Horst Birthelmer wrote:


On Nov 19, 2004, at 2:34 PM, Jim Rees wrote:

Is there any reason we shouldn't be using varargs now? I can't believe any
of our supported platforms wouldn't have it.


Is there any reason for the "needlock" parameter to be the last one??
...other than compatibility??

That's basically what Christof did ... If we change both things we might be a lot safer.

If you really wanted to make it safer, all the functions should have prototypes in header files. (and have the compiler enforce it.)

This may be an compile optimazation problem. One way to test is to
compile the specific routine but add the -S option so the input  to the
assembler is created, then look at what is being optimized. I tried
this with gcc on sun4x_59 and I don't see anything wrong but on the platform
that is failinig, there may be a problem.

It could also be a problem with the *aproc routine that gets called
and passed these paramters.

To find out, someone need to pin down the line of code doing the overlay,
not just the routine in which it occured.





Horst

_______________________________________________
OpenAFS-devel mailing list
[EMAIL PROTECTED]
https://lists.openafs.org/mailman/listinfo/openafs-devel




--

 Douglas E. Engert  <[EMAIL PROTECTED]>
 Argonne National Laboratory
 9700 South Cass Avenue
 Argonne, Illinois  60439
 (630) 252-5444
_______________________________________________
OpenAFS-devel mailing list
[EMAIL PROTECTED]
https://lists.openafs.org/mailman/listinfo/openafs-devel

Reply via email to