Potential deadlock between main and net #9215

fanquake commented at 10:31 AM on November 24, 2016: member

Running current master https://github.com/bitcoin/bitcoin/commit/93566e0c37c5ae104095474fea89f00dcb40f551

2016-11-24 10:23:42 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:42 Previous lock order was:
2016-11-24 10:23:42  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:42  (1) cs_main  main.cpp:5971
2016-11-24 10:23:42  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:42 Current lock order is:
2016-11-24 10:23:42  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:42  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:42  (1) cs_main  main.cpp:1785
2016-11-24 10:23:42 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:42 Previous lock order was:
2016-11-24 10:23:42  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:42  (1) cs_main  main.cpp:5601
2016-11-24 10:23:42  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:42 Current lock order is:
2016-11-24 10:23:42  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:42  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:42  (1) cs_main  main.cpp:1785
2016-11-24 10:23:43 Imported mempool transactions from disk: 392 successes, 456 failed, 0 expired
2016-11-24 10:23:43 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:43 Previous lock order was:
2016-11-24 10:23:43  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:5601
2016-11-24 10:23:43  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:43 Current lock order is:
2016-11-24 10:23:43  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:1785
2016-11-24 10:23:43 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:43 Previous lock order was:
2016-11-24 10:23:43  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:5601
2016-11-24 10:23:43  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:43 Current lock order is:
2016-11-24 10:23:43  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:1785
2016-11-24 10:23:43 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:43 Previous lock order was:
2016-11-24 10:23:43  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:5601
2016-11-24 10:23:43  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:43 Current lock order is:
2016-11-24 10:23:43  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:43  (1) cs_main  main.cpp:1785
2016-11-24 10:23:44 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:44 Previous lock order was:
2016-11-24 10:23:44  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:44  (1) cs_main  main.cpp:5971
2016-11-24 10:23:44  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:44 Current lock order is:
2016-11-24 10:23:44  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:44  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:44  (1) cs_main  main.cpp:1785
2016-11-24 10:23:44 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:44 Previous lock order was:
2016-11-24 10:23:44  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:44  (1) cs_main  main.cpp:5601
2016-11-24 10:23:44  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:44 Current lock order is:
2016-11-24 10:23:44  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:44  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:44  (1) cs_main  main.cpp:1785
2016-11-24 10:23:46 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:46 Previous lock order was:
2016-11-24 10:23:46  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:46  (1) cs_main  main.cpp:5971
2016-11-24 10:23:46  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:46 Current lock order is:
2016-11-24 10:23:46  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:46  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:46  (1) cs_main  main.cpp:1785
2016-11-24 10:23:48 Pre-allocating up to position 0x4000000 in blk00688.dat
2016-11-24 10:23:49 UpdateTip: new best=0000000000000000000bac2ac98aca7be77696a38eeb4d927851c3e134753e9a height=440332 version=0x20000000 log2_work=85.584191 tx=173286485 date='2016-11-24 09:49:28' progress=0.999990 cache=740.9MiB(780947tx) warning='1 of last 100 blocks have unexpected version'
2016-11-24 10:23:55 POTENTIAL DEADLOCK DETECTED
2016-11-24 10:23:55 Previous lock order was:
2016-11-24 10:23:55  pnode->cs_vRecvMsg  net.cpp:1846 (TRY)
2016-11-24 10:23:55  (1)
 cs_main  main.cpp:4899
2016-11-24 10:23:55  (2) pnode->cs_vSend  net.cpp:2642
2016-11-24 10:23:55 Current lock order is:
2016-11-24 10:23:55  (2) pnode->cs_vSend  net.cpp:1865 (TRY)
2016-11-24 10:23:55  (1) cs_main  main.cpp:6547 (TRY)
2016-11-24 10:23:55  (1) cs_main  main.cpp:1785

rebroad commented at 12:45 PM on November 24, 2016: contributor

also seeing these

jonasschnelli commented at 1:08 PM on November 24, 2016: contributor

IMO these "potential" deadlocks are pretty normal because the way how we detect "potential" deadlocks is not over optimized. In the log above, I can see a problematic potential deadlock.

laanwj added the label P2P on Nov 25, 2016

Sjors commented at 3:28 PM on March 16, 2018: member

@practicalswift does #12665 offer any relief here?

practicalswift commented at 3:51 PM on March 16, 2018: contributor

@Sjors No having #12665 merged wouldn't have helped here since #12665 doesn't change any run-time behaviour. It is strictly a compile-time thing: it tests our explicit run-time locking assumptions (in the form of AssertLockHeld(…)/AssertLockNotHeld(…)) also at compile-time.

practicalswift commented at 3:53 PM on March 16, 2018: contributor

@jonasschnelli Which one is that?

MarcoFalke commented at 2:33 PM on August 6, 2019: member

@fanquake Is this still an issue?

fanquake commented at 12:08 AM on August 7, 2019: member

@MarcoFalke I don't think I've seen this for a while, so can close for now.

fanquake closed this on Aug 7, 2019

MarcoFalke locked this on Dec 16, 2021