login
Header Space

 
 

Re: kgdboe?

Score:
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
To: Jesse Barnes <jbarnes@...>
Cc: Jason Wessel <jason.wessel@...>, lkml <linux-kernel@...>, David S. Miller <davem@...>, Christoph Hellwig <hch@...>
Subject: Re: kgdboe?
Date: Tuesday, May 6, 2008 - 9:16 am

* Jesse Barnes <jbarnes@virtuousgeek.org> wrote:


i think the proper way would be to extend the kgdb-over-console (kgdboc) 
concept and make it easy to utilize the existing netconsole code via 
KGDB. This brings nice synergy: netconsole is utilized by other debug 
infrastructure as well so anything that supports netconsole would 
automatically support KGDB as well. I dont really like the old kgdboe 
approach from a design POV as it created a parallel infrastructure for 
special-purpose kgdb IO.

likewise, the firewire debugging code could/should be extended to create 
a bidirectional console as well - that has other benefits as well. (we'd 
get printks over firewire - instead of the special-purpose firescope 
utility)

and the early console drivers could be reused to do early debugging via 
KGDB as well.

	Ingo
--
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
[GIT PULL] kgdb fixes for 2.6.26, Jason Wessel, (Mon May 5, 8:32 am)
Re: kgdboe?, Jesse Barnes, (Mon May 5, 2:10 pm)
Re: kgdboe?, Ingo Molnar, (Tue May 6, 9:16 am)
Re: kgdboe?, Jesse Barnes, (Tue May 6, 1:05 pm)
speck-geostationary