On exynos5 platforms the Arm Generic timer is available and is used for
MCS, but the mct.h is still used as the mct device needs to be
configured to implement the Arm Generic timer. mct.h is a common header
for register definitions while exynos4412-mct.h is used for implementing
timer functions on the exynos4 where the mct doesn't support being the
Arm generic timer.
Also remove misplaced initGenericTimer() declaration.
When SMP is enabled, the sscratch register is used to
contain the per-core kernel stack instead of the current
running thread. We get the current running thread from the
top of the kernel stack.
- Remove Remap function from seL4 API for arm, x86, riscv and the
respective invocation implementation.
- Update Map as replacement for Remap
- Update manual
This allows a change of rights if the frame being mapped is already
mapped in at the given vaddr. To map a page to a different address,
unmap it first.
Co-authored-by: Hesham Almatary <hesham.almatary@data61.csiro.au>
Co-authored-by: Anna Lyons <Anna.Lyons@data61.csiro.au>
Co-authored-by: Victor Phan <Victor.Phan@data61.csiro.au>
Co-authored-by: Kent McLeod <Kent.Mcleod@data61.csiro.au>
The ia32 kernel is supposed to pass through type 1 memory that is
outside of the kernel window as device untyped. In a previous refactor
this memory stopped getting created as untypeds. This behaviour is fixed
by pre-clamping the memory before being given to the free region list.
Prior to this commit the kernel would always trigger a full reschedule
on setPriority. This change allows the kernel to attempt a direct
switch, avoiding invoking the scheduler.
- Move location of setPriority call to after any cap deletion
This makes the proof easier as setPriority now changes
ksSchedulerAction, which is required to be ResumeCurrentThread or
ChooseNewThread by the cap delete functions.
- Update to attempt direct switch on set prio.
Without FORCE being added to `set` in declare_seL4_arch for
KernelSel4Arch, the set command doesn't seem to persist as the variable
must be already initialised in some cases. In this part of the script
processing, a platform has already been selected and is setting its seL4
arch and so using FORCE seems appropriate anyway.
This register is visible to software executing at EL0 but not writeable.
Storing it in the VCPU context required custom save/restore handling as
it had to be explicitly handled when switching from a VCPU thread to a
non-VCPU thread so that it didn't become a channel. It is possible to
now update this register via seL4_TCB_WriteRegisters for software
executing at EL0.
This also fixes a potential bug where if a vcpu-thread is switched for a
non-vcpu-thread and then switched to a different vcpu-thread the
original vcpu-thread's copy of this register will get set to 0.
Reserved IRQ reporting would print on every received reserved IRQ. This
feature is supposed to report spurious interrupts, yet most cases
reserved IRQs are not spurious. We change the printing to only print if
the reserved IRQ was unhandled. The intention is to allow more app
configurations to leave this feature enabled in development builds.
These configs are also used by x86 files and this fixes an configuration
race condition where projects try set KernelIOMMU after platform
settings are configured and so KernelIOMMU doesn't result in a correct
value.
