lseek: Fixed incorrect lseek behaviour with whence Set or Cur

[jounathaen]

I had buggy code that was using lseek.

Turns out, FUSE is a little odd in that regard (I might be wrong, but this is how it looks to me):

• All reads and writes in FUSE are referenced to an absolute offset.
• For that reason, reads and writes don't change the file offset in virtiofsd (it uses preadv64 internally)
• Lseek does exist in FUSE and in virtiofsd it maps to the host's lseek. It also returns the host's lseek return value.
• In Hermit, this also seems to return meaningful results for seeks from the start and the end, but since a read/write doesn't change the file offset, lseek with a Cur whence doesn't return a meaningful value.
• As mentioned before, read/write in virtiofsd is not affected by the file offset, therefore simply forwarding lseek to the deamon is not sufficient for seeking through a file (we already use an internal offset correctly in these cases).

This PR fixes the incorrect behaviour in lseek by:

• only calling virtiofsd for lseeks relative to the file's end (we don't know the size of the file in Hermit, so this inevitable)
• updating the offset in case of Set and Cur.

This should have a slightly positive effect on the performance, since it reduces virtio transactions.

Note that this is not tested for Data and Hole, so these might still be buggy.

[github-actions]

Benchmark Results

Benchmark	Current: c441518	Previous: 250f22e	Performance Ratio
startup_benchmark Build Time	137.67 s	124.48 s	1.11
startup_benchmark File Size	0.91 MB	0.91 MB	1.00
Startup Time - 1 core	0.94 s (±0.01 s)	0.93 s (±0.02 s)	1.01
Startup Time - 2 cores	0.95 s (±0.02 s)	0.93 s (±0.03 s)	1.02
Startup Time - 4 cores	0.96 s (±0.02 s)	0.95 s (±0.02 s)	1.01
multithreaded_benchmark Build Time	139.44 s	122.65 s	1.14
multithreaded_benchmark File Size	1.02 MB	1.02 MB	1.00
Multithreaded Pi Efficiency - 2 Threads	3.22 % (±15.47 %)	2.79 % (±13.37 %)	1.16
Multithreaded Pi Efficiency - 4 Threads	1.76 % (±8.46 %)	1.58 % (±7.57 %)	1.12
Multithreaded Pi Efficiency - 8 Threads	0.85 % (±4.10 %)	0.82 % (±3.93 %)	1.04
micro_benchmarks Build Time	171.86 s	145.61 s	1.18
micro_benchmarks File Size	1.02 MB	1.03 MB	1.00
Scheduling time - 1 thread	3.78 ticks (±18.15 ticks)	3.05 ticks (±14.65 ticks)	1.24
Scheduling time - 2 threads	1.57 ticks (±7.53 ticks)	1.56 ticks (±7.50 ticks)	1.00
Micro - Time for syscall (getpid)	0.20 ticks (±0.94 ticks)	0.18 ticks (±0.86 ticks)	1.09
Memcpy speed - (built_in) block size 4096	1250.00 MByte/s (±6000.00 MByte/s)	1313.03 MByte/s (±6302.52 MByte/s)	0.95
Memcpy speed - (built_in) block size 1048576	554.00 MByte/s (±2659.21 MByte/s)	692.52 MByte/s (±3324.10 MByte/s)	0.80
Memcpy speed - (built_in) block size 16777216	205.21 MByte/s (±984.99 MByte/s)	211.07 MByte/s (±1013.12 MByte/s)	0.97
Memset speed - (built_in) block size 4096	1290.32 MByte/s (±6193.55 MByte/s)	1290.32 MByte/s (±6193.55 MByte/s)	1
Memset speed - (built_in) block size 1048576	1339.18 MByte/s (±6428.07 MByte/s)	1331.78 MByte/s (±6392.54 MByte/s)	1.01
Memset speed - (built_in) block size 16777216	873.30 MByte/s (±4191.82 MByte/s)	866.63 MByte/s (±4159.85 MByte/s)	1.01
Memcpy speed - (rust) block size 4096	1304.35 MByte/s (±6260.87 MByte/s)	960.00 MByte/s (±4608.00 MByte/s)	1.36
Memcpy speed - (rust) block size 1048576	501.56 MByte/s (±2407.49 MByte/s)	562.78 MByte/s (±2701.33 MByte/s)	0.89
Memcpy speed - (rust) block size 16777216	206.11 MByte/s (±989.32 MByte/s)	203.06 MByte/s (±974.70 MByte/s)	1.01
Memset speed - (rust) block size 4096	1578.95 MByte/s (±7578.95 MByte/s)	1500.00 MByte/s (±7200.00 MByte/s)	1.05
Memset speed - (rust) block size 1048576	979.55 MByte/s (±4701.85 MByte/s)	1289.16 MByte/s (±6187.96 MByte/s)	0.76
Memset speed - (rust) block size 16777216	880.28 MByte/s (±4225.35 MByte/s)	882.37 MByte/s (±4235.38 MByte/s)	1.00
alloc_benchmarks Build Time	159.83 s	141.09 s	1.13
alloc_benchmarks File Size	0.98 MB	0.98 MB	1.00
Allocations - Allocation success	2.00 % (±13.86 %)	2.00 % (±13.86 %)	1
Allocations - Deallocation success	1.40 % (±9.69 %)	1.40 % (±9.74 %)	1.00
Allocations - Pre-fail Allocations	2.00 % (±13.86 %)	2.00 % (±13.86 %)	1
Allocations - Average Allocation time	261.32 Ticks (±1810.83 Ticks)	245.11 Ticks (±1698.52 Ticks)	1.07
Allocations - Average Allocation time (no fail)	261.32 Ticks (±1810.83 Ticks)	245.11 Ticks (±1698.52 Ticks)	1.07
Allocations - Average Deallocation time	17.32 Ticks (±119.99 Ticks)	17.05 Ticks (±118.16 Ticks)	1.02
mutex_benchmark Build Time	156.02 s	142.06 s	1.10
mutex_benchmark File Size	1.03 MB	1.03 MB	1.00
Mutex Stress Test Average Time per Iteration - 1 Threads	0.30 ns (±2.08 ns)	0.34 ns (±2.36 ns)	0.88
Mutex Stress Test Average Time per Iteration - 2 Threads	0.38 ns (±2.63 ns)	0.36 ns (±2.49 ns)	1.06

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