[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


On Mon, Jul 23, 2001 at 11:26:03PM +0200, Bjorn Wesen wrote:
> On Mon, 23 Jul 2001, Dragan Stancevic wrote:
> > The reason that it seems cut off is because printk uses a ring buffer and I
> > am sure that we had an overrun because I think that it's higly likely that
> > kernel was printing messages much faster than the console was abble to send
> > over the serial.
> My debug-outputs using printk never get stuck in a ring-buffer even if
> they swamp their output.. after all, they are synchronous when using a
> serial debug-console, the printk does not exit until it has completed
> the writing. Otherwise they would be totally worthless for debugging
> anything at all :) 

OK, maybe I wasn't clear enough, when I said "The reason that it seems cut
off is because printk uses" I meant that "the reason that those lines you
saw earlier were missing is because prink uses ring-buffer"

I didn't say that printk will get stuck in a ring-buffer, printk is a kernel
function that drops output into a ring buffer, another thread is picking it
up from that ring buffer, if printk is faster being called more often than
the app(eg, syslog) or console the ring will overrun, also it is komonly
known that if you use
printk for debugging it doesn't necessarly appear in your dmesg in the order
it was called.  Ayway that's besides the point, if you are referring to the
last line, why it wasn't printed out all the way, it is possible that the
system rebooted but since I am having problems booting at all at times the
board might have been stuck after a reboot/crash.

As far as the missing lines they were either overwriten in the ring buffer
or the system rebooted before they were abble to appear on a screen.

Doy you have IDE specs at hand to look up what are the states, I don't have
the specs.

Peace can only come as a natural consequence
of universal enlightenment. -Dr. Nikola Tesla