[illumos-Advocates] RTI: RFE #937 - WRITE_SAME support for COMSTAR

[Dan McDonald]

On Wed, Jun 01, 2011 at 11:24:57AM -0400, Eric Schrock wrote:
> Sounds to me like Dan has verified that the VMWare beta is indeed issuing
> the new commands and the code is reacting correctly in response.  That
> sounds good enough to me.  The only thing I can think of is to use libscsi
> to manually send bad requests to stress edge conditions and make sure we
> don't tip over or blow an assert on bad input, but that's probably overkill.

There is a larger more generic problem with some of these edge cases, which
in turn begs the question about a good regression battery for COMSTAR's
iSCSI.

Garrett's right --> I did a lot of performance tests with various
parameters.  I used VMware's new VM initialization with the blank-the-disk
option turned on.  That uses WRITE_SAME.

Dan