[comp.sys.ibm.pc] Intercepting parity errors

jjoshua@topaz.rutgers.edu (Jon Joshua) (08/08/88)

Would someone out there please point me towards a program that will
intercept parity errors and take care of them?   I have a few TSR
programs running in the background along with an extended memory
board.  Occasionally I will get parity errors.  The machine will tell
me this by writing an error message to the screen.  Most of the time I
just need to hit esc and the machine goes back to doing what it was
doing.  The funny thing is that I can get the parity errors to repeat
by running the same thing.  I was once spell-checkng a document and it
continuously blew up on the word 'VAX'.  


JOn.
-- 
 _____________________________________________________________________________
|o|   Jonathan Joshua    jjoshua@topaz.rutgers.edu          Why not...      |o|
|o|                                                   Nothing better to do  |o|
|o|                 Anything is possible,                                   |o|
\o|                     except for skiing through a revolving door.         |o|
  \_  ____/ \                                                               |o|
    \/       \       _______________________________  ______________________|o|
              \___  /                               \/
                  \/

johnson@c10sd1.StPaul.NCR.COM (Wayne D. T. Johnson) (08/10/88)

In article <Aug.7.15.01.17.1988.3125@topaz.rutgers.edu> jjoshua@topaz.rutgers.edu (Jon Joshua) writes:
>
>Would someone out there please point me towards a program that will
>intercept parity errors and take care of them?   I have a few TSR
>programs running in the background along with an extended memory
>board.  Occasionally I will get parity errors.  
>        The funny thing is that I can get the parity errors to repeat
>by running the same thing.  I was once spell-checkng a document and it
>continuously blew up on the word 'VAX'.  
 
On my old PC, the NMI (Non-Maskable Interupt) was used to signal parity
errors.  Occasionaly some of our software would accidently load and branch
to int vector 2 (the NMI vector).  It LOOKED like a parity error which put
the error "ERROR ON EXPANSION CARD" on the screen and halt the machine.  Not
nice.

We fixed the software and solved the problem.  Is it possible that some of
your software (maybe the TSRs) is a bit buggy?

To trap your errors, you might (no, I havn't tried this) change your 2nd
vector to the DOS ignore interrupt routine (the address pointed to by the 
unused interrupts).

For those who do not know what interrupt vectors are, they are 4 byte 
addresses located starting at address 0:0, 0:4, 0:8 (this is the second one),...
-- 
Wayne Johnson                 (Voice) 612-638-7665
NCR Comten, Inc.             (E-MAIL) W.Johnson@StPaul.NCR.COM or
Roseville MN 55113                    johnson@c10sd1.StPaul.NCR.COM
These opinions (or spelling) do not necessarily reflect those of NCR Comten.