Trigger links instabilities when the OH FPGA receives S-bit

I'm using this issue to keep track of a behavior recently discovered on the GEM OH triggers. Updates will be added here to keep track of the problem.

During the investigations of the spurious signals/induced noise upon L1A, we noticed an instability of the trigger links. We first thought about an effect of the L1A themselves, but this hypothesis has been ruled out. By instability, we mean 8b10b symbol errors and misses commas.

Contrarily to the first thoughts, the problem is triggered when the OH FPGA receives and processes S-bits. Masking the VFAT at the OptoHybrid level (with the BEFE.GEM.OH.OH{}.FPGA.TRIG.CTRL.VFAT_MASK registers) removes the instability. L1A are still being sent and S-bit produced and sent by the VFAT.

The issue was reproduced today at p5 where the trigger links to the GEM BE board showed instabilities up to the moment where all VFAT were masked from the trigger path.

It is unknown if the problem also