#259 Spectranet allowing retriggered nmi

v1.1
closed-fixed
None
5
2013-05-11
2012-05-24
Anonymous
No

The spectranet responds to the NMI event and pages itself in. In the hardware a flip flop is set preventing nmi being triggered again until RETN has been executed.
Fuse doesn't emulate this so multiple NMIs can be generated resulting in a crash.

Discussion

  • Philip Kendall

    Philip Kendall - 2013-02-10
    • milestone: --> v1.1
     
  • Philip Kendall

    Philip Kendall - 2013-02-26
    • assigned_to: Philip Kendall
     
  • Philip Kendall

    Philip Kendall - 2013-02-27

    Do we have a way to get a reproducible crash (or other bad behaviour) for this issue? Fixing it shouldn't be too hard, but I'd like to know I've fixed it rather than just guessing :-)

     
  • Sergio Baldoví

    Sergio Baldoví - 2013-03-03

    It appears to be reproducible with these steps:
    1) NMI
    2) NMI
    3) Select "[E] Exit"
    4) Machine crashes

     
  • Philip Kendall

    Philip Kendall - 2013-03-04

    Thanks. Not sure how I missed that!

     
  • Philip Kendall

    Philip Kendall - 2013-03-08

    Fixed in r4904 (libspectrum) and r4905 (Fuse). This does add an extra flag to the SZX Spectranet block, which I'll talk to Jon Needle about.

     
  • Philip Kendall

    Philip Kendall - 2013-03-08
    • status: open --> pending-fixed
     
  • Sergio Baldoví

    Sergio Baldoví - 2013-03-29

    Some stubbed routines are needed to allow disabling Spectranet by --disable-sockets flag on configure.

     
  • Fredrick Meunier

    • status: pending-fixed --> closed-fixed
     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks