quartus_map.exe Exception: Access Violation

[Altera_Forum]

Moin, 

 Returning back from my QII7.1 Experience to QII6.1, i'm faced with new "fun". I've a simple design here, consisting of an SOPC (containing NIOS-cpu,pll,ddr-sdram,jtag-uart,uart, tristate bridge, cfi-flash and lan91c111); running on the CyclonII-NIOS-Eval Board.  It already did compile and i was able to run (e.g. Hello World-)Software on the design. Suddenly, (i have no idea, if it had to do with my previous installed QII7.1) Quartus crashes, when i compile. Running SOPC-Builder seems to be ok but right after pressing the compile button in Quartus this happens: *** Fatal Error: Module: quartus_map.exe Exception: Access Violation 0x132E8 : qcu_read_acf_assignment + 0x107C (comp_qcu) 0xA936 : (quartus_map) 0x94EF : qexe_get_command_line + 0xE1 (comp_qexe) 0xB362 : qexe_run_tcl_option + 0x4E2 (comp_qexe) 0xB3AF : qexe_standard_main + 0x14 (comp_qexe) 0x2A32B : msg_exe_main + 0xDB (CCL_MSG) 0x9D4D : (quartus_map) 0x16FD6 : RegisterWaitForInputIdle + 0x48 (kernel32) 0x83C5FFFF: MEM_MANAGER::s_quartus_memory_counter + 0xFFDBEFC2 (MSVCR71) End-trace  Asking Google took me to the advice to run manually the "remove_add_constraints_for_ddr_sdram_0.tcl" script from QII. This script ran without error but didn't change anything. Error still comes right after pressing the compile button.  OOps - i almost forgot the line in the "Processing" window: Info: Command: quartus_map --read_settings_files=on --write_settings_files=off ov70eval -c ov70eval --analysis_and_elaboration  (I get the same result, if i start this command from a cygwin shell.)   Anybody got any further advice for me?  Cheers, WK

[Altera_Forum]

Fatal Errors are the worst, since there's absolutely nothing to point you in the right direction to fix it. Did this occur in Q6.1 or Q7.1? I thought I saw something similar to your issue, but it was in Q6.1 and fixed in Q7.1. One other last ditch thing to try as an experiment, can you close Quartus, delete the db, and recompile? Shouldn't have an affect unless incremental compilation or something is reusing the db, but just a shot in the dark.

[Altera_Forum]

Someone else had the very same error messages for a problem that was fixed in version 7.1. It's possible that derguteweka has a different problem causing these same messages, but most likely version 7.1 fixed it. 

 It looks like derguteweka is going to have to choose between problems specific to 6.1 and problems that are new in 7.1. I recommend using 7.1 SP1, working through any problems with that, rather than dealing with old problems or keeping a design in the old SOPC Builder version for a long time.  If the 7.1 problems in the http://www.alteraforum.com/forum/showthread.php?t=664 thread could have been caused by starting an SOPC Builder design in 6.1 and then upgrading to 7.1, then it might be worth the trouble to re-create the SOPC Builder system from scratch in 7.1 SP1 even though derguteweka had trouble with the first 7.1 attempt. Using 7.1 SP1 means you won't have to deal with any bugs like this fatal error that are supposed to be fixed already. You will still have to deal with any new 7.1 bugs that weren't fixed in SP1--the patch 0.20 mentioned in the other thread might still be needed if the design has a tri-state bridge. But using 7.1 SP1 also means that your design will be easier to maintain than it would be if you kept it in 6.1, which is before the major SOPC Builder redesign that was first released in 7.1. And the redesign in 7.1 provides major benefits for some designs.  If trying 7.1 SP1 again, check the version 7.1 (not 7.1 SP1) release notes for possible issues that I suspect are related to the SOPC Builder redesign in 7.1. The release notes have an "SOPC Builder Issues" section. (The release notes are at http://www.altera.com/literature/lit-qts.jsp.)  I suggested re-creating the SOPC Builder system from scratch if the 7.1 problems could have been caused by changing Quartus versions for a system created in 6.1, but you normally should not have to re-create the system. There is, however, something you should do one time when you first open the system in 7.1. I didn't find official documentation for that, but there is some mention of it at http://www.alteraforum.com/forum/showthread.php?p=2204#post2204.

[Altera_Forum]

Moin, 

 OK, Ryscs shot in the dark hit the target - in other words: When i delete the db/ directory, im able to compile my design again. It's also running on Hardware, (=i can download & run Software on the NIOS). If i open the SOPC again and rebuild it, even with no changes i'll get the same Access Violation again during the compile of the whole system. If i delete db/ before compiling, everything will work fine again. This behaviour is just on one particular project, on others in QII6.1 everything is ok (so far). I assume, something in an input file of SOPC got broken, which causes something in the db/- director to be lethal for quartus_map.exe if SOPC was running before.  I'm pretty lucky, that i'm currently doing just the first steps and so the time i waste with this kind of effects is just within days, not months :rolleyes: and also switching the QII-Version is still possible. But i want to keep this time as short as possible and start concentrating of the specific problems of my design - there'll be more than enough of it, but a stable Designsoftware is mandatory first.  In the past, i've never dared to migrate a project from one Version of quartus to another - i've yet faced enough problems without doing so :) - so i also won't do in the future.  I'll try to install 7.1SP1 - it'll take some days because of an evil combination between the huge files (no standalone SP1 available for Web Edition) to download, our firewall at work (eating up large downloads), slow download at home and a to small USB-Stick for passing the file(s) unsplit. :) But i've no good feeling at all with this, when i have a look e.g. at here: http://www.alteraforum.com/forum/showthread.php?t=682 In my design there will be a few (2-4) self-made SOPC components, so i'll have to expect similar effects :( At least those components do not yet exist outside of my brain, so i might still hope things won't get to ugly because components are done from scratch in the same QII Version. I've already faced the same problem (...Pausing target processor: not responding....) with 7.1 as rjackal does with 7.1SP1, which caused me to step back to 6.1. and in opposite to other people, i'm not a huge supporter of java language - there are just to many examples, where introducing java resulted in huge, slow and bad behaving software.  Cheers WK