Skip to main content

Full text of "xerox :: sdd :: memos 1978 :: 19780912 Pilot Status Report"

See other formats


XEROX 



BUSINESS SYSTEMS 
Systems Development Department 

To: Robert Kierr Date: September 12, 1978 

From: H. C. Lauer Org: SDD/SD 

Subject: Pilot Status report Filed: [Iris] < Lauer > PilotStatus.memo 

Reference: Pilot Pre-alpha and Alpha releases, dated September 5, 1978 from H. C. Lauer to Robert Kierr 



The following is a brief summary of the status of Pilot and of the components which, by agreement, 
will be included in the Alpha release of Pilot. It appears that it is possible to come very close to 
meeting our September 15 target for starling the release procedure, llie remaining unknowns are: 
(1) how rapidly we can get new microcode releases into use by the Pilot group, and (2) the number 
of "surprises" in demand paging. 

Demand paging 

All modifications to Pilot necessary to test demand paging are complete. 

We have mn VFSTest (formerly MockCIient) for an extended period in page faulting 
mode, but without paging out existing code or data. 

Charnley has identified a means of making BilBlt restartable after page faults (at tlie 
expense of speed) 

Garner is fixing a bug in XFER which prevents code segments from being demand paged. 

Remaining tasks: 

Fix tlie Map Log to allow larger log to be kept 
Continue testing 

OIS and PUP communications 

Yogen is on schedule and in the midst of testing Pup communications (sec September 5 
memo); he has ~3 days remaining work on his contributions to tlie Alplia release. 

Disk-bootable microcode 

Charnley claims that tliis problem is solved; a hardware fix is required to make it work. 

Charnely will produce a new release of microcode Wednesday containing: 
Long pointer disk controller 
Reslarlable Bit}31t (short or long pointers) 
"Identical" .BT and .MiJ versions 
(l^crliaps) Garner's fix for the XFi^'k bug 

Remaining tasks: 

Make a second microcode release containing long pointer lUTFP and last minute 
bug fixes 

Compatible long pointer IRDC 



Pilot status report 2 

Pitts has developed and tested a controller and driver. 

Redell is waiting for this to appear in an official version of microcode to proceed 

Remaining tasks: 

Convert DisklO for long pointer arguments 
Move buffers into hyperspace 

Mapping off tlie ends of files 

Remaining tasks: • . 

Minor bug fixes 
Obscure case testing 

Logging 

Testing has been bogged down in the cycle of getting Jalics's bug fixes bound into Pilot 

Tom Horsley will take over responsibility from Paul Jalics 

The public interface to the logging functions appears to have settled down 

Remaining tasks: 

Convert Jalics's existing implementation to final version of interface 
Finish testing 

Compatible long pointer lUTFP and Long BitBlt 

Charnley has converted BitBlt to accept either short or long pointers; Wick will produce a 
new version of BitBltDefs for using this. 

Remaining tasks: 

Convert Display controller per Pitts's specification 
Test 

The following other items are also of interest: 

Iris has been very flakey this past week. Since this is our principle vehicle for storing and 
transmitting files and since we use it heavily, an hour of down time per day is 
beginning to hurt. We are starling to find that we cannot do the task next on the 
list because Iris is down and we cannot get at the necessary files. // Iris gels any 
worse, it will impact our schedules. 

All sources and command files on CoPilot have been converted to the Librarian, and Pilot 
implcmentors arc using the Access utensil successfully. So far no serious problems 
have been encountered, and users seem happy with it. 

Greg Shaw and Diamond have been extremely helpfi.il in uncovering a host of minor bugs. 

For the forthcoming release of Pilot, we will overlap final testing of its accompanying 
microcode release with the six-day release schedule. In the future, however, we will 
require a microcode release to be in our hands at least one week prior to starting the 
Pilot release. 

Ilcms not affecting the Alpha release of Pilot: 

■ Victor Schwartz is at approximately day 6 of the seventeen day schedule for converting and 
integrating llie RS232C into Pilot (see my Sept 5 memo) 

'ilie Microcode Executive is awaiting the return of the c/.arina from Bora Bora 

The Pilot kit still needs to be defined 



Pilot status report 

c: W. C. Lynch, Pilotlmplcmcntors