Question

How to get the transmission time of jam signal?

Answer 1

What is Jam Signal?

A flag sent by a gadget on an Ethernet system to show that a crash has happened on the system.

Impacts happen on Ethernet systems since access to media (for the most part a link) depends on conflict - that is, on a first-come, first-served premise. On the off chance that two stations endeavor to take control of the medium in the meantime and start transmitting, the two stations will identify each other's flag and understand that a crash has happened.

The two stations at that point issue a stick flag, which advises every single other station on the system of the impact. They all must hold up a brief timeframe before endeavoring to transmit once more.

The time span is arbitrary for each station so the retransmissions won't cause more crashes. The stick flag sent by one transmitting station must begin with a 62-bit example of substituting 1s, trailed by a 32-bit grouping that gives a fake checksum incentive to the next transmitting station. This 32-bit arrangement can't be equivalent to the recurrent excess check (CRC) esteem for the edge going before the stick.

The primary thing done before endeavoring a transmit is a bearer sense, the CS part of CSMA. This searches for transmissions. [Even a flood of zeros will cause a positive transporter sense, since Ethernet's bit encoding is a Manchester encoding (at any rate for 10baseT and 100baseTx), with 0 being an upward progress from - 0.85V to 0.85V, and 1 being a descending change from 0.85V to - 0.85V. So you'd see consistent changes. This is altogether done baseband, btw, dissimilar to simple encodings] If there is a transmission happening, the hub will keep detecting until there is not any more flag, and 96 bit-times more (at 0.1 µs per bit for 10 Mbps Ethernet, and 10 ns for each piece for 100 Mbps Ethernet, that is 9.6 µs and 0.96 µs individually) before it endeavors transmission.

In the event that there's no sign of danger, transmission begins, and edges are conveyed onto the wire, utilizing the Manchester encoding.

On the off chance that the channel spread deferral made another hub begin transmitting 96 bit-times after the past hub's transmission, yet not so as to see our transmissions, we quit transmitting, and emanate a stick flag - 48 bit-times (that is 4.8 µs for 10 Mbps joins, and 0.48 µs for 100 Mbps connections, and 6 octets' worth) of '1' (descending) advances - and enters exponential backoff. The time of the backoff is (for the principal crash) an irregular postponement of up to 512 piece times, yet a different of 512 piece times (i.e., no deferral, or 512 piece times for the main impact). The following crash can cause a postponement of 0, 512, 1,024 or 2,048 piece times, picked arbitrarily. The third crash can cause a postponement of 0, 512, 1,024, 2,048, 4,096, 8,192, 16,384 or 32,768 piece times. The deferral won't develop past 262,144 piece times.

I accept (yet am not sure) this is taken care of by means of the MII (media autonomous interface) of the Ethernet controller. This suggests an Ethernet driver can't specifically cause a stick flag to be affected, unless the MII itself uncovered a custom interface to do this (most don't), in which case you'd have to compose or alter a driver (and presumably likewise larger amount interfaces, e.g. ioctls, to have the capacity to do this from userspace). At any rate, the BSD and Linux arrange stacks and equipment particular drivers that I've perused and hacked don't appear to deal with CSMA/CD in programming, and all the controller details and manuals I've perused recommend this is taken care of in equipment.

You could present an obsessive defense by setting an extra hub on the Ethernet section and working the fragment fifty-fifty duplex mode, and having the additional hub transmit a great deal of 64-octet outlines, dispersed 97 or so bit-times separated (sufficiently long for the bearer sense to demonstrate that the channel is free, yet sufficiently short for channel engendering deferrals to produce results; you may require a few feet of cabling to accomplish this, however; maybe with bunches of repeaters to disturb the proliferation delays). At that point you're very prone to see stick signals being conveyed a portion of the time.