talos-hcode/import/chips/p9/procedures/ppe, branch 07-25-2019

PM: Fix DB0 Hang

2019-09-05T09:01:28+00:00

Key_Cronus_Test=PM_REGRESS Change-Id: I706ec7b87e777b736153d5765ced0a3f6cea5d96 CQ: SW470688 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/81266 Tested-by: Jenkins Server Tested-by: PPE CI Tested-by: Cronus HW CI Tested-by: Hostboot CI Tested-by: FSP CI Jenkins Tested-by: HWSV CI Reviewed-by: YUE DU Reviewed-by: RANGANATHPRASAD G. BRAHMASAMUDRA Reviewed-by: Jennifer A Stofer

PM: Handle PIB Reset with data check handler on all GPEs

2019-05-10T15:55:04+00:00

Key_Cronus_Test=PM_REGRESS Change-Id: I121aaf5efb579fde88829f2ef5354ad4c8b6d77b CQ:SW447494 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/68622 Tested-by: Jenkins Server Tested-by: PPE CI Reviewed-by: Douglas R. Gilbert Tested-by: Cronus HW CI Tested-by: FSP CI Jenkins Reviewed-by: Gregory S. Still Reviewed-by: Jennifer A. Stofer

HCODE Make divide using DERP/DORP atomic

2018-08-24T15:38:53+00:00

Change-Id: Ib7e1c88f7ac934ba551082f27231fff1bb3961dc CQ: SW443103 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/64570 Reviewed-by: RANGANATHPRASAD G. BRAHMASAMUDRA Reviewed-by: RAHUL BATRA Tested-by: Jenkins Server Tested-by: PPE CI Reviewed-by: Gregory S. Still

PM: IOTA Fixes

2018-04-13T14:17:54+00:00

Key_Cronus_Test=PM_REGRESS Change-Id: Ie8be627853aa7452805408cbb159166e70f96e48 CQ: SW424742 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/57124 Reviewed-by: Gregory S. Still Reviewed-by: YUE DU Reviewed-by: Jennifer A. Stofer Tested-by: Jennifer A. Stofer

CME Code Size Reduction ATTEMPT#3

2018-03-22T19:08:27+00:00

-- some IOTA kernel cleanup -- also add checking for IOTA execution stack overflow -- re-coded to eliminate some math library macro usage -- added native 16-bit multiply -- re-coded to remove redundancy from external interrupt handler -- removed dec handler (optional define) and other minor cleanup -- fixed Interrupt initialization code in std_init (all PPE images) -- always inline pstate_db0_clip_bcast & update_vdm_jump_values_in_dpll -- optimized pls calculation code -- optimized pstate init, db1 handler, core good handling -- optimized pmcr requests and pmsr updates (always write for both cores) Key_Cronus_Test=PM_REGRESS Change-Id: I6965a5a581562f0bb1cd735642f592cb4954970b Original-Change-Id: If48fec5832bd5e46cb89f0d6a97d90a488e8ff7b CQ: SW415503 RTC: 178789 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/53381 Tested-by: PPE CI Tested-by: Hostboot CI Tested-by: Jenkins Server Tested-by: Cronus HW CI Tested-by: FSP CI Jenkins Reviewed-by: YUE DU Reviewed-by: RANGANATHPRASAD G. BRAHMASAMUDRA Reviewed-by: Gregory S. Still

PPE empty PK FFDC handler to save-off required registers

2018-02-06T20:34:39+00:00

- Defined PPE_CAPTURE_INTERRUPT_FFDC, by default NULL - Dependent Platform need to plug in a branch for FFDC Change-Id: I896af5b252c8fbf6531634d1a39fc84ed580a594 Original-Change-Id: Iea222662abe19c0b5f9da6eac3e161963b1c047c RTC:179374 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/48977 Tested-by: FSP CI Jenkins Reviewed-by: Douglas R. Gilbert Reviewed-by: Sachin Gupta Tested-by: Jenkins Server Tested-by: PPE CI Reviewed-by: Jennifer A. Stofer

Register FFDC call is handled within machine_check_handler

2018-02-01T23:15:30+00:00

Within the PK interrupt vector, if you have a function call from one function to another, it needs branch link instruction, but since the BL is already being used to jump from software context to the PK Vectored Interrupt context, the function call from within will corrupt the Link register. For interrupts like data_storage, instruction_storage, alignment_exception and program_exception, the save-off will capture the LR registers to indicate which interrupt was getting executed when halt happened. For Machine check handler, LR won't be valid. Change-Id: Iee17b37acd438c7bee2c956cac2de3ce64d04441 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/51587 Tested-by: FSP CI Jenkins Tested-by: Jenkins Server Tested-by: PPE CI Reviewed-by: Sachin Gupta Reviewed-by: Soma Bhanutej Reviewed-by: Douglas R. Gilbert Reviewed-by: Jennifer A. Stofer

PPE empty PK FFDC handler to save-off required registers

2018-02-01T23:15:30+00:00

- Defined PPE_CAPTURE_INTERRUPT_FFDC, by default NULL - Dependent Platform need to plug in a branch for FFDC Change-Id: Iea222662abe19c0b5f9da6eac3e161963b1c047c RTC:179374 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/48977 Tested-by: FSP CI Jenkins Reviewed-by: Douglas R. Gilbert Reviewed-by: Sachin Gupta Tested-by: Jenkins Server Tested-by: PPE CI Reviewed-by: Jennifer A. Stofer

PPE: Adjust the maximum decrementer count

2018-02-01T22:31:21+00:00

Change-Id: I4f59cabf7439590ac736e6f64e35acf11e6c5aa9 CQ: SW402715 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/46944 Tested-by: Jenkins Server Tested-by: FSP CI Jenkins Tested-by: PPE CI Reviewed-by: Martha Broyles Reviewed-by: William A. Bryan Reviewed-by: Michael S. Floyd Reviewed-by: Jennifer A. Stofer

Use HCode header timebase frequency for pk trace

2018-02-01T22:00:35+00:00

Change-Id: Id873da7ccb96192eb69e4aa5cff60ad33c713024 Original-Change-Id: I34a9d8d1a003ae6b07b50039a9ea57ff9fd5af4d RTC: 179852 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/46227 Reviewed-by: Douglas R. Gilbert Reviewed-by: Michael S. Floyd Tested-by: Jenkins Server Tested-by: PPE CI Tested-by: FSP CI Jenkins Reviewed-by: Juan R. Medina Dev-Ready: Gregory S. Still Reviewed-by: Gregory S. Still