09-06-2018 03:48 PM - edited 11-13-2018 08:52 AM
I'm using Xilinx 2018.2 and it just will randomly exit without any error messages. I'm using it on Windows 10, x64. It's a pretty new Windows installation and there's plenty of available memory. I noticed that in workspace/metadata/.log, it has this as the last entry, and I'm assuming that it's the cause. Is there a work around for this? I've tried running it both with admin and non-admin privileges.
!ENTRY org.eclipse.cdt.core 4 0 2018-09-06 15:43:36.294 !MESSAGE Error !STACK 0 java.lang.NullPointerException at org.eclipse.cdt.internal.core.ConsoleOutputSniffer.processLine(ConsoleOutputSniffer.java:178) at org.eclipse.cdt.internal.core.ConsoleOutputSniffer.access$0(ConsoleOutputSniffer.java:174) at org.eclipse.cdt.internal.core.ConsoleOutputSniffer$ConsoleOutputStream.checkLine(ConsoleOutputSniffer.java:99) at org.eclipse.cdt.internal.core.ConsoleOutputSniffer$ConsoleOutputStream.write(ConsoleOutputSniffer.java:58) at java.io.OutputStream.write(OutputStream.java:75) at org.eclipse.cdt.core.CommandLauncher.printCommandLine(CommandLauncher.java:287) at org.eclipse.cdt.core.CommandLauncher.waitAndRead(CommandLauncher.java:250) at org.eclipse.cdt.internal.core.BuildRunnerHelper.build(BuildRunnerHelper.java:273) at org.eclipse.cdt.make.core.MakeBuilder.invokeMake(MakeBuilder.java:219) at org.eclipse.cdt.make.core.MakeBuilder.build(MakeBuilder.java:108) at org.eclipse.core.internal.events.BuildManager$2.run(BuildManager.java:735) at org.eclipse.core.runtime.SafeRunner.run(SafeRunner.java:42) at org.eclipse.core.internal.events.BuildManager.basicBuild(BuildManager.java:206) at org.eclipse.core.internal.events.BuildManager.basicBuild(BuildManager.java:246) at org.eclipse.core.internal.events.BuildManager$1.run(BuildManager.java:301) at org.eclipse.core.runtime.SafeRunner.run(SafeRunner.java:42) at org.eclipse.core.internal.events.BuildManager.basicBuild(BuildManager.java:304) at org.eclipse.core.internal.events.BuildManager.basicBuildLoop(BuildManager.java:360) at org.eclipse.core.internal.events.BuildManager.build(BuildManager.java:383) at org.eclipse.core.internal.events.AutoBuildJob.doBuild(AutoBuildJob.java:144) at org.eclipse.core.internal.events.AutoBuildJob.run(AutoBuildJob.java:235) at org.eclipse.core.internal.jobs.Worker.run(Worker.java:55)
Edit: I resolved the issue by installing Xilinx SDK to a Linux distro since the Windows version is unusable.
09-11-2018 03:12 AM
09-11-2018 10:59 PM
09-13-2018 11:20 AM
Which .Xilinx folder are you referring too?
I'm experiencing the quiet crash when I change the serial_flash_family value from 1 to 5 and then click OK.
It never used to happen, until I changed to version of xilisf from 5.8 to 5.11.
SDK 2018.2, Vivado 2018.2, no other installations of these tools.
09-25-2018 02:02 AM
I have same problem with SDK 2018.1, 2018.2, 2018.2.1 with different PC with windows 10-x64.
A build of "C" project from scratch works fine. In case of remove file/folder from workspace the automatic rebuild cause crash. After crash the SDK cannot restart because workspace is corrupted. the workaround is (1) remove SDK workspace folder, (2) start SDK and disable automatic rebuild (3) import projects from current folder (4) execute build manually. In case of crash, the SDK can restart because automatic build is disabled.
This procedure is quite boring and finally I decided to uninstall Vivado 2018.x and use 2017.4.
09-25-2018 02:14 AM
As i wrote in another message i have exact the same problem. In my case after a crahs when i restart SDK it comes up
with an Java-Error, when i click "Restart" it can continue working. But as you said it is extremly boring !
09-25-2018 04:12 PM
Exactly same story : SDK 2018.2 crashes too frequently on new laptop Win10
But on old laptop Win7 it NEVER NEVER NEVER crashes.
Maybe Windows 10 related ?
09-27-2018 12:28 AM
Same thing here, the SDK 2018.2 exits systematically (without any message) when I try to delete a project from a workspace, or if I try to delete a file from a project. I can restart the SDK but the delete operation was not done, so it's pratically impossible to delete any file or project. I'm on Windows 10 64-bit.
10-01-2018 09:05 PM - edited 10-01-2018 09:44 PM
Same problem here - and it effectively make the tool impossible to use!
Just found a walk-around that worked so far:
Deselect "Build Automatically" under "Project" as soon as you start SDK after exporting hardware and before creating any C project using wizard.
10-02-2018 02:39 AM
I've also disabled the "Build Automatically" feature, but it makes no difference. I've also some crashes when adding files to a project, or when lauching the debugger...
10-04-2018 06:47 AM
Same here. Crashes quietly (just "poof!"). No error messages...
Also running Windows 10 64-bit. Mine crashes when I attempt to create a new application (e.g. a new "Hello World"), or when I delete a previous application and the BSP associated with it.
Not sure if it's relevant, but the SDK project I'm using (and having trouble with) was created in a previous version (2015.4) and upgraded to 2018.2.
10-04-2018 11:51 AM
Not sure if this helps, but it seems to be an Eclipse access violation issue with MSCTF.DLL:
Faulting application name: eclipse.exe, version: 0.0.0.0, time stamp: 0x573c7dbf
Faulting module name: MSCTF.dll, version: 10.0.17134.285, time stamp: 0x7cd940c6
Exception code: 0xc0000005
Fault offset: 0x000000000001ebf9
Faulting process id: 0x10f0
Faulting application start time: 0x01d45c0dfa2fb4bc
Faulting application path: C:\Xilinx\SDK\2018.2\eclipse\win64.o\eclipse.exe
Faulting module path: C:\WINDOWS\System32\MSCTF.dll
Report Id: f92e09dc-40dd-429f-a2fd-5907fdf0a53c
10-10-2018 06:25 PM
10-11-2018 02:28 AM
This might be due to the windows periodic updates are not applied to the PCs. At our work also we have never experienced any crash on Windows 10. Please do apply the windows updates and let us know if you still see the crash.
10-12-2018 03:49 AM
I've all windows updates applyied, and it still crashes regulary.
I can easilly reproduce the problem by deleting a project in a workspace: the crash is systematic. Let me know if you need more information that would help you to reproduce and solve that problem.
10-13-2018 10:51 AM
I've been seeing this as well since installing 2018.2. I'm on a Windows10 x64 VMware virtual machine with 5 GB RAM. I already usually disable "Build Automatically". I see the issue about 50% of the time when adding an application project (File | New | Application Project). It also happened just now when I asked it to regenerate BSP after I changed Repositories to point to a different local FreeRTOS customized repo.
FYI, I have done a lot of Eclipse-based development and usually keep my workspaces in a location under C:\Users\Jsam\xil_wksp\001\ 002, 003, and so forth. Eclipse workspaces in my judgment are very lightweight, are "easy-come, easy-go" and should never be checked in to version control and should be kept away from the real project files (.project, .cproject) and source code. This was a very successful policy for me when using TI CodeComposerStudio, which is very much like Xilinx SDK.
10-15-2018 09:56 PM
Same issue here.
The 2018.2 Vivado SDK (and only the SDK) will crash quietly when creating new a project or renaming a file in a project. I wonder if the issue is related to accessing the file system specifically on Windows 10.
10-16-2018 12:45 PM
Some additional observations in case this helps the Xilinx development team.
On Windows 10, I see this crash much more readily if I have opened a build log in Wordpad to view results and not closed Wordpad before making project changes or building again.
In another case, I believe a project Clean... resulted in SDK shutting down because Windows Explorer was displaying the workspace directory where the build log files reside (thus locking them from deletion?).
Perhaps SDK is not gracefully handling file access errors thrown up by Windows 10?
10-24-2018 11:03 PM
I got the same problem, already with 2018.1.
I am running Xilinx tools on windows 10 PC.
Unfortunately not only the SDK crash - Vivado too.
Please your support to solve this!!
10-29-2018 06:31 AM
I have further "empirical" evidence that, on my system, the "quiet crashing" of SDK is related to limited Windows memory.
I run SDK on a Windows 10 x64 VMware virtual machine with only 5 GB RAM. My laptop has only 8 GB total, so 5 is all I can allocate to the VM.
I had been leaving the Microsoft Edge browser open always so I can Google for help or search the Xilinx Forums. With that browser open, I get the SDK crashing multiple times in a day and, in some cases, my workspace and project files are irreparably corrupted.
Fortunately, I backup my projects very regularly and also know how to recreate from scratch if necessary. And Eclipse workspaces are super easy to abandon and create new ones.
I have now adopted a policy of not letting the web browser be open while I have SDK open. Not ideal, but the time lost after crashes and corruptions was much more than the time to close one app and open the other.
KEY INSIGHT: I had been following the "no browser" policy for a full week with no SDK crashes. Yesterday, I forgot to close the browser and was running SDK when it crashed for the first time in many days. And yup, I realized I had left the browser open.
If SDK is crashing due to inability to allocate RAM for something, I can understand that, but it would be nice if it could be more graceful and not corrupt its own project and workspace files. That may not be exactly what is happening, but this theory fits my observations.
10-29-2018 08:25 AM
New PC I9-7900X 3.3GHz 10 cores (20 logical cpu), 64 GB DDR. Windows 10x64 version 10.0.17134 Build 17134.
All updates from Microsoft applied.
Same problem with SDK 2018.2 (Vivado 2018.2.2). No problem with SDK 2017.4.1
Waiting for solution from Xilinx I suggest CentOS 7.5 x86_64. I installed on VMware workstation 14 and verified the following tools:
10-29-2018 10:27 PM - edited 10-29-2018 10:29 PM
Unfortunately we are not able to reproduce the problem on Windows 10
For windows 10, while installing the SDK, you can see the supported builds (1709 and 1803). Screenshot is attached.
If you see any file in your bin directory with hs_err* files, please supply them. We will look into the files to find out the Java crash where it is actually crashing
To know the build of your machine, click on start, type "About your PC" in the search and click on "About your PC" and see the build
10-30-2018 02:00 AM
There are a lot of "bin" folders. Which one do you mean ?
I cannot find any hs_err... files.
And here is my WinVersion:
10-30-2018 08:16 AM
Sorry for the confusion. The bin folder of SDK installation.
10-30-2018 08:22 AM
In the meantime i have searched my computer for a hs_err.. file but there is none.
What i have i a .log file with a lot of entries like this:
!ENTRY org.eclipse.cdt.core 4 0 2018-10-30 14:42:26.340
Does this help ? ( i can send you the complete log if you want )...
10-30-2018 08:28 AM
Please send the log file.
11-01-2018 09:02 AM
Could you please let us know whether you tried the following to avoid this crash problem ?
1. Delete the .Xilinx folder from your c:\users\<yourUserName>\.Xilinx
2. Launch xsdk.bat from command line (c:\<InstallFolder>\SDK\2018.2\bin\xsdk.bat) and see any errors in the command prompt