CRM incorrect processing for CRM_MPLS_NEXTHOP

[msosyak]

In void CrmOrch::getResAvailableCounters() for case CrmResourceType::CRM_MPLS_NEXTHOP: we do
static_cast<sai_object_type_t>(crmResSaiAvailAttrMap.at(res.first)); but after this PR #2008 the value was changed from SAI_OBJECT_TYPE_NEXT_HOP to SAI_SWITCH_ATTR_AVAILABLE_IPV4_NEXTHOP_ENTRY so the logic in void CrmOrch::getResAvailableCounters() for case CrmResourceType::CRM_MPLS_NEXTHOP: become broken and start produce the followin errors in SONiC:

Nov 23 16:23:54.673681 sonic ERR swss#orchagent: :- objectTypeGetAvailability: attr SAI_L2MC_ENTRY_ATTR_PACKET_ACTION is not resource type
Nov 23 16:23:54.673681 sonic ERR swss#orchagent: :- getResAvailableCounters: Failed to get availability for object_type 53 , rv:-5

Link to related source code https://github.com/Azure/sonic-swss/blob/01c243a1cfa397065f9d8f5519168bc556cf6328/orchagent/crmorch.cpp#L589
https://github.com/Azure/sonic-swss/blob/01c243a1cfa397065f9d8f5519168bc556cf6328/orchagent/crmorch.cpp#L67

[msosyak]

@smaheshm @qbdwlr @prsunny @kcudnik Please take a look

@kperumalbfn FYI

[qbdwlr]

@smahesh @prsunny @kcudnik @kperumalbfn The fix in #2008 is completely wrong and likely still triggers the same error that it was supposed to fix.
Look at the code for CrmOrch::getResAvailableCounters(): CRM_MPLS_NEXTHOP uses the SAI API sai_object_type_get_availability(); it does not call sai_switch_api->get_switch_attribute().
There are no SAI_SWITCH_ATTRs for MPLS Nexthops, because during the SAI Community review, I was asked to remove the SAI_SWITCH_ATTRs that I had defined and use sai_object_type_get_availability with SAI_OBJECT_TYPE instead.
Not sure what platform was being tested when the issue for #2008 was seen, but the correct solution is to revert the change in #2008 and add support for the SAI_OBJECT_TYPE in sai_object_type_get_availability(). For VS platform, see the support I provided in VirtualSwitchSaiInterface::objectTypeGetAvailability().
Reviewing other recent changes in CrmOrch::getResAvailableCounters(), it appears that CRM_SRV6_NEXTHOP will have similar issues as CRM_MPLS_NEXTHOP and should be addressed as I have recommended.
CRM_SRV6_MY_SID_ENTRY has no unique processing and should be grouped with CRM_MPLS_INSEG and CRM_NEXTHOP_GROUP_MAP; it does not require a separate case.
I see no support for either CRM_SRV6_MY_SID_ENTRY or CRM_SRV6_NEXTHOP in vslib. Support needs to be added in VirtualSwitchSaiInterface::objectTypeGetAvailability().

[prsunny]

@smaheshm , would you take a look. Looks like we may have to revert #2008

[smaheshm]

#2057