Low-level trace-buffers in CAM

[Pokala, Ravi]

-----Original Message-----


From: Slawa Olhovchenkov <slw at zxy.spb.ru>
Date: 2015-10-28, Wednesday at 07:03
To: Ravi Pokala <rpokala at panasas.com>
Cc: "freebsd-geom at freebsd.org" <freebsd-geom at freebsd.org>, "freebsd-scsi at freebsd.org" <freebsd-scsi at freebsd.org>, "freebsd-hackers at freebsd.org" <freebsd-hackers at freebsd.org>, "ken at freebsd.org" <ken at freebsd.org>, "imp at freebsd.org" <imp at freebsd.org>, "scottl at freebsd.org" <scottl at freebsd.org>
Subject: Re: Low-level trace-buffers in CAM

>On Tue, Oct 27, 2015 at 02:22:33AM +0000, Pokala, Ravi wrote:
>
>> Hi folks,
>> 
>> ----------------------------------------------------------------
>> This is an updated re-send of a message I originally sent about a year ago, during MeetBSD 2014. A few people expressed interest in person, but no one ever followed up on the lists. I'm bringing this up again, in the hopes that I can get some feedback before / during next week's Dev/Vendor Summit. I'm CCing some folks who expressed interest at that time, and some folks that I was told would be interested.
>> ----------------------------------------------------------------
>
>What about also export average queue length? May be request wait time in queue.

Those are certainly good to have, but are at a higher layer than what I'm trying to capture. I'm really interested in what's going into and out of the hardware - the values of the CDB fields just before they get sent to the hardware, and the elapsed time and return-CDB fields when the SIM driver's interrupt handler fires at command completion. Admittedly, most people don't need that level of detail, but I work for a storage company, so ... ;-)

-Ravi