Smart trap problem with interrupt driven stdio
Posted: Mon Aug 03, 2009 5:40 pm
Hi All, but mostly the Netburner software gurus.
I haven't tested this extensively, but I had a problem recently where I had an array access past the end of the array, that was causing a reboot with no trap output.
I was closing and re-opening the stdio ports to set them up as interrupt driven. This seems to work fine for everything but the smart trap output.
When I removed the re-open stuff, leaving stdio as polled I/O, smart trap output appeared on the appropriate serial port.
I'm guessing that when a trap occurs the interrupts are disabled. But that doesn't seem like it would cause the LocalByteOut routine in bsp.c to not work, unless they are causing an interrupt on top of trying to do polled I/O which maybe causes a second trap, leading to an immediate reboot.
I'm probably missing something fundamental about how the MCF5270 processors work, so any insight on this would be appreciated.
This is using Rel23_rc7a.
Thanks!
Mike Chadwick
I haven't tested this extensively, but I had a problem recently where I had an array access past the end of the array, that was causing a reboot with no trap output.
I was closing and re-opening the stdio ports to set them up as interrupt driven. This seems to work fine for everything but the smart trap output.
When I removed the re-open stuff, leaving stdio as polled I/O, smart trap output appeared on the appropriate serial port.
I'm guessing that when a trap occurs the interrupts are disabled. But that doesn't seem like it would cause the LocalByteOut routine in bsp.c to not work, unless they are causing an interrupt on top of trying to do polled I/O which maybe causes a second trap, leading to an immediate reboot.
I'm probably missing something fundamental about how the MCF5270 processors work, so any insight on this would be appreciated.
This is using Rel23_rc7a.
Thanks!
Mike Chadwick