pop - like popping a fuse, except, popped fuses are a lot more economical to replace compared to network interfaces. Nothing at all ominous about that, at least, not ot me. But, each individual has their own sensitivities, so to each their own. I once heard about the cable guy who, on a hot, sweaty summer day, accidentally crossed his metal utility ladder across a pair of aerial power lines. He was killed, apparent cause was officially stated as accidental electrocution. Witnesses claim he literally smoked, and the discharging juice knocked out a large hunk of his buttock in the process of seeking it's path of least resistance. To me, that is at least in the neighborhood of ominous.
At the physical layer, Ethernet operating over twisted pair wire is designed with the intention of having 2 and only 2 devices attached to a given twisted pair wire link such that 1 device is connected from each end of the wire link. You're considering attaching 3 devices to a given wire link, 1 at one end, 2 at the other end. All I'm saying is, Ethernet was not designed for this. You might get lucky but even if you don't, I would not expect the damage to approach that of the unlucky cable guy. Even if you do, there are several other layers above the physical layer where problems, or "conflicts" may occur that, again, might cause your setup to not work.
Theoretically, there may be a toggle option, just not sure how practical it is. For instance, there's been a gizmo available for decades called an RJ45 AB Switch, AKA manual switch. Applictions are numerous. Back when Ethernet only ran 10Mbps, I experimented with one of these, but keep in mind, this gizmo came into use prior to the present wave of Ethernet, during an era when analog voice "telephone traffic" was the most likely candidate for traversing twisted pair wires with RJ45 plugs. A lot of the techs suggested my experiment would not work because switching contacts begin to break down once the signal gets about 5 to 6 MHz. This was the theory that most all of us techs were familiar with. At the same time, they were curious to see how this actually played out in practice. I've observed that frequently there is significant difference between the theory of how things are supposed to work compared to the practical real life every day playing out of things. I tried it. It worked, sort of. I was able to get my workstation to reliably communicate on the network through this mechanical toggle switch. Each time I toggled the switch, the PC being switched to generally needed to be rebooted so that it sort out the nework protocol. I seem to recall this was the case regardless of which protocol was in use (IPX or IP). Back then, the state of Microsoft PC-OS was only Win95. By the time 100Mbps Fast Ethernet came along, IPX had nearly become obsolete, the MS-OS had gotten a lot better with how it handled it's network protocol routine, but, unfortunately, there was still the potential problem with switching contacts. The 100Mps setup was very erratic and the general conclusion was that the manual switch was the cause. There were some manufacturers who claimed to have developed a manual switch that was designed to offset the problems that occur when such a high speed signal traverses a physical/mechanical electrical connection like those inside manual switches, but I do know if they work as advertised.