std::make_unique for arrays is equivalent to doing:
std::unique_ptr<T>(new typename std::remove_extent<T>::type[size]())
(note the ending () after the array size specifier). This means that the
default value within memory for the constructed types will be whatever
the default constructor for that type does. Given the built-in
type for std::uint8_t doesn't have a constructor, this is equivalent to
forcing zero-initialization, so the memory will already be zeroed out on
construction. Because of that, there's no need to zero it out again.
All usage of GetPointerFromVMA is to recover the pointer that is nulled by changing page type to RasterizerCachedMemory. Our rasterizer cache only works on linear heap and vram, so we can recover the pointer directly by address computation, instead of going through VMA table. Also removed a sanity check pointer!=nullptr in RasterizerMarkRegionCached(RasterizerCachedMemory=>Memory), as now the pointer is never null. The sanity check was added in f2a5a77 (#2797), which was originally necessary during VMA unmapping process, because the function is invloked by VMA after unmapping the page, which in turn invokes back to query the memory, forming a circular dependency. Now the dependency is resolved so the check is not necessary
There is no external use of PhysicalToVirtualAddress any more, so it there is no need to have a generic function that handles all physical regions. Also, the previous APT change makes it possible that linear heap has some regions mapped to old and new VAddr regions at the same time, so we need to check both region and mark cached for the mapped one. RasterizerMarkRegionCached would skip the unmapped one in its loop
* Kernel: reimplement memory management on physical FCRAM
* Kernel/Process: Unmap does not care the source memory permission
What game usually does is after mapping the memory, they reprotect the source memory as no permission to avoid modification there
* Kernel/SharedMemory: zero initialize new-allocated memory
* Process/Thread: zero new TLS entry
* Kernel: fix a bug where code segments memory usage are accumulated twice
It is added to both misc and heap (done inside HeapAlloc), which results a doubled number reported by svcGetProcessInfo. While we are on it, we just merge the three number misc, heap and linear heap usage together, as there is no where they are distinguished.
Question: is TLS page also added to this number?
* Kernel/SharedMemory: add more object info on mapping error
* Process: lower log level; SharedMemory: store phys offset
* VMManager: add helper function to retrieve backing block list for a range
Two functional change:
QueryProcessMemory uses the process passed from handle instead current_process
Thread::Stop() uses TLS from owner_process instead of current_process
