Ticket #13045 (closed defect: fixed)
VirtualBox Debugger telnet console issues => Fixed in SVN
Reported by: | mhanor | Owned by: | |
---|---|---|---|
Component: | VMM | Version: | VirtualBox 4.3.12 |
Keywords: | Cc: | ||
Guest type: | all | Host type: | Windows |
Description
Host: Windows 7 SP1 x64
Issue 1:
A telnet client that doesn't do line buffering, meaning that it sends each character as it's typed (e.g. the Windows telnet client), doesn't work with the telnet console of the VirtualBox debugger. Although I don't know if it has any other impact, the GUI console and the telnet console (buffered or not) seem to work fine after commenting out the else branch, of the main "if" statement, inside the dbgcProcessInput() function, vbox\src\VBox\Debugger\DBGConsole.cpp, lines 566,568:
else /* Received nonsense; just skip it. */ pDbgc->iRead = pDbgc->iWrite;
Issue 2:
When the telnet client forcefully exits, the server's socket remains in TIME_WAIT status, while the VirtualBox console thread hogs one CPU core, caught up in the do-while loop, inside the dbgcInputRead() function (same source file as above). The reading from the socket returns success, with zero bytes read, because the RTSocketRead() function doesn't do anything about this situation. Also, the do-while condition is met (RTSocketSelectOne()), inside dbgcInputRead(), so there's no exit condition. Setting rc to VERR_NET_SHUTDOWN, if the read was zero bytes, corrects this problem.
Change History
comment:2 Changed 8 years ago by vushakov
- Summary changed from VirtualBox Debugger telnet console issues to VirtualBox Debugger telnet console issues => Fixed in SVN
This should be fixed as suggested in the forthcoming 5.0.2. Will be backported to 4.3 so that the next 4.3 dot-dot release has it too.
Both problems are still occurring, even in VirtualBox 5.0 RC1.