Re: [ECOS] Thread messages problem.

[Jonathan Larmour <jlarmour at redhat dot com>]

felixwong@i-technologies.cc wrote:
> 
> I am writing a hard disk device driver for eCos system on the Atmel EB40 board
> with my specific ATA interface hardware.  In debug process, I enabled the
> CYGDBG_IO_INIT keyword for debugging messages in serial port #0  (diag_printf).
> 
> The problem I met was quite annoying, some thread messages was displayed on
> the serial port #0 terminal.  Sometime appear just after 1 or 2 devices initialized,
> another time appear after all devices initialized.  They look like:
> $T05thread:00000001;0f:ac680402;0d:9c680402;#03
> $T05thread:00000001;0f:84680402;0d:68680402;#7d
> on the screen(of terminal program).  The the program cannot proceed to the
> "main" where I set my breakpoint there.

That shouldn't happen. That implies the stub stopped, e.g. because it hit a
breakpoint. It's intriguing that the values you cite all end in 680402, yet
for a $T GDB packet those are the PC and SP. I think you might be
encountering stack corruption, or stack overflow.

Jifl
-- 
Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062
Maybe this world is another planet's Hell -Aldous Huxley || Opinions==mine