Supported Gigabit Adapter for ESXi 6.0 no longer recognized in 6.5u1

[tateconcepts]

Hi all,

So I managed to get my hands on an older firmware version compatible with ESXi and vSphere as per HCL. It seems that 6.5u1 does not detect the Thunderbolt to Gigabit adapters. See below

[root@localhost:~] esxcli hardware platform get

Platform Information

   UUID: 0x74 0x8d 0xa 0x29 0x9b 0x8a 0x3d 0x57 0x83 0x39 0x72 0x53 0x65 0x2a 0xe9 0xcc

   Product Name: MacPro6,1

   Vendor Name: Apple Inc.

   Serial Number: F5KR40******

   IPMI Supported: false

[root@localhost:~] esxcli network nic list

Name    PCI Device    Driver  Admin Status  Link Status  Speed  Duplex  MAC Address         MTU  Description                                            

------  ------------  ------  ------------  -----------  -----  ------  -----------------  ----  --------------------------------------------------------

vmnic0  0000:0b:00.0  tg3     Up            Up            1000  Full    00:3e:e1:c8:98:06  9000  Broadcom Corporation NetXtreme BCM57762 Gigabit Ethernet

vmnic1  0000:0c:00.0  tg3     Up            Up            1000  Full    00:3e:e1:c8:98:05  9000  Broadcom Corporation NetXtreme BCM57762 Gigabit Ethernet

Can your custom VIB be used with 6.5u1 and I thought this issue was resolved with 6.0?

I cannot locate where it has been depricated so I'm a little baffled. I actually needed this to test an RHEL 7.4 guest that sees the vNIC but cannot get an IP from the single vSwitch0 that all other guests are able to use with onboard pNIC. I wouldn't think we would need anything from Apple (as I know they are mum on most topics with SE) since we are using the same hardware and specs that they have already provided.

Does anyone have any suggestions here to what to do to get these adapters functional or are we at the end of the line for all Apple devices (I have an older MacMini and older release of ESXi but they are useless because they do not support ECC memory, do not support more that 16GB and moreover do not have 12 core Xeon CPU which is needed by my workload)?

[tateconcepts]

The custom VIB I am referring to is that of William Lam's blog here

https://www.virtuallyghetto.com/2013/09/running-esxi-55-on-apple-mac-mini.html

and also reported to work OOB with 6.0 here

https://www.virtuallyghetto.com/2015/02/esxi-6-0-works-ootb-for-apple-mac-mini-mac-pro.html

My question is what changed between 6.0 and 6.5 where this device would no longer be detected or recognized (it's the same hardware and vendor ID to my knowledge within the adapter)?

[daphnissov]

Well ESXi 6.0 is not ESXi 6.5 U1. There have been quite a lot of changes, including in the driver code. I see you've posted a comment on his blog and so that's probably the best place to start. He also has a profile at VMTN as lamw​ and so maybe he'll see this and respond.

[tateconcepts]

daphnissov​ Thanks for the reply. I wish I could tell what specifically changed. I tried reading his blog, the many comments and even some from Pau's TinkerTry site (which is really good I might add). At this point, if I've thrown this much money into a "trashcan" literally - I might as well dump my findings for all to exploit as I cannot reach VMware for any support (I'm not a VMware admin in this day and age), Apple product security is treating me like Felix Krause and one is essentially pwned either way.

Therefore I am left with no options at this point and run a risk of a whopping amount of vulnerabilities, with an EFI that is known to be exploitable OR upgrade to 6.5u1 and lose everything that is expected from the HCL OR I do away with ESXi, permit 10.13 to put an end of the Apple of my EFI and you lose any Apple hardware forever, meaning you lose market share and are essentially throwing in the towel. I bet I can get Apple SE to reconsider this IF someone from VMware also brings their A game. With the new Mac Pro ready to launch later this year along with a massive surplus of hardware for the previous versions, I expect to see more on this to come. I'd say a botnet of Mac Pro's running insecure hypervisors that can get pwned at ring -2 (because of Apple not working with VMware) and -1 (for VMware not making an attempt to remedy wonky code) is bad enough that they should either update the HCL with a disclaimer pointing back at Apple and their developer sites or permit a paid customer to bring litigation to the table. Money always gets people talking and is easy addressed in a BIA, but gossip and bad PR risk can only be measured qualitatively.

Do any admins run this community site actually work at VMware? You've got some good talent here and at no cost!

(I never could figure out why VMware wanted Tanium in their ecosystem, I would have thought that DDP with Cylance would be the dog eating its food BTW).