Re: [PATCH for review] [123/145] i386: make fault notifier unconditional and export it

[Keith Owens <kaos@ocs.com.au>]

Andi Kleen (on Sun, 13 Aug 2006 22:17:48 +0200) wrote:
>On Sunday 13 August 2006 17:28, Adrian Bunk wrote:
>> > It's needed for external debuggers and overhead is very small.
>> >...
>> 
>> We are currently trying to remove exports not used by any in-kernel 
>> code.
>
>That ``we'' doesn't include me at least.
>
>> 
>> The patch description also lacks the name of what you call "external 
>> debuggers" (I assume the exports are not for a theoretical usage but for 
>> an already existing debugger?).
>
>The fault chain is needed for pretty much any debugger, including
>kgdb, kdb, nlkd. The one in this case was NLKD.

No.  The page fault event was only used by kprobes, but it slowed down
all code.  That is why it was changed to only be present if kprobes was
in the system.  kdb and AFAIK kgdb do not need the page fault chain.
nlkd might need it, but that does not seem to have been explicitly
stated.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/