Minutes, IBIS Quality Task Group 11 January 2022 12:00-13:00 EST (09:00-10:00 PST) ROLL CALL ANSYS Curtis Clark Intel Technology Michael Mirmak Micron Technology * Randy Wolff Siemens EDA Weston Beal Signal Integrity Software: * 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: - Bob Ross said there was a new testcase to discuss. Review of previous meeting minutes: Minutes from the December 21, 2021 meeting were reviewed. Bob Ross moved to accept the minutes. Randy Wolff seconded the motion. Without objection, the minutes were approved. ARs: - AR: Mike LaBonte to publish new ibischk7 web page Done. - AR: Randy Wolff to announce ibischk710 release Done. - AR: Bob Ross to distribute ibischk7.1.0 source code to ibischk7 licensees Done. There were 12. - AR: Randy Wolff to test new ibischk710 on RedHat Linux Done./ It worked. - AR: Mike LaBonte to write ibischk bug report to eliminate system calls Not done yet. - AR: Mike LaBonte to update ibischk710 user guide to correct EMS references Not done. Mike will double-check. - AR: Bob Ross to fix brackets around Priority in ibischk BUG227 Done. - AR: Weston Beal to prepare presentation on IBIS Quality Specification and checklist No report. - AR: Mike LaBonte to document TSCHK2 Windows PowerShell problem This was done, on the TSCHK2 web page. - AR: Bob Ross to upload prior IBISCHK source code distributions Done. Bob said it began with 2. NEW ITEMS: IBISCHK710 development: Bob Ross reported that the IBISCHK developer had been paid December 28, 2021, as promised. We were still waiting for the message spreadsheet, needed to complete the new user guide. New parser bug reports: Bob Ross said we had deferred classifying BUG227 in the previous IBIS Open Forum meeting, to see if there would be a new related BIRD. Mike LaBonte said the BIRD would make the difference between the message being a warning or an error. Bob said the issue had been caused by an expectation set in EDA tools. Mike said Michael Mirmak had proposed enhancing [Test Data] to support AMI testing, even by EDA tools. Mike was not sure EDA tools would implement the testing if ibischk was there to do it. Bob said discussions around that could become complicated. Randy said some EDA tools found the root name mismatch problem readily. Bob said this should be discussed in ATM group meetings. He said it would be a IBISCHK 7.2.0 parser change. Mike felt we might want to make the new fix in IBISCHK 7.1.1 to reduce the total amount of work for needed for IBISCHK 7.2.0, considered the BIRD changes. IBIS version bug: Bob Ross said a testcase received from Arpad Muranyi was still too large, after attempting to reduce the testcase, which was originally from Micron. Randy Wolff said it would take a lot of work to make a small testcase. He said the testcase [Component] had [IBIS Ver] 7.0 or 7.1 errors that should be reported, but the file in this case was [IBIS Ver] 5.0. When the file was referenced by a .emd file, the EMD version number was being used instead of the .ibs file version. Parsing the IBIS 5.0 file directly produced no error messages. Changing the .ibs file to [IBIS Ver] 7.0 produced 179 errors. Bob said the issue found might be a valid issue when the 5.0 file is used as part of an EMD. Randy said the .emd file did not referenced power and ground, so it had no bearing on the flagged issue. Bob said the source of the problem in the files was difficult to track down. Randy said [EMD Parts] was triggering the problem. Mike LaBonte asked if a testcase might be built from scratch. Bob said there might be two EMD groups. He said [IBIS Ver] 7.0 and 7.1 files should give the same behavior for that error. Randy said they did. Bob said he wanted a testcase that could go into the parser source code QA test suite. Randy said BIRD 182 introduced the new rule that resulted in error E5808. He showed BIRD182 and the IBIS file with the [Pin Mapping] problem. Pin A2 had signal_name VSSQ in [Pin], which became the bus_label name. In [Pin Mapping] A2 pulldown_ref was connected to bus_label VSS, which created a conflict. IBISCHK found 179 errors in the IBIS 5.0 file checked directly when it was changed to [IBIS Ver] 7.1, but only 25 when checked from the .emd file. Bob said VSSQ was dangling in the file, not really defined as a bus_label. Randy said the problem could be fixed by changing VSSQ to VSS, but VSSQ was the databook name. Bob said that arrangement was originally not anticipated, so the rule came later. Randy said the opposite case, where multiple bus_labels map to one signal_name, was allowed. Mike asked if that check could have been applied to files before IBIS Ver 7.0. Randy said we did not do that, to avoid impacting existing files. Randy said the checking should be the same, regardless of whether called from EMD. Bob asked if the BIRD182 checks should be applied retroactively. Mike said we would need a testcase for the 25 vs. 179 error problem. Randy said the problem component was not even used by the EMD, it was an extra in the file. Significant work would be required to make the testcase valid. Bob said it was difficult to trace through the files. Mike wondered if the 25 errors were only for components called by the EMD. Randy said some were not. Mike said he would compare the error code sets from both tests. Randy said E5808 was only produced by EMD checking. He felt we needed to consider retroactive checking. AR: Mike LaBonte to investigate IBISCHK error differences calling from EMD vs. directly IBIS quality checklist: No discussion. Tabled topics (no discussion without motion): - BIRD181.2 - IBISCHK security fixes Randy Wolff moved moved to adjourn. Lance Wang seconded. Without objection the meeting ended. Meeting ended: 13:24 ET Next meeting January 18, 2022