Memory Browser window issue

[lpcware]

Content originally posted in LPCWare by robekras on Fri Mar 07 02:20:53 MST 2014
Memory Browser window shows question marks instead of real data.

Steps to reproduce:

Program is running.
Go to Memory Browser window and try to show memory area of a specific area e.g. 0x20000000 (I have a lpc1778).
As the LPCXpresso can't access the memory, because target program is running, it shows only question marks.
If I suspend/break the target program I would expect the memory window should be filled with real data.
But that will not happen. If I scroll up or down the window, the memory before (or after) is shown correctly, but never the area which is initialy filled with '????'.
When I open the Memory Browser, when target program is suspended, the area is correctly shown. Then when I start the target program and I enlarge the Memory Browser window also (of course) the new memory area which is comming into sight is filled with '????' and will never be updated by real data.

Scrolling up and down, clicking on refresh (context menu), closing and reopening the memory tab does not help.

If I go to the Memory Monitor window I also see only question marks for that area.

This behaviour is also true for previous versions (at least for 6.1.4_194) of LPCExpresso.

Robert

[lpcware]

Content originally posted in LPCWare by robekras on Mon Mar 10 23:32:35 MST 2014

According 2.
It is not related to changes I made, but changes made by the running program. I start the target program, and on the next break/stop/suspend I see changed memory locations, but not highlighted in red.

More strange, sometimes it works und more often it doesn't. Yesterday I had changed memory values in one locations, they were highlighted in red and there was a frame drawed around every changed value. When I scrolled down (or up) to other memory locations where values were changed between running and stopping of program, I see the changes in values but no highlighting.
Going back to the original location the highlighting works when I had changes (between running and stopping program).

Another strange issue I observed last week: I changed a memory location by typing a new value (changes to green), hit Enter to 'enter' the change. It changed to red, and this red did not go after running and stopping the program.

Robert

[lpcware]

Content originally posted in LPCWare by lpcxpresso-support on Mon Mar 10 09:52:00 MST 2014
1. Entering Debug mode does not stop DMA. If the memory is changed (by DMA, or some other external event) after the debugger has read the memory, then it will not be re-read

2. When changing memory, you have to hit Enter/Return to 'enter' you changes (or Esc to cancel your changes). If you do not hit Enter, the values are not changed. This seems to describe your circumstances

[lpcware]

Content originally posted in LPCWare by robekras on Mon Mar 10 07:36:52 MST 2014
And one issue in addition :-).

If values in Browser Memory window changes I expect that the changed values are shown in red.
In fact sometimes some memory locations are shown in red, but sometimes (or very often?) changed locations are not highlighted in red.

I added some screenshots:
The changes are in the line with address 0x20005c1c, column 9, 10 and 14.

Robert

[lpcware]

Content originally posted in LPCWare by robekras on Mon Mar 10 07:23:53 MST 2014
Thanks for the info.

The second issue related the lpc_entdata.
Seems to be related to DMA!?
DMA is working although processor is suspended.
As the memory window is read before (or after) the watch expression the displayed data differs.
I did a single step, and the data in memory browser window were updated to the same as in expression window.
Which in next step leads to the suspicion that the refresh in memory browser window doesn't work.

Robert

[lpcware]

Content originally posted in LPCWare by lpcxpresso-support on Fri Mar 07 10:32:31 MST 2014
We have fixed the first issue (displaying '??' always). This was caused by a longstanding bug in Eclipse/CDT (Bug 388142, if you are interested).

The other issues you have raised are still being investigated.

This fix will be made available soon.

[lpcware]

Content originally posted in LPCWare by robekras on Fri Mar 07 07:50:53 MST 2014
And some more issues related to the Memory Browser (and Memory) window:

Please take a look at the screenshot of the watch expression window.
The structure lpc_enetdata is located at address 0x20007620. The single variables within prxs[0...5] shows some reasonable values.
But the Memory windows (also the Memory Browser window) shows all 0xff.
I did it more than once. Sometimes I get the same value on some variables, but not on all.
Same strange behaviour on 6.1.4 and 7.0.0.

Robert