Mario. wrote: > Heisenbug. Surely ;-) > Okay, I had similar problem in assembly, but that was probably > hardware... I included a single NOP instruction which solved a lot of > problems in a MCU by moving the rest of the program one instruction > away. But one thing the disabled debug messages surely do is that it > speeds up the program a bit and potantially makes it much more > responsive due to cache size - when there is no debug, it cannot > possibly take the valuable CPU cache. Well, this shouldn't affect the operation of a correct program. John Kasunich and I went over the relevant code last night and we couldn't find any smoking guns in there. This driver reads all the inputs at the beginning, and then writes all the outputs at the end, so changes in the timing shouldn't affect anything. I'm more afraid there is some very dark monster lurking, ie. that something is address sensitive, and adding the debug statements moves some piece of data to a different place, where a haywire pointer can't find it. These can be the devil to find, even in user-space C code, and an absolute nightmare in kernel space.
Jon ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
