In #3997, laanwj wrote:
so it appears that there are some locking inconsistencies in the network code. This is unrelated to the activechain-related locking issues (which only dealt with cs_main and cs_wallet) but worth creating a new issue for
This is that issue.
Here's some debug.log output from a bitcoind (not -qt) that I built from 89bbd54f:
2014-04-24 15:29:24 UpdateTip: new best=00000000000000001fb86ae8e043340da370bb3275a0e4ab4e71c375b3d62f05 height=297495 log2_work=78.191077 tx=37474592 date=2014-04-24 15:29:11 progress=0.999999
2014-04-24 15:29:25 AddToWallet 1b38742b2c3315b9d02d80c9d107e6235205f11ee3d645d0a017c1adedabbeba
2014-04-24 15:29:25 ProcessBlock: ACCEPTED
2014-04-24 15:36:39 ResendWalletTransactions()
2014-04-24 15:36:39 Relaying wtx 9ea94ab3f30363af36a2ec89e47d93285e3612c5e5084111f0d827d2719ba0e8
2014-04-24 15:36:39 POTENTIAL DEADLOCK DETECTED
2014-04-24 15:36:39 Previous lock order was:
2014-04-24 15:36:39 (1) pnode->cs_vSend net.cpp:1556
2014-04-24 15:36:39 cs_main main.cpp:4206
2014-04-24 15:36:39 cs_wallet wallet.cpp:934
2014-04-24 15:36:39 (2) cs_mapRelay net.cpp:1845
2014-04-24 15:36:39 Current lock order is:
2014-04-24 15:36:39 pnode->cs_vRecvMsg net.cpp:1537
2014-04-24 15:36:39 cs_main main.cpp:3246
2014-04-24 15:36:39 (2) cs_mapRelay main.cpp:3326
2014-04-24 15:36:39 (1) cs_vSend net.h:449
2014-04-24 15:45:46 ERROR: AcceptToMemoryPool : nonstandard transaction: dust