fix: rework TLS, fix alignments

[mkroening]

This PR completely reworks and unifies our thread-local storage implementation:

1. A memory leak on ARM was fixed.
2. We now test TLS on all architectures (was mistakenly disabled in 711a85c).
3. We now have only one unified Tls struct instead of three architecture-dependent TaskTLS structs. This should make future changes much easier.
4. On x86-64 and RISC-V, we now have only one layer of allocation indirections for managing the structure. This was already the case on ARM.
5. We now allocate and free the TLS structure with the correct alignment. This was done wrong on all architectures in different ways before. Closes Large TLS alignments don't work #1677.
6. TLS allocations are now managed through a separate Allocation abstraction.
7. We no longer allocate an (unused) dtv structure on ARM (not done on the other platforms anyway). We can still add proper dtv support once we have the need for multiple TLS data blocks per thread in the future.
8. We can now add proper Hermit-specific TLS data for in-kernel use that is accessed via the thread pointer instead of going through several indirections via the core pointer as we do now. This will be done in a follow-up PR.
9. The code should now be properly documented and cited.

[github-actions]

Benchmark Results

Benchmark	Current: 87ffd83	Previous: 376b966	Performance Ratio
startup_benchmark Build Time	113.87 s	110.37 s	1.03
startup_benchmark File Size	0.91 MB	0.90 MB	1.00
Startup Time - 1 core	0.91 s (±0.03 s)	0.93 s (±0.03 s)	0.99
Startup Time - 2 cores	0.93 s (±0.03 s)	0.93 s (±0.03 s)	0.99
Startup Time - 4 cores	0.91 s (±0.04 s)	0.93 s (±0.03 s)	0.98
multithreaded_benchmark Build Time	113.13 s	114.11 s	0.99
multithreaded_benchmark File Size	1.01 MB	1.01 MB	1.00
Multithreaded Pi Efficiency - 2 Threads	87.82 % (±9.31 %)	87.54 % (±7.72 %)	1.00
Multithreaded Pi Efficiency - 4 Threads	42.65 % (±3.09 %)	43.85 % (±3.71 %)	0.97
Multithreaded Pi Efficiency - 8 Threads	24.69 % (±1.84 %)	25.11 % (±2.07 %)	0.98
micro_benchmarks Build Time	106.88 s	106.92 s	1.00
micro_benchmarks File Size	1.01 MB	1.01 MB	1.00
Scheduling time - 1 thread	63.81 ticks (±4.16 ticks)	67.03 ticks (±3.54 ticks)	0.95
Scheduling time - 2 threads	35.64 ticks (±4.26 ticks)	38.31 ticks (±5.49 ticks)	0.93
Micro - Time for syscall (getpid)	3.33 ticks (±0.51 ticks)	2.98 ticks (±0.45 ticks)	1.12
Memcpy speed - (built_in) block size 4096	79404.09 MByte/s (±54932.31 MByte/s)	78287.67 MByte/s (±54207.29 MByte/s)	1.01
Memcpy speed - (built_in) block size 1048576	42840.65 MByte/s (±29650.14 MByte/s)	42758.95 MByte/s (±29603.37 MByte/s)	1.00
Memcpy speed - (built_in) block size 16777216	28783.83 MByte/s (±23658.97 MByte/s)	28796.76 MByte/s (±23654.74 MByte/s)	1.00
Memset speed - (built_in) block size 4096	79293.26 MByte/s (±54844.14 MByte/s)	78337.85 MByte/s (±54246.52 MByte/s)	1.01
Memset speed - (built_in) block size 1048576	43067.81 MByte/s (±29806.83 MByte/s)	43001.55 MByte/s (±29767.07 MByte/s)	1.00
Memset speed - (built_in) block size 16777216	29535.47 MByte/s (±24094.08 MByte/s)	29553.13 MByte/s (±24095.41 MByte/s)	1.00
Memcpy speed - (rust) block size 4096	70444.41 MByte/s (±49435.58 MByte/s)	70495.32 MByte/s (±49405.95 MByte/s)	1.00
Memcpy speed - (rust) block size 1048576	42628.03 MByte/s (±29535.56 MByte/s)	42856.55 MByte/s (±29682.27 MByte/s)	0.99
Memcpy speed - (rust) block size 16777216	29035.32 MByte/s (±23864.77 MByte/s)	28900.71 MByte/s (±23739.86 MByte/s)	1.00
Memset speed - (rust) block size 4096	71078.44 MByte/s (±49887.51 MByte/s)	69943.60 MByte/s (±49128.48 MByte/s)	1.02
Memset speed - (rust) block size 1048576	42890.02 MByte/s (±29712.30 MByte/s)	43095.53 MByte/s (±29843.91 MByte/s)	1.00
Memset speed - (rust) block size 16777216	29792.76 MByte/s (±24301.19 MByte/s)	29647.15 MByte/s (±24171.90 MByte/s)	1.00
alloc_benchmarks Build Time	104.78 s	105.40 s	0.99
alloc_benchmarks File Size	0.97 MB	0.97 MB	1.00
Allocations - Allocation success	100.00 %	100.00 %	1
Allocations - Deallocation success	69.98 % (±0.28 %)	70.03 % (±0.30 %)	1.00
Allocations - Pre-fail Allocations	100.00 %	100.00 %	1
Allocations - Average Allocation time	12284.97 Ticks (±215.57 Ticks)	12298.03 Ticks (±230.53 Ticks)	1.00
Allocations - Average Allocation time (no fail)	12284.97 Ticks (±215.57 Ticks)	12298.03 Ticks (±230.53 Ticks)	1.00
Allocations - Average Deallocation time	725.45 Ticks (±92.80 Ticks)	722.17 Ticks (±103.75 Ticks)	1.00
mutex_benchmark Build Time	104.50 s	105.06 s	0.99
mutex_benchmark File Size	1.01 MB	1.02 MB	1.00
Mutex Stress Test Average Time per Iteration - 1 Threads	12.54 ns (±0.88 ns)	12.74 ns (±0.74 ns)	0.98
Mutex Stress Test Average Time per Iteration - 2 Threads	13.98 ns (±1.26 ns)	14.80 ns (±1.47 ns)	0.94

This comment was automatically generated by workflow using github-action-benchmark.