[Info-vax] Any chance Perfecttune Perfectfile Perfectdisk will be ported to x86?

[Robert A. Brooks]

On 5/30/2019 1:16 PM, VAXman- at SendSpamHere.ORG wrote:
> In article <00B3FA56.34DA8358 at SendSpamHere.ORG>, VAXman-  @SendSpamHere.ORG writes:
>> In article <qcnejn$ivu$1 at dont-email.me>, "Robert A. Brooks" <FIRST.LAST at vmssoftware.com> writes:
>>> On 5/29/2019 6:14 PM, etmsreec at gmail.com wrote:
>>>
>>>> When I last looked at it, about ten years ago, PerfectCache was able to
>>>> retain coherency, but was not able to co-exist with FastPath.
>>>
>>> Fastpath, or Multipath?
>>
>> Multipath.
> 
> Sorry, I received a phone call from Jim Hibbits.  It was the Fastpath that
> needs to be disabled for PerfectCache.  I could elaborate upon why but for
> now it's sufficient to just state which *path option doesn't play well.

Yeah, I knew that.

I blindly debugged the PerfectCache code to make it work with Multipath.
Bill Davy would describe the behaviour and I'd try to guess what he was doing wrong.

I don't remember the specifics, but the problems centered around switching to 
the MSCP-served path.

To make life easier, we added the 3rd-party intercept mechanism to allow a
supported way to intercept I/O.  It was documented in the V7.3-2 New Features
manual, and will properly documented in the forthcoming I/O User's Guide.

-- 
                  -- Rob