net: stop wolfSSL's CloseSocket macro from turning P2P closes into double closes

wolfssl/wolfio.h (included by <wolfssl/ssl.h> in net.h) does
`#ifndef CloseSocket / #define CloseSocket(s) close(s)`. Every TU that
includes net.h after that point compiles CloseSocket(x) to a bare
close(x). That includes all 15 calls in net.cpp, among them the ones
in CloseSocketDisconnect and ~CNode. util/sock.cpp's
CloseSocket(SOCKET&), the only thing that sets the handle to
INVALID_SOCKET, was never called from net.cpp. This has been the case
since OpenSSL was replaced by wolfSSL (Hush 9ff756590, 2020), so every
DragonX release is affected.

So CNode::hSocket kept its number after the socket was closed. Every
P2P disconnect closed that same number again, 2 to 4 times:
- CloseSocketDisconnect (net.cpp:572) runs once per disconnect trigger
  (ThreadSocketHandler, SocketSendData, the TLS socket handler,
  ThreadMessageHandler), and each run still saw a valid-looking hSocket;
- ~CNode (net.cpp:2819) then closed it once more.
Meanwhile the socket handler kept FD_SETting, send()ing and recv()ing on
the stale number. In between, other threads get that number from the
kernel's lowest-free-fd rule:
- BDB reopening wallet.dat or database/log.* (ThreadFlushWalletDB, ~11/s);
- fopen("rb+") of a block file for getdata/getblock;
- an HTTP RPC accept.
The stray close then closes it. BDB keeps doing I/O on the dead number:
'wallet.dat: write failed for page N' / 'Illegal seek' / 'BDB0060
PANIC', or silent wallet-page writes into a blocks/blk*.dat that got
the number next. Its failing reads later threw out of ThreadNotifyWallets
(std::terminate). A closed HTTP connection fd makes libevent re-arm
and later close whatever owns that number next ('Epoll ADD ... failed').

Reproduced on isolated regtest with an LD_PRELOAD close() guard:
- one disconnectnode gives exactly one EBADF double close (49
  disconnects, 50 EBADF; with strace and no shim, 29 disconnects gave
  30 'close(N) = -1 EBADF');
- P2P churn alone gives non-owner closes of block files and wallet.dat;
- RPC-only and Tor-only loads give none.
Disassembly of CloseSocketDisconnect and ~CNode shows 'call close@plt'
before this change and 'call CloseSocket(unsigned int&)' after.

Undefine the macro right after the wolfSSL include in net.h; wolfio.h
is include-guarded, so it can't come back. net.cpp gets an #error
tripwire in case it ever does. NetCloseSocket.ClosedNodeNeverClosesAReusedFdNumber
fails on the old code (a closed CNode closes the fd number's new owner)
and passes now.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

src/net.cpp

@@ -45,6 +45,13 @@
 #include <wolfssl/options.h>
 #include <wolfssl/ssl.h>
 #include <hush/tlsmanager.cpp>
+
+// CloseSocket must be util/sock.cpp's CloseSocket(SOCKET&), which invalidates the handle it closes. If
+// wolfSSL's `#define CloseSocket(s) close(s)` survived into this file (see net.h), each close here would leave
+// its stale fd number behind to be closed again after another thread reused it.
+#ifdef CloseSocket
+#error "CloseSocket is a macro here (wolfssl/wolfio.h): it would bypass CloseSocket(SOCKET&) in util/sock.cpp"
+#endif
 using namespace hush;
 extern int32_t HUSH_TESTNODE;
 extern uint8_t ASSETCHAINS_CLEARNET;