Back to Minutes page
Back to IEEE P1532 home page
The following individuals attended the meeting:
Neil Jacobson Dave Bonnett Mark Moyer Kurt Guntheroth Sung Chung Ted Eaton Dave Thompson Howard Tang Brad Ishihara Ken Parker Glen Rosendale (7-19-99 Only) Yadvindra Dhami Arthur Khu Alan Herrmann(7-20-99 Only)
IEEE P1532 Working Group Minutes of meeting at Atmel July 19th and 20th
From 7-19-99
The ITC meeting will be held on Tuesday 9-28-99 6-7Pm and will be an informational meeting only.
Motion: Minutes
Minutes from the previous meeting be approved.
Motion: Mark Moyer
Second: Ken Parker
Carried: 11 - 0 - 0
Section 5.1.11
The word "and" will be added on Page 40 between time and amount.
Kurt expressed a concern that to support "sectional" programming of all devices that auto-incrementing devices would require arbitrary instruction initialization ability.
Action Item(s): Ken Parker
Ken will expand the definition of the ISC_NOOP Instruction to include "The contents of the ISC registers will not be affected by ISC_NOOP".
A rule will be added to 6.1.1 to specify no minimum or maximum time requirements between completion of an ISC instruction and the start of another. (From Kurt’s notes)
Ken will add wording to specify that the PData register does not necessarily contain address and data. (From Kurt’s notes).
Rule 6.4.g will be moved to general rules?
The word Programming will be removed from the second sentence of 6.4.12
The words "for example" will be added prior to "(via the ISC_ADDRESS_SHIFT instruction)" in sentence 1 of Section 6.6.5
Rule 6.6.5.1.b will be removed.
Permission 6.6.5.1.c: ISC_ERASE will be replaced by ISC_INCREMENT
Figure 20 Xs will be added the ISC_PData column for ISC_PROGRAM_DONE, ISC_ERASE_DONE,
and ISC_PROGRAM_SECURITY
Note: This will require updates to 6.5.1,6.5.2,6.5.3
Motion: ISC_PROGRAM_SECURITY
The ISC_Program_Security instruction will have its own register.
Motion: Ted
Second: Kurt
Defeated: 3 - 5 - 1
Prior to the previous vote Ted called a point of order.
Due to recent company mergers we had exceeded the allowable voting members allowed by a single company. It was later agreed that the rule would be enforced.
Action Item: Ken Parker
It will be explicitly described that 1149.1 instructions as well as the ISC_NOOP instruction have no effect on ISC register contents.
Motion: Program Resume
(A) Mechanism (will) be defined for resuming ISC operations after ISC_Disable issued before (the) last address (is) accessed.
Motion: Brad
Second: Kurt
Withdrawn
From 7-20-99
Section 8.3.2.2
Neil brought up the idea of multiple duration times within a single chip.
Proposed Change:
Attribute ISC_Instruction_Parameters of <> : entity is
" ISC_Enable, *,*, wait_duration",
" ISC_Program,xx1,xx1,wait_duration",
" ISC_Program,xx1,x1x,wait_duration",
Tabled for later.
Proposed Change:
Section 8.3.2.5 Wording will be added to require each section of the algorithm will consist of 0-n operation triples.
I.e.:
IRSCAN
WaitDuration
DRScan(NullDRScan)
This was later revisited and deemed not accurate. Ken will investigate wording to appropriately identify the requirements.
Pg. 86
The group focused the discussion on how much or how little structural algorithm information was required in the BSDL.
It was noted that Stapl(Jam) has two inherent layers defined. I consists of a base set of building blocks such as Program, read_back, erase..., as well as a higher level sequencing of these building blocks.
The basic consensus of the group (with several notable holdouts) was that to facilitate 100% vendor independent software development we would be required to add this higher level of sequencing.
This discussion was tabled.
Motion: ISC_Address capture value
The Capture-Dr value in the ISC_Address register of the ISC_ADDRESS_SHIFT instruction will be the current address value
(The Capture - DR value of ISC_ADDRESS_SHIFT will be the device’s current address value.)
Motion: Ken
Second: Dave
Carried: 11-0-0
The discussion then switched to Kurt’s "Proposal #2" which was posted to the reflector.
The group discussed the pros and cons of the method. It was noted that Kurt had incorporated many of the features of Ken’s proposed data file moving the operator control information into the BSDL. The group decided that it was worth while to pursue the format.
Action Item: Kurt
Kurt will investigate the addition of expected and/or mask information in order to support combined program/verify operations. He will also explore a method of describing a reentrant attribute of the ISC_ACCESSED mode.
In addition Kurt will take a stab at the definition of operation_sequences.
Next Meeting:
The Next meeting was scheduled for Sept. 13-14th and Xilinx will host.
Motion: Hardware Standard first
Motion: Brad
Second: Mark
Motion to Table: Ted
Second: Kurt
Action Item(s): Neil
Neil will include the previous motion as an agenda item for next meeting.
Neil will include a discussion of Architecture shift on the agenda for next meeting.