drivers: Modified code in virtio transport

[udaykiriti]

I've made some small, improvements to the virtio drivers. :| This commit optimizes queue selection for both MMIO and PCI to skip redundant register writes.

[github-actions]

Benchmark Results

Details

Benchmark	Current: b72167f	Previous: 67c1b2b	Performance Ratio
startup_benchmark Build Time	117.82 s	122.88 s	0.96 ❗
startup_benchmark File Size	0.87 MB	0.87 MB	1.00 ❗
Startup Time - 1 core	0.97 s (±0.03 s)	1.00 s (±0.02 s)	0.96
Startup Time - 2 cores	0.98 s (±0.02 s)	1.00 s (±0.02 s)	0.98
Startup Time - 4 cores	0.99 s (±0.03 s)	0.99 s (±0.03 s)	1.00
multithreaded_benchmark Build Time	120.84 s	123.06 s	0.98 ❗
multithreaded_benchmark File Size	0.97 MB	0.97 MB	1.00 ❗
Multithreaded Pi Efficiency - 2 Threads	89.07 % (±8.72 %)	88.68 % (±8.46 %)	1.00
Multithreaded Pi Efficiency - 4 Threads	43.76 % (±3.85 %)	44.47 % (±3.36 %)	0.98
Multithreaded Pi Efficiency - 8 Threads	25.66 % (±2.05 %)	25.46 % (±2.13 %)	1.01
micro_benchmarks Build Time	181.76 s	172.67 s	1.05 ❗
micro_benchmarks File Size	0.98 MB	0.98 MB	1.00 ❗
Scheduling time - 1 thread	73.98 ticks (±4.36 ticks)	75.22 ticks (±3.59 ticks)	0.98
Scheduling time - 2 threads	38.61 ticks (±2.16 ticks)	38.66 ticks (±2.11 ticks)	1.00
Micro - Time for syscall (getpid)	4.67 ticks (±0.25 ticks)	4.73 ticks (±0.33 ticks)	0.99
Memcpy speed - (built_in) block size 4096	56888.48 MByte/s (±40235.13 MByte/s)	57776.95 MByte/s (±40955.91 MByte/s)	0.98
Memcpy speed - (built_in) block size 1048576	28767.23 MByte/s (±24062.43 MByte/s)	29119.47 MByte/s (±24404.01 MByte/s)	0.99
Memcpy speed - (built_in) block size 16777216	25581.13 MByte/s (±21660.23 MByte/s)	24320.25 MByte/s (±20600.90 MByte/s)	1.05
Memset speed - (built_in) block size 4096	57194.58 MByte/s (±40430.99 MByte/s)	57846.57 MByte/s (±40990.95 MByte/s)	0.99
Memset speed - (built_in) block size 1048576	29611.17 MByte/s (±24552.52 MByte/s)	29837.87 MByte/s (±24806.06 MByte/s)	0.99
Memset speed - (built_in) block size 16777216	25908.84 MByte/s (±21774.53 MByte/s)	25074.19 MByte/s (±21076.19 MByte/s)	1.03
Memcpy speed - (rust) block size 4096	53168.09 MByte/s (±39021.55 MByte/s)	51978.32 MByte/s (±37541.67 MByte/s)	1.02
Memcpy speed - (rust) block size 1048576	28723.40 MByte/s (±24155.27 MByte/s)	28975.81 MByte/s (±24306.78 MByte/s)	0.99
Memcpy speed - (rust) block size 16777216	24933.33 MByte/s (±21170.78 MByte/s)	24942.26 MByte/s (±21157.69 MByte/s)	1.00
Memset speed - (rust) block size 4096	53484.69 MByte/s (±39187.19 MByte/s)	52746.25 MByte/s (±38138.36 MByte/s)	1.01
Memset speed - (rust) block size 1048576	29569.57 MByte/s (±24661.14 MByte/s)	29825.10 MByte/s (±24812.07 MByte/s)	0.99
Memset speed - (rust) block size 16777216	25224.88 MByte/s (±21268.11 MByte/s)	25391.70 MByte/s (±21352.68 MByte/s)	0.99
alloc_benchmarks Build Time	163.00 s	155.86 s	1.05 ❗
alloc_benchmarks File Size	0.94 MB	0.94 MB	1.00 ❗
Allocations - Allocation success	100.00 %	100.00 %	1
Allocations - Deallocation success	100.00 %	100.00 %	1
Allocations - Pre-fail Allocations	100.00 %	100.00 %	1
Allocations - Average Allocation time	9901.86 Ticks (±446.35 Ticks)	8052.09 Ticks (±709.81 Ticks)	1.23 ❗
Allocations - Average Allocation time (no fail)	9901.86 Ticks (±446.35 Ticks)	8052.09 Ticks (±709.81 Ticks)	1.23 ❗
Allocations - Average Deallocation time	1073.33 Ticks (±558.57 Ticks)	929.37 Ticks (±301.24 Ticks)	1.15
mutex_benchmark Build Time	156.68 s	165.12 s	0.95 ❗
mutex_benchmark File Size	0.97 MB	0.97 MB	1.00 ❗
Mutex Stress Test Average Time per Iteration - 1 Threads	15.74 ns (±0.91 ns)	15.62 ns (±0.66 ns)	1.01
Mutex Stress Test Average Time per Iteration - 2 Threads	15.30 ns (±0.90 ns)	15.68 ns (±0.95 ns)	0.98

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

[mkroening]

This does not work at this level, since only one virtq can be selected at the same time, but self.selected is not set to false when other virtqs are selected.

[mkroening]

The refactor required to properly support this would be rather big and nontrivial. I am not sure if it would be worth it. Are you okay with closing the PR or what would you prefer? :)

[udaykiriti]

Close the pr if is not that good:| and let me about remaining commits