Kernel/Threads: Dynamically allocate the TLS region for threads in the BASE region of the linear heap.

Each thread gets a 0x200-byte area from the 0x1000-sized page, when all 8 thread slots in a single page are used up, the kernel allocates a new page to hold another 8 entries.

This is consistent with what the real kernel does.
This commit is contained in:
Subv 2016-04-19 17:12:48 -05:00
parent 3e7e8daf59
commit d192fb066d
5 changed files with 74 additions and 28 deletions

View file

@ -140,8 +140,11 @@ public:
MemoryRegionInfo* memory_region = nullptr;
/// Bitmask of the used TLS slots
std::bitset<300> used_tls_slots;
/// The Thread Local Storage area is allocated as processes create threads,
/// each TLS area is 0x200 bytes, so one page (0x1000) is split up in 8 parts, and each part
/// holds the TLS for a specific thread. This vector contains which parts are in use for each page as a bitmask.
/// This vector will grow as more pages are allocated for new threads.
std::vector<std::bitset<8>> tls_slots;
VAddr GetLinearHeapAreaAddress() const;
VAddr GetLinearHeapBase() const;