Topic: Antbleed: A remote shutdown backdoor in antminers - page 6. (Read 8076 times)

You don't understand. This phones home to the server every 10 minutes, and if the server says "false", it shuts down. Thats it. It can't be switched off on the miner, it will always try to phone home and always shutdown if it receives the command. Only way to turn it off is to block bitmains server with a firewall.

Yes a hard fork would be required for any changes to the scrypt algo.


In the distant future the cryptographic proof will all need to be changed as ECDSA will become insecure but it is presumed to be secure until 2050.
In addition at some point we will need to slowly transition in a scrypt change on SHA-256 to what I presume will be SHA-3 if weaknesses are discovered in 256 but it will be well before it is 100% Broken. This is all speculation though and until any advantages to Sha-3 vs 256 affecting cryptanalytic attacks are found it is not a concern.
But I can imagine the fracas then although the logic that one hard fork doing a lot of things at once is better than 100 mini hard-forks for every little thing from 2013 still has weight.

Either way were far off from those forks we just need to look at the one now ^^.

https://bitcointalksearch.org/topic/should-bitcoin-move-to-sha-3-146191
https://bitcointalksearch.org/topic/m.1301958
https://books.google.ca/books?id=UsRsCQAAQBAJ&pg=PA128&lpg=PA128&dq=ecdsa+secure+till+2050&source=bl&ots=9qMTUSGSOG&sig=-qIrJ-MDblLYsdXV5oJMh-mimKA&hl=en&sa=X&ved=0ahUKEwiUoo-5jMTTAhUY92MKHSv_BKMQ6AEIMDAC#v=onepage&q=ecdsa%20secure%20till%202050&f=false
http://blog.oleganza.com/post/42523601710/how-to-steal-all-coins

Yes and yes.

Well, if that happen, it will be done in an emergency.
It may happen if the chain split, and the Core side is being attacked and vulnerable.
I think some few devs are already working on this in case we need to deploy the PoW HF.
It's still an extreme scenario.

small clarification question: IF we did end up deciding a POW change was needed, wouldn't that require a hard fork? like, even more-so than any aspect of the scaling debate?

serious question. how could we possibly ever change the algo at this point? we know full well unless it was 100% proven broken we'd never get agreement a change was needed, and even if that happened, we'd never get agreement on WHAT to change to.

can this backdoor be used for malicious things? to steal funds or something?


i can't see why not, they are the best one, and it also make sense because it match the percentage of chinese miners, which is around 70%, probably all pool are running s9 now

if someone else is running s5 or s7, would be a very minority, and i'm actually surprised that there are 30% of miners running those old one

its because segwit as a softfork has last month been found to hit a hurdle, so now people are hurrying to find any reason to push mining aside to drag segwit in without having to change segwit as a soft fork.(even though segwit is not even active and can easily be redone as a proper community uniting version that includes other features the community desire.. rather than the tier network cesspit version)

funny part is that segwit is using its own backdoor (going soft) to bypass node consensus and its admitted that they want to make it easier to implement changes even easier in the future. (code backdoors that outsiders can exploit to add in their own features by 'going soft')

EG imagine if hearne coded segwit as a soft fork(line for line the same) and said how he wants to add more ways to add new things in without node consensus, would you remain positive about it?

EG imagine if hearne tried to blame pools for holding things back(after giving only pools the vote) and wanted to find ways to ban/orphan blocks/pools based on brand bias.. not tx/block validity

A lot of it is probably due to people wanting to find as many ways to discredit someone else, i.e. some Core supporters looking to discredit BU, Bitmain, etc. and some BU supporters looking to discredit Core. Some of it is likely from projects that were started several months ago but only now reached their conclusion (like reverse engineering an asic and related closed-source firmware for asicboost as that takes a long time).

the first thing that came to my mind after reading this was a question! "Why now". why is it that these days, all these news about mining, miners' equipment, backdoor, boost, etc are coming out. ASICs weren't created yesterday!

I wasn't specifically calling you out or Jonald but more or less the many btcu shills while quoting the other reply, and the Chinese thing is because Jihan is Chinese and he is willing to throw money around to get his way.

I don't want a pow change I would like there to definitely be more asic producers and them come from all around the world. The issue is the production is just going to be cheaper in China due to the many factories and cheap labor, the other issue is Jihan buying out his competition there. I don't mine Bitcoin directly anymore because it's just past me with noise and power consumption for a home miner, but I do worry about the network. Jihan seems like a bad player in the Bitcoin world and it seems like he is starting to sabre rattle to protect his interest and this could just be a new weapon in his arsenal especially with all future miners.

Along with the fix we know large mines won't take down their miners and flash them, too much work for a large mine until an attack happens.

It may be related to minerlink, but it itself is not governed by minerlink. It does its own entirely separate thing. However it could potentially be used by minerlink to remotely stop an antminer from mining, but the way that they implemented that is rather odd and convoluted so it seems unlikely. Perhaps it was meant for something else related to minerlink but was never completed.

Achow, are you saying antbleed is not miner link at all?  But completely different code?

As I said earlier, this is completely unrelated to minerlink except for the fact that it shares a domain name. The phone home code runs regardless of user input or settings.

Sure larger mining farms might have something like that, but for the casual "I have have cheap electricity so I got a couple R4's/S9's" miner, I highly doubt they have such settings. Those who still mine from home with antminers are vulnerable. Additionally, most networking stuff by default doesn't restrict outbound connections, so unless the miner was specifically concerned about something already on a machine inside the network sending something out of the network, this default would not be changed.

like someone said... very old news,

minerlink is disabled by default...

... and most miners only allow outbound traffic from their rig to go to pool stratum anyway, not to Bitmain or anyone else.

just more pointless FUD.

We can't even handle larger blocks without a threat of splitting the chains, I think the only true thing that would push the main chain into a fully supported algorithm change is finding out the Chinese government is in control of Bitmain.

I was thinking along the lines of if your not mining and if someone hacks minerlink and turns off all the not patched antiminer's for a while the odds of generating a new block go up relative to the decrease in difficulty until they turn them on again. But true enough specific targeted attacks could change signalling kind of like finding a bug in the nodes you can turn them off for a few days only this could also brick stuff and take the hash out of the game.

The key thing is changing the manual hosts entry to 127.0.0.1 auth.minerlink.com since phone home is hardcoded ... what a PITA at least what I get from reading that reddit and the related posts here.
https://www.reddit.com/r/Bitcoin/comments/67qwqv/antbleed_exposing_the_malicious_backdoor_on/dgsk6cf/
(Speaking of Tesla and the current patch I wonder if you can change the host file to phone home to the Tesla car instead ha-ha)

Since a staff posted the OP I agree with Franky posts go to increase awareness of a patch than keeping the thread silent.

you may realise other countries are in the manufacturing game and its not (current 67% nay/abstainers) of asics are bitmain or chinese.
thats the reddit hysteria of blaming the segwit nays'/abstaining on one person(illogical facepalm).

many people who are independent like myself who just think critically rather than ass kissery see some flaws in segwits 2 merkle approach. even gmax sees now that 2merkle approach has hit issues last month

looking deeper at the matter i can see a few countries with low labour, low facility costs and low electric costs. and some manufacturers are already using them.

but it does not matter about blaming china(as a country).. what people should be doing is defending a decentralised diverse peer network not just pointing fingers to find excuses of why bitcoin should centralise to one one brand by pretending its getting attacked by another brand

it might be worth you taking a step back from the reddit hysteria of blaming the (current 67% nay/abstainers) on jihan because those numbers dont rationally add up,
the community is wider than that

The issue of voting with your wallet when it comes to mining equipment is that whoever sells the best possible chance at ROI will dominate and with the Chinese labor and access of materials it's going to be Bitmain.

Half in the bag already? I can't imagine any large business in China gets anywhere without some government heavy breathing.

mention my name and i appear.
i was going to keep myself quiet because this indeed is a exploit risk unlike asicboost which is a efficiency

anyway achow and frodocooper the antbleed website has explained the issue and the fix

but as you all aware i do like to waffle my opinion when i can. which i was going to hold back on due to this topic originally being an actual risk that needs highlighting without derailing the topic.
but my name was mentioned.

seeing as the OP and first couple posts of topic has the fix aswell as an explaination and people have already started derailing the topic with "kill asics now" "PoW change" "sitting on landmines" we might aswell consider all posts just "post bumps" to keep the topic on front page so people tomorrow and so on know of the issue.

anyway
if you want to change PoW
then that involves a proper network consensus upgrade, which is a great opportunity to then do things properly like a 1merkle segwit which would give a 4mb baseblock and not need the 1mb base 4mb weight tier network cesspit creating crap.

(id prefer the 4mb block to be dynamic to not rely on any dev team to spoon feed limits in future)
aswell as a opportunity to really handle quadratics by keeping maxtxsigops down, eg:4k or below so that the native key users dont gain more spammer causing headache (0.14: maxtxsigops 16k)

thus getting the community everything everyone wants. where native key users get 4mb block and segwit key users get to use segwit where everyone on the same level playing field.

yep no more core tier network concern.
no more band camp debate. but finally a community peer network not filled with half gestures and empty promises

P.S
you can try pigeon holing me into "chinese" or "btcu" all you want.
my opinion is about open diverse decentralised single peer network of many implementations. which has not changed in years

if the only thing you can do is try turning this topic into a "hate the chinese"(as a country) or point fingers at anything not core. then thats your bias.

but if you just want to waste a oppertunity of a proper ful network consensus upgrade just to kill asics. just realise that within a couple months the utopian dream asic killers think they will get of solo mining from their basement again wont last long because someone somewhere will find efficiency methods of any new system and within 6 months the its back to a new form of pool/syndicate/sybil mining groups

They can go full censorship too, anyone who is pro segwit or has been posting anti Bitmain things could get their miner turned off. They link the miner to their customer records and I am sure they are able to find via email and a small bit of googling their Bitcointalk or Reddit username.