Re: sptimecounter01 use of stack allocator/deallocator

2017-11-21 Thread Sebastian Huber
On 21/11/17 16:23, Joel Sherrill wrote: On Tue, Nov 21, 2017 at 1:12 AM, Sebastian Huber > wrote: On 20/11/17 17:53, Joel Sherrill wrote: The purpose of defining these to NULL isn't documented in this test and renders this test

Re: sptimecounter01 use of stack allocator/deallocator

2017-11-21 Thread Joel Sherrill
On Tue, Nov 21, 2017 at 1:12 AM, Sebastian Huber < sebastian.hu...@embedded-brains.de> wrote: > On 20/11/17 17:53, Joel Sherrill wrote: > >> >> The purpose of defining these to NULL isn't documented in this test and >> renders this test broken for what seems to be a reason orthogonal to the >> tes

Re: sptimecounter01 use of stack allocator/deallocator

2017-11-20 Thread Sebastian Huber
On 20/11/17 17:53, Joel Sherrill wrote: The purpose of defining these to NULL isn't documented in this test and renders this test broken for what seems to be a reason orthogonal to the test's apparent purpose. This test runs in a minimal environment (test runs in boot_card()) to verify the

sptimecounter01 use of stack allocator/deallocator

2017-11-20 Thread Joel Sherrill
Hi Why does sptimecounter01 define CONFIGURE_TASK_STACK_ALLOCATOR and CONFIGURE_TASK_STACK_DEALLOCATOR to NULL? I ask because I have an out of tree paravirtualized BSP that has to provide BSP specific stack management to account for the memory protection it operates within. Not using those alloca