Minutes from the 15 Mar 2022 IBIS Quality Task Group meeting are attached.
Mike
Minutes, IBIS Quality Task Group
15 March 2022
12:00-13:00 EST (09:00-10:00 PST)
ROLL CALL
ANSYS Curtis Clark
Cadence Design Systems * Yingxin Sun
Intel Technology * Michael Mirmak
Micron Technology * Randy Wolff
Siemens EDA Weston Beal
MathWorks: * Mike LaBonte
Teraspeed Labs: * Bob Ross
Zuken USA: * Lance Wang
Everyone in attendance marked by *
NOTE: "AR" = Action Required.
-----------------------MINUTES ---------------------------
Mike LaBonte conducted the meeting.
Call for IBIS related patent disclosures:
- None
Call for opens:
- None
Review of previous meeting minutes:
Minutes from the March 1 and 8, 2021 meetings were reviewed. Randy Wolff moved
to approve
the minutes. Bob Ross seconded. Without objection the minutes were approved.
ARs:
- AR: Weston Beal to prepare presentation on IBIS Quality Specification and
checklist
No report.
NEW ITEMS:
Parser bug reports:
Yingxin Sun had sent information about an IBISCHK problem by email. Yingxin
said a full
Windows path caused an error message about "illegal character with ASCII code
58".
Yingxin demonstrated the problem using a tool "T2B", which integrated the
IBISCHK source
code. A test IBIS file showed no errors when checked, but it had two errors
when checked
a second time. Yingxin said the different behavior was related to a
"geFileVersion > 7.1"
conditional.
Lance Wang asked if the IBISCHK program had the same behavior when run
directly. Yingxin
had not tried that. Mike LaBonte suggested that T2B could enable the IBISCHK
"-numbered"
option to get more helpful error codes.
Mike asked if Yingxin was requesting an enhancement for IBISCHK to allow being
called with
a full path. Mike was not sure why ibischk did not allow full paths. Bob Ross
said the
IBIS specification did not allow full paths. Mike said that might apply to
[File Name],
but the IBISCHK command line should not have that restriction. Michael Mirmak
said it was
easier to write IBISCHK if the passed file name could be matched literally
against [File
Name].. Mike suggested T2B could chdir to the enclosing directory, and then
pass a
relative path to the IBISCHK PARSE function.
Yingxin pointed out that his email showed a code snippet in which a variable
"ptr" was set
to the relative path, but never used. The check function called below that was
passed the
full path variable "sFileName", and that function printed the errors. Yingxin
said it
looked like "ptr" should be used in place of "sFileName" in that call.
Lance did not see why IBISCHK would not reproduce the error. Mike asked why
T2B did not
give the error the first time. Yingxin said the variable geFileVersion was
null the first
time, and the check was not run. Mike asked if that was another bug. Yingxin
said it
was.
Yingxin showed that changing the file from [IBIS Ver] 2.1 to 5.1 gave different
behavior.
The checks were performed as though the [IBIS Ver] of the previously checked
file was in
effect. Mike said that was a global variable problem, and he might be able to
reproduce
the problem with a modified IBISCHK that ran the check twice. He said once a
fix was
found for the geFileVersion bug, we could let Yingxin know.
Bob said the testcase file could be reduced. Mike said only two factors
mattered: calling
the check with a full path, and checking in the same session with different
[IBIS Ver]
values. Yingxin said he would supply a reduced testcase file.
AR: Yingxin Sun to provide reduced testcase file for full path error bug.
IBISCHK bug web page:
Mike LaBonte said webmaster Steven Parker would add a Supporting Version column
to the
table in the IBISCHK bug page.
IBIS quality checklist:
No discussion.
Tabled topics (no discussion without motion):
- BIRD181.2
- IBISCHK security fixes
IBISCHK BUG227:
Michael Mirmak said IBIS-AMI root name matching requirements would be discussed
in the ATM
meetings. He asked what the IBIS Quality Task Group needed to proceed on
BUG227. Mike
LaBonte said an enhanced [Test Data] and [Test Load] keywords was the other
issue.
Otherwise, IBISCHK might send the DLL all defaults in AMI_parameters_in, and
that would
not test the actual parsing of the string. Randy Wolff said only the root name
needed to
be checked. Mike asked if BUG227 could be considered fixed with only a root
name check.
Michael said only the tool could check one side, and only the model could check
the other
side. He preferred to reduce parser work, and only check the root name.
Mike asked if AMI_GetWave should be checked. Michael said it should be only
necessary to
check AMI_Init. He said if we had [Test Data] enhancements, then we might check
AMI_GetWave. Randy said he preferred to reduce the development cost,
especially since
there would be no additional funding for work on IBISCHK 7.1.1. Bob Ross
suggested not
modifying [Test Load] and [Test Data], suggesting that a new keyword would be
better.
Michael agreed, saying those keywords did have issues, and it may not be worth
fixing
them. Mike felt IBIS-ISS could be used for [Test Load], if it was ever
enhanced.
Mike asked if the bug report needed an update.
Bob felt we could update the bug report later.
Lance Wang moved moved to adjourn. Bob Ross seconded. Without objection the
meeting ended.
Meeting ended: 12:04 ET
Next meeting March 15, 2022