanup soon and will be removed.

kick the line out anyways, there's no cc110x in master anymore. :)

jnohlgard · 2015-07-15T18:19:56Z

I think Mulle would need USEMODULE += periph_gpio because it controls on-board LEDs and power switches using the periph_gpio driver.
Edit: And periph_rtt for the clock

haukepetersen · 2015-07-15T21:14:54Z

The thing is, that all Freescale CPUs do still build all the peripherals anyway, as the structure for these boards differs from the structure of the other CPUs. I would actually touch them in a follow-up PR...

haukepetersen · 2015-07-31T19:03:45Z

rebased.

PeterKietzmann · 2015-08-03T12:26:03Z

@kaspar030 did you already start your review? For #3538 it would be cool to have this one merged soon. At the first glance the PR looks ok to me, but I didn't read the changes that detailed.

haukepetersen · 2015-08-12T08:57:42Z

@kaspar030 @PeterKietzmann could you review this? Would be nice to have it off the table.

Also rebased.

kaspar030 · 2015-08-12T09:20:29Z

Looking at it now, is there a need for all "periph" *_MODULEs to also be _FEATURES?

haukepetersen · 2015-08-12T14:12:14Z

Don't know if I understand the questions correctly: do you mean if we still need the features_provided declarations? Then the answer is yes.

kaspar030 · 2015-08-14T08:50:45Z

boards/saml21-xpro/Makefile.features

timer in there two times

haukepetersen · 2015-08-17T09:21:35Z

hm, the new periph_common clashes with this PR, as this PR does automatic parsing of everything that starts with periph_...

kaspar030 · 2015-08-17T11:02:34Z

Do you want to rename that to "common_periph"?

haukepetersen · 2015-08-17T11:10:03Z

just did.

haukepetersen · 2015-08-17T11:10:37Z

Can you look over the changes once more? I will squash once you give me your ok.

kaspar030 · 2015-08-17T11:13:42Z

drivers/Makefile.dep

haukepetersen · 2015-08-25T18:32:05Z

fixed obvious errors and look into travis bugs now.

haukepetersen · 2015-08-25T18:47:12Z

Travis things seemed unrelated to me, let's see what the current run tells us.

PeterKietzmann · 2015-08-25T20:05:36Z

drivers/Makefile.dep

This driver is based on i2c. Don't you need to add the USEMODULE += periph_i2c line here?

PeterKietzmann · 2015-08-25T20:47:06Z

I don't know where this comes from, but some applications don't compile because it seems that a USEMODULE is missing. E.g. now you can't compile periph_adc for stm boards. Adding USEMODULE += periph_adc to the tests Makefile, it works

PeterKietzmann · 2015-08-25T20:50:46Z

Travis things seemed unrelated to me, let's see what the current run tells us

@haukepetersen am I mistaken somewhere? For me, the build fails really don't seem unrelated to the PR. Please adapt some of my proposed changes. I hope the number of fails will decrease and then I'll have a look at your PR again. Ok?

haukepetersen · 2015-09-02T09:56:46Z

I think you are right, I mussed have messed up something in the build system... I look into it as soon as I find the time...

haukepetersen · 2015-11-24T17:01:40Z

remove the hack'n'ack label, as this PR still needs work

haukepetersen · 2016-01-26T19:16:25Z

ok won't get this working anytime soon (or ever?), so I will close this PR for now and open a new one whenever (if ever) I make progress...

haukepetersen added Type: enhancement The issue suggests enhanceable parts / The PR enhances parts of the codebase / documentation Area: build system Area: Build system Area: drivers Area: Device drivers labels Jul 15, 2015

haukepetersen assigned kaspar030 Jul 15, 2015

haukepetersen changed the title ~~make: eneable selection of periph drivers via USEMODULE~~ make: enable selection of periph drivers via USEMODULE Jul 15, 2015

jnohlgard reviewed Jul 15, 2015
View reviewed changes

haukepetersen mentioned this pull request Jul 16, 2015

auto_init: add initialization for random module #3425

Closed

PeterKietzmann mentioned this pull request Jul 17, 2015

drivers/hih6310: Initial commit of HIH6310 driver #3300

Merged

haukepetersen force-pushed the add_periph_usemodule branch from bd77b3d to e16e501 Compare July 31, 2015 19:03

haukepetersen mentioned this pull request Jul 31, 2015

Introducing SAUL, the [S]ensor [A]ctuator [U]ber [L]ayer #3538

Merged

haukepetersen force-pushed the add_periph_usemodule branch from e16e501 to 5a5a210 Compare August 12, 2015 08:57

kaspar030 added the CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR label Aug 14, 2015

kaspar030 reviewed Aug 14, 2015
View reviewed changes

boards/saml21-xpro/Makefile.features Outdated

Copy link
Copy Markdown

Contributor

kaspar030 Aug 14, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

timer in there two times

haukepetersen force-pushed the add_periph_usemodule branch from 5a5a210 to 3fb5436 Compare August 17, 2015 11:09

kaspar030 reviewed Aug 17, 2015
View reviewed changes

drivers/Makefile.dep

Copy link
Copy Markdown

Contributor

kaspar030 Aug 17, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I like!

haukepetersen added 2 commits August 25, 2015 20:31

s make: fixed Makefile dep

7504a0b

s zep: added missing semisolon

eeacfc9

haukepetersen force-pushed the add_periph_usemodule branch from 3b8cb23 to eeacfc9 Compare August 25, 2015 18:31

PeterKietzmann reviewed Aug 25, 2015
View reviewed changes

drivers/Makefile.dep

Copy link
Copy Markdown

Member

PeterKietzmann Aug 25, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This driver is based on i2c. Don't you need to add the USEMODULE += periph_i2c line here?

OlegHahm modified the milestone: Release NEXT MAJOR Sep 2, 2015

This was referenced Oct 21, 2015

make: moved driver deps to drivers/Makefile.dep #4127

Merged

boards: cleaned up entries in Makefile.features #4128

Merged

drivers: renamed periph_common #4129

Closed

haukepetersen removed the Community: Hack'n'ACK candidate This PR is a candidate for review and discussion during one of RIOT's monthly Hack'n'ACK parties label Nov 24, 2015

haukepetersen mentioned this pull request Nov 27, 2015

drivers: Optimized periph PWM interfaces #3644

Closed

OlegHahm modified the milestones: Release 2015.12, Release 2016.03 Dec 3, 2015

haukepetersen mentioned this pull request Dec 14, 2015

drivers/periph: updated DAC driver interface #4431

Merged

haukepetersen closed this Jan 26, 2016

haukepetersen mentioned this pull request Mar 14, 2016

cpu/make: added global Makefile.periph #5065

Closed

jnohlgard mentioned this pull request Jul 7, 2016

cpu: Fix array-bounds errors with GCC 6 #5602

Merged

haukepetersen deleted the add_periph_usemodule branch March 8, 2017 10:32

Conversation

haukepetersen commented Jul 15, 2015

Uh oh!

jnohlgard Jul 15, 2015

Choose a reason for hiding this comment

Uh oh!

haukepetersen Jul 15, 2015

Choose a reason for hiding this comment

Uh oh!

kaspar030 Aug 12, 2015

Choose a reason for hiding this comment

Uh oh!

jnohlgard commented Jul 15, 2015

Uh oh!

haukepetersen commented Jul 15, 2015

Uh oh!

haukepetersen commented Jul 31, 2015

Uh oh!

PeterKietzmann commented Aug 3, 2015

Uh oh!

haukepetersen commented Aug 12, 2015

Uh oh!

kaspar030 commented Aug 12, 2015

Uh oh!

haukepetersen commented Aug 12, 2015

Uh oh!

kaspar030 Aug 14, 2015

Choose a reason for hiding this comment

Uh oh!

haukepetersen commented Aug 17, 2015

Uh oh!

kaspar030 commented Aug 17, 2015

Uh oh!

haukepetersen commented Aug 17, 2015

Uh oh!

haukepetersen commented Aug 17, 2015

Uh oh!

kaspar030 Aug 17, 2015

Choose a reason for hiding this comment

Uh oh!

haukepetersen commented Aug 25, 2015

Uh oh!

haukepetersen commented Aug 25, 2015

Uh oh!

PeterKietzmann Aug 25, 2015

Choose a reason for hiding this comment

Uh oh!

PeterKietzmann commented Aug 25, 2015

Uh oh!

PeterKietzmann commented Aug 25, 2015

Uh oh!

haukepetersen commented Sep 2, 2015

Uh oh!

haukepetersen commented Nov 24, 2015

Uh oh!

haukepetersen commented Jan 26, 2016

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants