Heap and main thread stack checking #2367

c1728p9 · 2016-08-04T00:45:51Z

Turn on main thread stack checking. This PR is on top of #2319

c1728p9 · 2016-08-04T00:46:12Z

/morph test

c1728p9 · 2016-08-04T00:46:46Z

@mbed-bot: TEST

HOST_OSES=windows
BUILD_TOOLCHAINS=GCC_ARM,ARM,IAR
TARGETS=K64F,NRF51_DK,NRF51_MICROBIT,NUCLEO_F411RE

mbed-bot · 2016-08-04T01:15:09Z

[Build 769]
SUCCESS: Building succeeded and tests were run! Be sure to check the test results

c1728p9 · 2016-08-04T01:59:53Z

/morph test

mbed-bot · 2016-08-04T02:24:28Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 599

Test failed!

mbed-bot · 2016-08-04T04:10:23Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 601

Test failed!

bridadan · 2016-08-04T14:50:31Z

@c1728p9 Test looks ok, false failure 👍

0xc0170 · 2016-08-08T12:25:23Z

Commit 4055d08 looks good. How did you test this?

c1728p9 · 2016-08-08T21:21:40Z

/morph test

c1728p9 · 2016-08-08T21:22:05Z

@mbed-bot: TEST

HOST_OSES=windows
BUILD_TOOLCHAINS=GCC_ARM,ARM,IAR
TARGETS=K64F,NRF51_DK,NRF51_MICROBIT,NUCLEO_F411RE

mbed-bot · 2016-08-08T21:50:55Z

[Build 783]
SUCCESS: Building succeeded and tests were run! Be sure to check the test results

mbed-bot · 2016-08-08T21:59:27Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 614

Build failed!

c1728p9 · 2016-08-08T22:12:30Z

/morph test

c1728p9 · 2016-08-08T22:14:24Z

@mbed-bot: TEST

HOST_OSES=windows
BUILD_TOOLCHAINS=GCC_ARM,ARM,IAR
TARGETS=K64F,NRF51_DK,NRF51_MICROBIT,NUCLEO_F411RE

geky · 2016-08-08T23:18:34Z

hal/common/retarget.cpp

@@ -714,3 +726,34 @@ extern "C" void __env_unlock( struct _reent *_r )
 #endif

 } // namespace mbed
+
+void* operator new(std::size_t count)


nitpick:

~~void* operator new(std::size_t count)~~
void *operator new(std::size_t count)

~~void* operator new[](std::size_t count)~~
void *operator new[](std::size_t count)

~~void operator delete(void* ptr)~~
void operator delete(void *ptr)

~~void operator delete[](void* ptr)~~
void operator delete[](void *ptr)

mbed-bot · 2016-08-08T23:19:00Z

[Build 785]
SUCCESS: Building succeeded and tests were run! Be sure to check the test results

mbed-bot · 2016-08-09T00:07:24Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 616

Test failed!

c1728p9 · 2016-08-09T03:07:51Z

/morph test

c1728p9 · 2016-08-09T03:08:05Z

@mbed-bot: TEST

HOST_OSES=windows
BUILD_TOOLCHAINS=GCC_ARM,ARM,IAR
TARGETS=K64F,NRF51_DK,NRF51_MICROBIT,NUCLEO_F411RE

mbed-bot · 2016-08-09T04:28:22Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 617

Test failed!

mbed-bot · 2016-08-09T05:28:33Z

[Build 787]
SUCCESS: Building succeeded and tests were run! Be sure to check the test results

0xc0170 · 2016-08-09T16:15:39Z

Please refrain from adding non-relevant commits (targets changes), that could be a separate patch. Can you please fix formatting of * (void * pointer -> void *pointer).

0xc0170 · 2016-08-09T16:26:04Z

rtos/rtx/TARGET_CORTEX_M/TESTS/memory/heap_and_stack/main.cpp

@@ -0,0 +1,218 @@
+#include <stdio.h>


this file should contain license on the top, please add it

0xc0170 · 2016-08-09T16:30:28Z

+1 for consolidating stack and heap, and enabling checking for the main thread.

Tests fail for cfstore (2 errors there, for 2 targets and only GCC_ARM), I dont think its related to this changeset. @simonqhughes Any pointers?
Edit: looks like it is, Russ please provide some info what needs to be done

bridadan · 2016-08-09T16:31:15Z

Please refrain from adding non-relevant commits (targets changes), that could be a separate patch.

@0xc0170 Pretty sure he made that target patch so the test would pass (it previously failed due to the other changes in this PR). So I think that's justified.

Set well defined limits for the heap and configure GCC and ARMCC to correctly check these. IAR already correctly checked its heap. This also statically declares the main thread stack in IAR so the linker is responsible for its placement. This must be statically declared since the majority of IAR .icf files make no guarantees where free space is located. For GCC_ARM and ARM the main thread stack is placed inside the heap and the heap size is adjusted to account for this.

Decrease the main stack size for the nrf51 so there is more heap space available.

When new or new[] fails to allocate space trigger an error.

With the latest K64F linker file the initial stack is out of sync with INITIAL_SP when uVisor is not present. This patch removes the incorrect declaration.

Since the heap and stack are no longer shared, stack checking on the main thread can be turned back on. This allows stack overflows on the main thread to be caught quickly.

Test the following components of the heap and stacks: -Heap, Interrupt stack and thread stack are at the expected locations -Entire range of main stack can be used without corrupting memory -Entire heap can be used -Heap limit is properly enforced and returns NULL when out of of memory

Decrease the interrupt stack of the NUC472 from 12KB to 8KB. This frees up enough space to allow the core tests to run.

Set the initial stack pointer to the end of ram so it matches INITIAL_SP defined in RTX_CM_lib.h.

The config store tests use more than 2K of the interrupt stack, causing an overflow. This patch bumps the K64F interrupt stack size to 4K for ARMCC and GCC. The IAR interrupt stack is left untouched since it is 32K.

c1728p9 · 2016-08-09T19:40:32Z

/morph test

c1728p9 · 2016-08-09T19:40:42Z

@mbed-bot: TEST

HOST_OSES=windows
BUILD_TOOLCHAINS=GCC_ARM,ARM,IAR
TARGETS=K64F,NRF51_DK,NRF51_MICROBIT,NUCLEO_F411RE

mbed-bot · 2016-08-09T20:10:33Z

[Build 789]
SUCCESS: Building succeeded and tests were run! Be sure to check the test results

mbed-bot · 2016-08-09T20:59:22Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 622

Test Cleanup failed!

bridadan · 2016-08-09T21:15:34Z

Interesting error (Test Cleanup), looking into it now. But tests all passed 👍

c1728p9 · 2016-08-09T21:16:46Z

@mbed-bot: TEST

HOST_OSES=windows
BUILD_TOOLCHAINS=GCC_ARM,ARM,IAR
TARGETS=K64F,NRF51_DK,NRF51_MICROBIT,NUCLEO_F411RE

mbed-bot · 2016-08-09T21:45:21Z

[Build 791]
SUCCESS: Building succeeded and tests were run! Be sure to check the test results

c1728p9 mentioned this pull request Aug 8, 2016

Heap checking #2319

Closed

c1728p9 force-pushed the main_thread_stack_checking branch from 4055d08 to 09c98a3 Compare August 8, 2016 21:21

c1728p9 changed the title ~~Main thread stack checking~~ Heap and main thread stack checking Aug 8, 2016

geky reviewed Aug 8, 2016
View reviewed changes

c1728p9 force-pushed the main_thread_stack_checking branch from ff8f789 to 66bc6a2 Compare August 9, 2016 03:07

0xc0170 reviewed Aug 9, 2016
View reviewed changes

c1728p9 and others added 8 commits August 9, 2016 14:36

Adjust stack size for the nrf51

81515c9

Decrease the main stack size for the nrf51 so there is more heap space available.

Override new and delete operators to trap errors

8d4b981

When new or new[] fails to allocate space trigger an error.

Correctly set INITIAL_SP for K64F

8fa5a7f

With the latest K64F linker file the initial stack is out of sync with INITIAL_SP when uVisor is not present. This patch removes the incorrect declaration.

Turn on main thread stack checking

3fa60cb

Since the heap and stack are no longer shared, stack checking on the main thread can be turned back on. This allows stack overflows on the main thread to be caught quickly.

Decrease NUC472 interrupt stack size

f08262a

Decrease the interrupt stack of the NUC472 from 12KB to 8KB. This frees up enough space to allow the core tests to run.

Fix nrf52 initial stack pointer in ARMCC

031b2eb

Set the initial stack pointer to the end of ram so it matches INITIAL_SP defined in RTX_CM_lib.h.

Increase interrupt stack size for K64F

1d2d9bd

The config store tests use more than 2K of the interrupt stack, causing an overflow. This patch bumps the K64F interrupt stack size to 4K for ARMCC and GCC. The IAR interrupt stack is left untouched since it is 32K.

c1728p9 force-pushed the main_thread_stack_checking branch from 66bc6a2 to 1d2d9bd Compare August 9, 2016 19:40

c1728p9 mentioned this pull request Aug 9, 2016

Main thread stack checking alt impl #2402

Merged

c1728p9 closed this Aug 11, 2016

Heap and main thread stack checking #2367

Heap and main thread stack checking #2367

Uh oh!

Conversation

c1728p9 commented Aug 4, 2016

Uh oh!

c1728p9 commented Aug 4, 2016

Uh oh!

c1728p9 commented Aug 4, 2016

Uh oh!

mbed-bot commented Aug 4, 2016

Uh oh!

c1728p9 commented Aug 4, 2016

Uh oh!

mbed-bot commented Aug 4, 2016

Result: FAILURE

Output

Uh oh!

mbed-bot commented Aug 4, 2016

Result: FAILURE

Output

Uh oh!

bridadan commented Aug 4, 2016

Uh oh!

0xc0170 commented Aug 8, 2016

Uh oh!

c1728p9 commented Aug 8, 2016

Uh oh!

c1728p9 commented Aug 8, 2016

Uh oh!

mbed-bot commented Aug 8, 2016

Uh oh!

mbed-bot commented Aug 8, 2016

Result: FAILURE

Output

Uh oh!

c1728p9 commented Aug 8, 2016

Uh oh!

c1728p9 commented Aug 8, 2016

Uh oh!

geky Aug 8, 2016

Choose a reason for hiding this comment

Uh oh!

mbed-bot commented Aug 8, 2016

Uh oh!

mbed-bot commented Aug 9, 2016

Result: FAILURE

Output

Uh oh!

c1728p9 commented Aug 9, 2016

Uh oh!

c1728p9 commented Aug 9, 2016

Uh oh!

mbed-bot commented Aug 9, 2016

Result: FAILURE

Output

Uh oh!

mbed-bot commented Aug 9, 2016

Uh oh!

0xc0170 commented Aug 9, 2016 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

0xc0170 Aug 9, 2016

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Aug 9, 2016 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bridadan commented Aug 9, 2016 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

c1728p9 commented Aug 9, 2016

Uh oh!

c1728p9 commented Aug 9, 2016

Uh oh!

mbed-bot commented Aug 9, 2016

Uh oh!

mbed-bot commented Aug 9, 2016

Result: FAILURE

Output

0xc0170 commented Aug 9, 2016 •

edited

Loading

0xc0170 commented Aug 9, 2016 •

edited

Loading

bridadan commented Aug 9, 2016 •

edited

Loading