Unified nicehash+eth-proxy #417

hackmod · 2018-12-27T12:47:44Z

Nicehash(EthereumStratum/1.0.0) support based on Wolfgang Frisch's working prototype.

Author: Wolfgang Frisch <wfrisch@riseup.net>
Date:   Mon Sep 25 19:24:09 2017 +0200

    NiceHash support: working prototype

unified nicehash + ethproxy and cleanup by @hackmod

rebased and squashed
unified processShare*()
unified nicehash + ethproxy
~~height field added to support extended getWork()~~
support mining.extranonce.subscribe for nicehash
fixed crash and cleanup
remove DifficultyNiceHash using DiffFloatToInit()
remove all nicehash specific options
BUG: do not modify shareDiff to show hashrate correctly
check len(params)
Do not terminate connection if miner submits one duplicate share (by nicehashdev, manually applied by hackmod)

Note

it use modified version of the ethash lib. https://github.com/hackmod/ethereum-ethash

this work supported by Ethersocial Network (ESN)
see also Ethersocial/Proposals#15
cc: @kimmyeonghun @magnalucus

duyk16 · 2018-12-27T13:14:27Z

Your work is great 👍

hackmod · 2019-01-14T23:39:19Z

@sammy007 // hello~
thanks for your great project!

if you have any interest or some free time to review this code, please let me know or I will make a temporal repo to support both nicehash(EthereumStratum/1.0.0) and EthereumStratum/2.0.0 for gangnam testnet. (I'm already prepared for the next PR to support EthereumStratum/2.0.0)

aemet93 · 2019-01-22T17:15:19Z

Hello @hackmod
I am trying to use your pull request, but no success. Have you tested it with real nicehash workers? Worker just connected and nothing more happened. It never creates session, submit work etc. Even if I tried to run your code without changes or merges to my code. It says Nicehash subscribe + ip and nothing till I shut down order. Then it says Client disconnected. Could you please help?
thank you in advance.

hackmod · 2019-02-10T03:41:15Z

@aemet93 // sorry for late answer.
it works nicely with my local pool setup.
what is your setup or any error log?

Ingham27 · 2019-02-22T09:06:30Z

@aemet93

type JSONRpcResp struct {
Id json.RawMessage json:"id"
Version string json:"jsonrpc,omitempty"
Result interface{} json:"result"
Error interface{} json:"error,omitempty"
}

delete omitempty in Error

type JSONRpcResp struct {
Id json.RawMessage json:"id"
Version string json:"jsonrpc,omitempty"
Result interface{} json:"result"
Error interface{} json:"error"
}

works

sammy007 · 2019-02-23T18:26:06Z

Could you squash? Using develop branch of course.

Author: Wolfgang Frisch <wfrisch@riseup.net> Date: Mon Sep 25 19:24:09 2017 +0200 NiceHash support: working prototype unified nicehash + ethproxy and cleanup by hackmod * rebased and squashed * unified processShare*() * unified nicehash + ethproxy * height field added to support extended getWork() * support mining.extranonce.subscribe for nicehash * fixed crash and cleanup * remove DifficultyNiceHash using DiffFloatToInit() * remove all nicehash specific options * BUG: do not modify shareDiff to show hashrate correctly * check len(params) * Do not terminate connection if miner submits one duplicate share (by nicehashdev, manually applied by hackmod) * fixed stale job bug using cached stale jobs. * BUG: old behavior always drop stale jobs * sendJob() again after stale share found. * cache stale jobs info. * use cached headHash/seedHash from stales cache * reuse Compute() result instead of calling Verfiy() again * confirm uniq Extranonce * set ExtranonceSub session flag to support non extranonce.subscribe capable miners for NiceHash * check the length of params * always calc actual difficulty * debug config option added to toggle some logs * disable some log conditionally Co-authored-by: Wolfgang Frisch <wfrisch@riseup.net>

@Ingham27

, fixed by @Ingham27

mikeyb · 2019-02-23T19:04:12Z

FYI, you should be able to squash on merge. Just hit the dropdown arrow on the merge button

sammy007 · 2019-02-23T21:03:51Z

I just wanna look into it and it's easier to read it all at once.

neptunix · 2019-03-31T22:01:38Z

Hi there! I've tried to run it with nicehash and the first thing i've encountered was nonce size mismatch:

	// 4/1/2019 12:07:46 AM2019/03/31 21:07:46 Malformed PoW result from eth@10.42.198.17
	//  nonce: 0xeee00d4b765f
	//  hash1: 0xfffb566c75d59b8bbaf7a4bfaa4e41c83ea1498ef3202f671b38fa61d1bde016
	//  hash2: 0x271ec813d6c4f8c53d526488cb51abc749bbb818dd6fcad76a80a334ccc7c7c1

it indeed does not match the regex (it's len is 12 instead of 16).
var noncePattern = regexp.MustCompile("^0x[0-9a-f]{16}$")

If I skip that check, I get the following error:

		4/1/2019 12:41:08 AM2019/03/31 21:41:08 share difficulty too low, 0.000000 < 260546931191, from eth@10.42.198.17

And that's probably related to nonce length. Maybe I should add leading 0x0 to the nonce value?

hackmod · 2019-03-31T23:29:16Z

Hi!

extranonce (randomHex(4)) + nonce(12 len hex string) makes total 16 bytes len.

(extranonce)
185febe#diff-40ea75f4bd2f5f500a0cc37a432246faR551

(nonce = extranonce + param[2](miner nonce from miner))
185febe#diff-40ea75f4bd2f5f500a0cc37a432246faR265

I guess the extranonce was missing or not properly prepended at pool side.

Hi there! I've tried to run it with nicehash and the first thing i've encountered was nonce size mismatch:
	// 4/1/2019 12:07:46 AM2019/03/31 21:07:46 Malformed PoW result from eth@10.42.198.17
	//  nonce: 0xeee00d4b765f
	//  hash1: 0xfffb566c75d59b8bbaf7a4bfaa4e41c83ea1498ef3202f671b38fa61d1bde016
	//  hash2: 0x271ec813d6c4f8c53d526488cb51abc749bbb818dd6fcad76a80a334ccc7c7c1
it indeed does not match the regex (it's len is 12 instead of 16).
var noncePattern = regexp.MustCompile("^0x[0-9a-f]{16}$")

If I skip that check, I get the following error:
		4/1/2019 12:41:08 AM2019/03/31 21:41:08 share difficulty too low, 0.000000 < 260546931191, from eth@10.42.198.17
And that's probably related to nonce length. Maybe I should add leading 0x0 to the nonce value?

neptunix · 2019-04-01T21:23:25Z

Hi!

Hi, thanks for the reply!

I guess the extranonce was missing or not properly prepended at pool side.

I've dumped the beginning of the communication, could you please give me a hint if something looks wrong here...

4/2/2019 12:13:15 AM004★ new client from 10.42.198.17:59524
4/2/2019 12:13:15 AM004✓ connected to backend
4/2/2019 12:13:15 AM004→ {"id":1,"method":"mining.subscribe","params":["NiceHash/1.0.0","EthereumStratum/1.0.0"]}<0a>
4/2/2019 12:13:15 AM004← {"id":1,"result":[["mining.notify","b9374b251df5b4ed","EthereumStratum/1.0.0"],"3968"],"error":null}<0a>
4/2/2019 12:13:15 AM004→ {"id":2,"method":"mining.authorize","params":["0x3F390F01C44aFe91A40cE9C6Fd03bf7436A0ddC0","x"]}<0a>
4/2/2019 12:13:15 AM004← {"id":2,"result":true,"error":null}<0a>
4/2/2019 12:13:15 AM004← {"method":"mining.set_difficulty","params":[1.1641354547009541],"id":null}<0a>
4/2/2019 12:13:15 AM004← {"method":"mining.notify","params":["32dbbb46","fffb566c75d59b8bbaf7a4bfaa4e41c83ea1498ef3202f671b38fa61d1bde016","7af88f0ffa1444f34a34b7ed37f03030a0c99bd2df499bb50cfae807da6f2033",true],"id":null}<0a>
4/2/2019 12:13:24 AM004→ {"id":3,"method":"mining.submit","params":["0x3F390F01C44aFe91A40cE9C6Fd03bf7436A0ddC0","32dbbb46","4ec000558d7a"]}<0a>
4/2/2019 12:13:25 AM004← {"id":3,"result":false,"error":null}<0a>
4/2/2019 12:13:26 AM004→ {"id":4,"method":"mining.submit","params":["0x3F390F01C44aFe91A40cE9C6Fd03bf7436A0ddC0","32dbbb46","4ec0046b64b5"]}<0a>
4/2/2019 12:13:26 AM004← {"id":4,"result":false,"error":null}<0a>

Backend skipped the wrong nonce length (len 12) and failed with the incorrect difficulty error in this case

3968 - is the extra nonce here, right?

hackmod · 2019-04-01T23:29:39Z

Hi!

Hi, thanks for the reply!

I guess the extranonce was missing or not properly prepended at pool side.

I've dumped the beginning of the communication, could you please give me a hint if something looks wrong here...

4/2/2019 12:13:15 AM004★ new client from 10.42.198.17:59524
4/2/2019 12:13:15 AM004✓ connected to backend
4/2/2019 12:13:15 AM004→ {"id":1,"method":"mining.subscribe","params":["NiceHash/1.0.0","EthereumStratum/1.0.0"]}<0a>
4/2/2019 12:13:15 AM004← {"id":1,"result":[["mining.notify","b9374b251df5b4ed","EthereumStratum/1.0.0"],"3968"],"error":null}<0a>
4/2/2019 12:13:15 AM004→ {"id":2,"method":"mining.authorize","params":["0x3F390F01C44aFe91A40cE9C6Fd03bf7436A0ddC0","x"]}<0a>
4/2/2019 12:13:15 AM004← {"id":2,"result":true,"error":null}<0a>
4/2/2019 12:13:15 AM004← {"method":"mining.set_difficulty","params":[1.1641354547009541],"id":null}<0a>
4/2/2019 12:13:15 AM004← {"method":"mining.notify","params":["32dbbb46","fffb566c75d59b8bbaf7a4bfaa4e41c83ea1498ef3202f671b38fa61d1bde016","7af88f0ffa1444f34a34b7ed37f03030a0c99bd2df499bb50cfae807da6f2033",true],"id":null}<0a>
4/2/2019 12:13:24 AM004→ {"id":3,"method":"mining.submit","params":["0x3F390F01C44aFe91A40cE9C6Fd03bf7436A0ddC0","32dbbb46","4ec000558d7a"]}<0a>
4/2/2019 12:13:25 AM004← {"id":3,"result":false,"error":null}<0a>
4/2/2019 12:13:26 AM004→ {"id":4,"method":"mining.submit","params":["0x3F390F01C44aFe91A40cE9C6Fd03bf7436A0ddC0","32dbbb46","4ec0046b64b5"]}<0a>
4/2/2019 12:13:26 AM004← {"id":4,"result":false,"error":null}<0a>

Backend skipped the wrong nonce length (len 12) and failed with the incorrect difficulty error in this case

3968 - is the extra nonce here, right?

yes right. the extranonce seems correctly responsed by pool server.

{"id":3,"method":"mining.submit","params":["0x3F390F01C44aFe91A40cE9C6Fd03bf7436A0ddC0","32dbbb46","4ec000558d7a"]}
so, for this case, pool server got 3968 + 4ec000558d7a (parms[2]) as 16 bytes len accepted nonce.

this PR already handle it. (I guess you are trying to make your own fix or pool server)

neptunix · 2019-04-02T07:10:07Z

Thanks, @hackmod.

As I see from the code

			// check Extranonce subscription.
			extranonce := cs.Extranonce
			if !cs.ExtranonceSub { extranonce = "" }
			nonce := extranonce + params[2]

the extranonce is not concatenated because cs.ExtranonceSub is false.
and it's set in mining.extranonce.subscribe which was not called by nicehash in my case

case "mining.extranonce.subscribe":
			var params []string
			err := json.Unmarshal(req.Params, &params)
			if err != nil {
				return errors.New("invalid params")
			}
			if len(params) == 0 {
				if err := cs.sendStratumResult(req.Id, true); err != nil {
					return err
				}
				cs.ExtranonceSub = true
				req := JSONStratumReq{
					Id:     nil,
					Method: "mining.set_extranonce",
					Params: []interface{}{
						cs.Extranonce,
					},
				}
				return cs.sendTCPReq(req)
			}
			return cs.sendStratumError(req.Id, []string{
				"20",
				"Not supported.",
			})

Do I miss something else?

this PR already handle it. (I guess you are trying to make your own fix or pool server)
Yes, I'm trying to run the pool with the code from your PR.

Thanks

hackmod · 2019-04-02T07:48:01Z

Thanks, @hackmod.

As I see from the code
			// check Extranonce subscription.
			extranonce := cs.Extranonce
			if !cs.ExtranonceSub { extranonce = "" }
			nonce := extranonce + params[2]
the extranonce is not concatenated because cs.ExtranonceSub is false.
and it's set in mining.extranonce.subscribe which was not called by nicehash in my case

claymore + nicehash mode or ethminer + nicehash mode always call mining.extranonce.subscribe.

how about to apply the following fix?

diff --git a/proxy/stratum.go b/proxy/stratum.go
index fda127c..10fb42a 100644
--- a/proxy/stratum.go
+++ b/proxy/stratum.go
@@ -165,6 +165,7 @@ func (cs *Session) handleTCPMessage(s *ProxyServer, req *StratumReq) error {
                        return cs.sendStratumError(req.Id, "unsupported stratum version")
                }

+               cs.ExtranonceSub = true
                cs.setStratumMode("EthereumStratum/1.0.0")
                log.Println("Nicehash subscribe", cs.ip)
                result := cs.getNotificationResponse(s)

case "mining.extranonce.subscribe":
			var params []string
			err := json.Unmarshal(req.Params, &params)
			if err != nil {
				return errors.New("invalid params")
			}
			if len(params) == 0 {
				if err := cs.sendStratumResult(req.Id, true); err != nil {
					return err
				}
				cs.ExtranonceSub = true
				req := JSONStratumReq{
					Id:     nil,
					Method: "mining.set_extranonce",
					Params: []interface{}{
						cs.Extranonce,
					},
				}
				return cs.sendTCPReq(req)
			}
			return cs.sendStratumError(req.Id, []string{
				"20",
				"Not supported.",
			})

Do I miss something else?

this PR already handle it. (I guess you are trying to make your own fix or pool server)

Yes, I'm trying to run the pool with the code from your PR.

Thanks

ExtranonceSub := false by default to supprt non extranonce.subscribe capable miners for NiceHash.

Please see EthersocialNetwork@e2f191a

neptunix · 2019-04-02T08:16:28Z

Thank you. I got the point.
I think I'd better do it that way:
if (the nonce length == 12 && extranonce is enabled ) => concatenate the values

StayClassic · 2019-06-20T07:07:30Z

any update on this?

aemet93 · 2019-06-20T12:16:23Z

Hello all again.
Everything works with this fixes, but we have another problem, which we cannot identify. It is delta.
I have checked everything many times, all share way from nicehash to our pool, verification and till database. We have stable -15% delta. All recommendations from nicehash support, such as connection, stable uptime etc. was complied. Our pool have very little amount of stale shares, it is much more with regular miners (not from nicehash), and it is ok. Strange thing, that nicehash does not see any rejection from our side, but surely says that speed at pool 15% less than we paying. Will be much thankful for any ideas, tips, and advices, because it becomes real problem for us. Thank you very much.

korjavin · 2019-06-21T07:23:15Z

We have seen this delta before with another implementation of NH protocol.

aemet93 · 2019-06-21T11:52:29Z

Also, maybe anybody know what is this "old stratum" version? I got this message when try to mine on nanopool (this pool works good and doesn't have delta).

aemet93 · 2019-06-24T19:14:12Z

If anyone interested, continue debugging this stuff, I assume that not all shares comes to pool. I have logged all shares, and for 100% share pool respond to NH as true. Speed calculation is obviously based on shares amount, so looks like NH send more shares than we see on pool. Now most interesting question is where is this lost shares?.. I will give a report here, if will find something interesting.

hackmod · 2019-06-25T03:37:11Z

this is the original unified nicehash-v1 branch https://github.com/EthersocialNetwork/open-ethereum-pool/tree/unified-nicehash-v1-original

I guess, lost shares may be caused by the following code. (original NH fix always drop these shares)
EthersocialNetwork@254d409

hackmod and others added 2 commits February 24, 2019 03:41

remove "omitempty" attribute from JSONRpcResp.Error reported by @aemet93

62392c6

, fixed by @Ingham27

hackmod force-pushed the unified-nicehash-v1 branch from e3a3fb3 to 62392c6 Compare February 23, 2019 18:52

baby636 mentioned this pull request Nov 28, 2023

[Snyk] Fix for 33 vulnerabilities baby636/open-ethereum-pool#38

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Unified nicehash+eth-proxy #417

Unified nicehash+eth-proxy #417

hackmod commented Dec 27, 2018 •

edited

duyk16 commented Dec 27, 2018

hackmod commented Jan 14, 2019

aemet93 commented Jan 22, 2019 •

edited

hackmod commented Feb 10, 2019

Ingham27 commented Feb 22, 2019

sammy007 commented Feb 23, 2019 •

edited

mikeyb commented Feb 23, 2019

sammy007 commented Feb 23, 2019

neptunix commented Mar 31, 2019

hackmod commented Mar 31, 2019

neptunix commented Apr 1, 2019 •

edited

hackmod commented Apr 1, 2019

neptunix commented Apr 2, 2019

hackmod commented Apr 2, 2019 •

edited

neptunix commented Apr 2, 2019

StayClassic commented Jun 20, 2019

aemet93 commented Jun 20, 2019 •

edited

korjavin commented Jun 21, 2019

aemet93 commented Jun 21, 2019

aemet93 commented Jun 24, 2019

hackmod commented Jun 25, 2019

Unified nicehash+eth-proxy #417

Are you sure you want to change the base?

Unified nicehash+eth-proxy #417

Conversation

hackmod commented Dec 27, 2018 • edited

Note

duyk16 commented Dec 27, 2018

hackmod commented Jan 14, 2019

aemet93 commented Jan 22, 2019 • edited

hackmod commented Feb 10, 2019

Ingham27 commented Feb 22, 2019

sammy007 commented Feb 23, 2019 • edited

mikeyb commented Feb 23, 2019

sammy007 commented Feb 23, 2019

neptunix commented Mar 31, 2019

hackmod commented Mar 31, 2019

neptunix commented Apr 1, 2019 • edited

hackmod commented Apr 1, 2019

neptunix commented Apr 2, 2019

hackmod commented Apr 2, 2019 • edited

neptunix commented Apr 2, 2019

StayClassic commented Jun 20, 2019

aemet93 commented Jun 20, 2019 • edited

korjavin commented Jun 21, 2019

aemet93 commented Jun 21, 2019

aemet93 commented Jun 24, 2019

hackmod commented Jun 25, 2019

hackmod commented Dec 27, 2018 •

edited

aemet93 commented Jan 22, 2019 •

edited

sammy007 commented Feb 23, 2019 •

edited

neptunix commented Apr 1, 2019 •

edited

hackmod commented Apr 2, 2019 •

edited

aemet93 commented Jun 20, 2019 •

edited