Something was messy with the DELL Firmware.
I’ve refreshed it and completely erased the NVRAM. Reconfigured the firmware and now the host behave as expected:
[root@c4140 ~]# !vdsm
vdsm-client Host getCapabilities | egrep "cpuFlags|cpuModel|Skylake"
"cpuFlags": "pln,msr,acpi,sse2,smx,rdrand,cqm_occup_llc,xsaveopt,rdseed,rtm,epb,sse,hypervisor,ibrs,cmov,nopl,cpuid_fault,pse,f16c,spec_ctrl,adx,constant_tsc,bts,rdt_a,pae,nx,tsc,x2apic,sep,pat,cqm_mbm_total,pebs,xsave,smep,ds_cpl,fma,ospke,mca,mmx,pge,pku,pcid,aperfmperf,ssse3,flexpriority,cqm,avx512dq,avx512vl,fpu,umip,flush_l1d,ssbd,lm,syscall,movbe,vpid,ht,xsavec,invpcid_single,3dnowprefetch,tsc_deadline_timer,cx8,rep_good,tm2,avx,cx16,rdtscp,ss,popcnt,lahf_lm,stibp,arch_perfmon,smap,clflushopt,invtsc,vmx,dts,xsaves,md_clear,dtherm,avx512f,bmi2,mpx,arch-capabilities,dtes64,avx512cd,mba,avx2,ept,pts,vme,vnmi,fxsr,pschange-mc-no,dca,avx512bw,tsc_adjust,cqm_llc,pclmulqdq,cat_l3,bmi1,monitor,pti,arat,abm,cpuid,clflush,mce,sse4_2,erms,nonstop_tsc,apic,cdp_l3,fsgsbase,sdbg,art,xgetbv1,tpr_shadow,cqm_mbm_local,clwb,pdpe1gb,xtpr,ida,de,pbe,intel_pt,ibpb,est,intel_ppin,tm,pni,aes,amd-ssbd,md-clear,skip-l1dfl-vmentry,hle,pdcm,invpcid,mtrr,pse36,sse4_1,xtopology,model_core2duo,model_pentium2,model_Skylake-Server-IBRS,model_Haswell,model_Skylake-Server,model_IvyBridge-IBRS,model_Penryn,model_Broadwell-noTSX-IBRS,model_qemu64,model_n270,model_kvm32,model_coreduo,model_Broadwell-IBRS,model_Skylake-Client-noTSX-IBRS,model_Haswell-IBRS,model_Broadwell-noTSX,model_Skylake-Client,model_SandyBridge,model_Skylake-Server-noTSX-IBRS,model_SandyBridge-IBRS,model_Broadwell,model_kvm64,model_Nehalem-IBRS,model_IvyBridge,model_pentium,model_Skylake-Client-IBRS,model_Conroe,model_Haswell-noTSX,model_Opteron_G2,model_Westmere,model_qemu32,model_486,model_pentium3,model_Opteron_G1,model_Westmere-IBRS,model_Haswell-noTSX-IBRS,model_Nehalem",
"cpuModel": "Intel(R)
Xeon(R) Gold 6130 CPU @ 2.10GHz",
Not sure what really happened, but those actions solved the issue.