Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 17 Dec 2005 13:13:00 -0700
From:      Scott Long <scottl@samsco.org>
To:        Nate Lawson <nate@root.org>
Cc:        cvs-src@freebsd.org, src-committers@freebsd.org, cvs-all@freebsd.org
Subject:   Re: cvs commit: src/sys/sys ktr.h src/sys/kern kern_clock.c      kern_switch.c
Message-ID:  <43A4714C.1080202@samsco.org>
In-Reply-To: <43A547F2.2090401@root.org>
References:  <43A547F2.2090401@root.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Nate Lawson wrote:
> Scott Long wrote:
> 
>> Nate Lawson wrote:
>>
>>
>>> njl         2005-12-17 03:57:10 UTC
>>>
>>>  FreeBSD src repository
>>>
>>>  Modified files:
>>>    sys/sys              ktr.h    sys/kern             kern_clock.c 
>>> kern_switch.c  Log:
>>>  Clean up unused or poorly utilized KTR values.  Remove KTR_FS, 
>>> KTR_KGDB,
>>>  and KTR_IO as they were never used.  Remove KTR_CLK since it was only
>>>  used for hardclock firing and use KTR_INTR there instead.  Remove
>>>  KTR_CRITICAL since it was only used for crit enter/exit and use
>>>  KTR_CONTENTION instead.
>>>  
>>>  Revision  Changes    Path
>>>  1.183     +1 -1      src/sys/kern/kern_clock.c
>>>  1.118     +2 -2      src/sys/kern/kern_switch.c
>>>  1.35      +12 -12    src/sys/sys/ktr.h
>>
>>
>>
>> Um, I was using KTR_CRITICAL for schedgraph.  It was actually quite
>> useful.  Compressing the option space only makes the options less
>> useful.  Surely there has to be a better solution.  Or, at least you
>> could call for comments before you alter this stuff.
> 
> 
> You didn't speak up about that in the previous discussion on arch@, 
> starting 10/31/2005.  The only comment was jhb@ saying it was not useful 
> alone, and he's the only one doing work on critical sections lately.

Um, cut me a little slack for WORKING ON A DAMN RELEASE that week.

> 
> If you can think of another use for this besides one event (enter/exit), 
> feel free to add it back.  Or, consider adding KTR_SUBSYS as a one-off 
> use like KTR_DEV is for other parts of the system.  KTR_CRITICAL would 
> be conditionally defined as KTR_SUBSYS when needed.
> 

It may only be a single event pair, but they are fired quite often. 
Cluttering up the option with more events only makes it harder to 
collect a useful amount of data.

Btw, fix your system clock.

Scott



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?43A4714C.1080202>