Skip to content

cpu/cortexm: disable interrupts before sleeping#6499

Merged
PeterKietzmann merged 1 commit intoRIOT-OS:masterfrom
haukepetersen:fix_stm32l1_hardfaultonisr
Jan 27, 2017
Merged

cpu/cortexm: disable interrupts before sleeping#6499
PeterKietzmann merged 1 commit intoRIOT-OS:masterfrom
haukepetersen:fix_stm32l1_hardfaultonisr

Conversation

@haukepetersen
Copy link
Copy Markdown
Contributor

@haukepetersen haukepetersen commented Jan 27, 2017

fixes #6484 (though temporarily for now)

A real fix will be provided with #6239

@haukepetersen haukepetersen added the Type: bug The issue reports a bug / The PR fixes a bug (including spelling errors) label Jan 27, 2017
@haukepetersen haukepetersen added this to the Release 2017.01 milestone Jan 27, 2017
@haukepetersen haukepetersen mentioned this pull request Jan 27, 2017
Closed
@haukepetersen haukepetersen force-pushed the fix_stm32l1_hardfaultonisr branch from 0a0ba13 to 6e364b9 Compare January 27, 2017 11:51
@haukepetersen haukepetersen force-pushed the fix_stm32l1_hardfaultonisr branch from 6e364b9 to 3da9ceb Compare January 27, 2017 12:03
@haukepetersen haukepetersen changed the title cpu/stm32: disable sleep-mode for STM32L1 cpu/cortexm: disable interrupts before sleeping() Jan 27, 2017
@haukepetersen haukepetersen changed the title cpu/cortexm: disable interrupts before sleeping() cpu/cortexm: disable interrupts before sleeping Jan 27, 2017
@haukepetersen
Copy link
Copy Markdown
Contributor Author

haukepetersen commented Jan 27, 2017

Acutally, I changed the fix: The ARM Cortex-M reference manual advises to disable global interrupt before going to sleep. This prevents interrupts that might be triggered right between the execution of the __DSB and the __WFI instruction and would be subsequently lost.

Further there seems to be some hardware issue on some Cortex MCUs (e.g. on the stm32l1), that crash when waking up from __WFI with interrupts enabled...

@PeterKietzmann PeterKietzmann added Process: needs backport Integration Process: The PR is required to be backported to a release or feature branch CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR labels Jan 27, 2017
PeterKietzmann
PeterKietzmann approved these changes Jan 27, 2017
View reviewed changes
Copy link
Copy Markdown
Member

@PeterKietzmann PeterKietzmann left a comment

Choose a reason for hiding this comment

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

@haukepetersen great! Please provide a backport PR.

This was referenced Jan 27, 2017
Closed
Closed
@PeterKietzmann PeterKietzmann merged commit 40a6664 into RIOT-OS:master Jan 27, 2017
@PeterKietzmann PeterKietzmann mentioned this pull request Jan 27, 2017
Merged
@PeterKietzmann
Copy link
Copy Markdown
Member

PeterKietzmann commented Jan 27, 2017

Did the backport for you in #6505

@haukepetersen haukepetersen deleted the fix_stm32l1_hardfaultonisr branch January 30, 2017 08:48
@immesys
Copy link
Copy Markdown
Contributor

immesys commented Feb 22, 2017

@haukepetersen, who did this fix, and @kaspar030 who did the new samd21/periph/pm.c. Is this a change that should be made universally? There are 4 places in the code where cortex chips execute __DSB(); __WFI(); and this is only one of them. The other three are in various periph/pm.c files. Should they be changed too?

@haukepetersen
Copy link
Copy Markdown
Contributor Author

haukepetersen commented Mar 28, 2017

@immesys: yes, it should be generalized -> see #6812

@miri64
Copy link
Copy Markdown
Member

miri64 commented Oct 18, 2017

There is a new release pending so at this point the backporting-needed label is only noise ;-)

@miri64 miri64 removed the Process: needs backport Integration Process: The PR is required to be backported to a release or feature branch label Oct 18, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@PeterKietzmann PeterKietzmann PeterKietzmann approved these changes

Assignees

@PeterKietzmann PeterKietzmann

Labels

CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR Type: bug The issue reports a bug / The PR fixes a bug (including spelling errors)

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

boards/nucleo-l1: hard faults with shell?

4 participants

@haukepetersen @PeterKietzmann @immesys @miri64