|Anonymous | Login | Signup for a new account||2020-09-24 01:22 UTC|
|Main | My View | View Issues | Roadmap | Repositories|
|View Issue Details|
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0000717||EMStudio||Serial Communications||public||2012-09-24 11:49||2013-09-15 06:08|
|Target Version||0.0.1||Fixed in Version|
|Summary||0000717: Multiple tables appear after ECU interrogation|
|Description||After installation of jenkins-EMStudio-Windows-Mingw-Unstable-8, git hash fa5aaf340fbef198d43c68a3e39b18356c9bd605 and with the default Windows7 64bit USB Receive and Transmit Transfer Sizes of 4096 bytes, it takes a long time to interrogate the ECU and after it finishes it displays duplicate VE, Lambda, Timing... etc tables.|
If I change the USB Receive and Transmit Transfer Sizes to 64 bytes as recommended by http://forum.diyefi.org/viewtopic.php?f=12&t=1844 [^] the ECU interrogation is much faster and there are no duplicate tables displayed.
I have attached a screenshot showing the duplicate tables.
|Steps To Reproduce||If the USB Receive and Transmit Transfer Size are set to the Windows 7 default of 4096 bytes the duplicate tables are shown.|
If the USB Receive and Transmit Transfer Size are set to the recommended setting of 64 bytes the duplicate tables are not shown.
|Additional Information||This is on a Windows 7 64-bit notebook.|
|Tags||No tags attached.|
|Attached Files|| EMStudio-Windows7.png [^] (144,355 bytes) 2012-09-24 11:49
Per #freeems IRC conversation:
(11:48:53) Fred: there are two issues
(11:49:01) Fred: 1) duplicate IDs = fuck off
(11:49:14) Fred: 2) broken packet management/routing system
|It looks like what's happening, is I'm sending a read request, timing out, sending a second one, then BOTH responses are coming back. So, as Fred mentioned, I need to handle duplicate packets, and fix how packets are handled a bit.|
Fixed in 59da9c8fefaa5b16d897737b6fccac4d2401407c,
I see this is resolved. Are both fixes in place? IIRC it needed each of these:
1) proper packet routing service that remembers what's been asked for and handles retries etc.
2) error throwing protection on trying to add duplicate locations to the list
If two was done first you could verify it by getting Andy to try it, and see errors dumped out or popped up. Then one could be done and those would go away, then you'd know it was solid.
|Bump fore response to questions.|
Just reviewed the above commit and it shows no sign of resolving the core underlying issues at work here at all. :-(
Proper packet routing/accounting and retries is required. This belongs at a packet level a manager not high up.
A debug line on the console is not sufficient for duplicate IDs, it should flash neon lights.
|To agree with that, though I didn't notice duplicates, the requirement for Tx/Rx transfer size changes is still required in builds as recent as gbffce31.|
|Copyright © 2000 - 2011 MantisBT Group|