How to send 1MB I/O size in a single I/O request without split

[Sibananda Sahu]

Thanks Scott for your great explanations.
I really appreciate your time that helped me much in understanding the OS
behavior.

Thanks,
Sibananda

-----Original Message-----
From: Scott Long [mailto:scott4long at yahoo.com]
Sent: Friday, January 30, 2015 11:12 PM
To: Sibananda Sahu
Cc: freebsd-scsi at freebsd.org
Subject: Re: How to send 1MB I/O size in a single I/O request without split


> On Jan 30, 2015, at 2:41 AM, Sibananda Sahu <sibananda.sahu at avagotech.com>
> wrote:
>
> Hey Scoot,
>
> Thanks very much for your reply.
>
>>> -          How can I get more sge count in an I/O request?
>>>
>
>> If you want to test multiple segments, I suggest leaving you system
>> running for a long time with multiple processes freeing and
>> allocating memory so that the system becomes >fragmented.  You can
>> also modify the bus_dma_tag in your driver to specify a maximum
>> segment size of 4k instead of (presumably in your case) something
>> larger.  That will force >busdma to stop merging adjacent segments.
>
> As you have said:
>
> "The busdma API will see that the allocation is contiguous and attempt
> to merge the contiguous segments.  This is usually desirable since few
> segments reduces processing overhead in the driver and the hardware."
>
> If I will modify the bus_dma_tag in our driver to have a max segment
> of 4K then what is the performance impact?
>

Most drivers have a for() loop that iterates through the list of segments
from busdma and stores them into SGE elements for the hardware to use.
Obviously, fewer segments means fewer trips through this loop.  It also
means a smaller SGE list, which means less command data transferred over the
PCI bus.  Though I cannot speak for the architecture details of your
hardware, often times the same optimizations apply to the hardware
internally.  With modern storage controllers that are doing millions of i/o’s
per second, these small improvements can add up.

Scott