Ensure we never benchmark less than two salts for "Many salts".

[magnumripper]

Closes #3912.

[magnumripper]

@solardiz I never found a format to positively test this with... Just dry-running, I could see why it could happen and why it should not with this fix - but I can't say I'm 100% sure.

[solardiz]

Here's my test prior to this PR's changes:

[solar@localhost src]$ grep BenchmarkMany ../run/john.conf 
BenchmarkMany = Y

[solar@localhost src]$ GOMP_CPU_AFFINITY=0-255 GOMP_SPINCOUNT=10000 ../run/john -test -form=ecryptfs
Will run 256 OpenMP threads
Benchmarking: eCryptfs (65536 iterations) [SHA512 512/512 AVX512F 8x]... (256xOMP) DONE
Warning: "Many salts" test limited: 1/256
Many salts:     6113 c/s real, 24.6 c/s virtual
Only one salt:  6206 c/s real, 24.6 c/s virtual

[solar@localhost src]$ GOMP_CPU_AFFINITY=0-255 GOMP_SPINCOUNT=10000 ../run/john -test -form=ecryptfs
Will run 256 OpenMP threads
Benchmarking: eCryptfs (65536 iterations) [SHA512 512/512 AVX512F 8x]... (256xOMP) DONE
Warning: "Many salts" test limited: 1/256
Many salts:     6206 c/s real, 24.6 c/s virtual
Only one salt:  6159 c/s real, 24.6 c/s virtual

[solardiz]

Same test on branch 3912:

[solar@localhost src]$ git branch
* (detached from bleeding-jumbo/3912)
  bleeding-jumbo
[solar@localhost src]$ GOMP_CPU_AFFINITY=0-255 GOMP_SPINCOUNT=10000 ../run/john -test -form=ecryptfs
Will run 256 OpenMP threads
Benchmarking: eCryptfs (65536 iterations) [SHA512 512/512 AVX512F 8x]... (256xOMP) DONE
Warning: "Many salts" test limited: 2/256
Many salts:     6159 c/s real, 24.6 c/s virtual
Only one salt:  6068 c/s real, 24.6 c/s virtual

[solar@localhost src]$ GOMP_CPU_AFFINITY=0-255 GOMP_SPINCOUNT=10000 ../run/john -test -form=ecryptfs
Will run 256 OpenMP threads
Benchmarking: eCryptfs (65536 iterations) [SHA512 512/512 AVX512F 8x]... (256xOMP) DONE
Warning: "Many salts" test limited: 2/256
Many salts:     6159 c/s real, 24.6 c/s virtual
Only one salt:  6206 c/s real, 24.6 c/s virtual

[solardiz]

This passed my testing. Thanks!