talos-hostboot/src/runtime, branch 07-25-2019

Always send NV_STATUS to Opal when opal-prd starts

2019-12-12T21:13:53+00:00

Opal does not persist NVDIMM status state so HBRT must send this info to Opal when opal-prd starts. Change-Id: I0771c63ff2857f9b70fbfad5760267e3da463009 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/87541 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Roland Veloz Reviewed-by: Matt Derksen Reviewed-by: Daniel M Crowell

Developer Improvement: Get code coverage tool working with Hostboot

2019-05-13T14:10:45+00:00

This commit fixes GCOV code coverage for P9 with GCC 4.9.2 Change-Id: Ie1e7c35f67414531dbd6e7a771ac1529a9ebd59d RTC: 208351 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/76812 Tested-by: Jenkins Server Reviewed-by: Nicholas E. Bofferding Reviewed-by: Ilya Smirnov Tested-by: Jenkins OP Build CI Tested-by: FSP CI Jenkins Tested-by: Jenkins OP HW Reviewed-by: Daniel M. Crowell

Fix bug that was scrambling traces in errorlogs during HBRT

2019-03-26T21:57:44+00:00

There was a bug in how HB was interpreting the clock time that PHYP returned in response to a get_clocktime request. The nsec member of the struct returned was getting the first 32bits filled in and Hostboot was expecting the value to be in the last 32bits. This change will shift the nsec value down 4 bytes if we find the back half of the nsec value to be all 0's. I believe the reason the wrong bits were filled in is because PHYP is returning a packed struct whose 2nd member is a 32 bit value so it does not get filled in correctly. Change-Id: I71f1c2cb672b9979e0b6bdc8b1852544f230b885 CQ: SW448397 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/74823 Reviewed-by: Matt Derksen Reviewed-by: Corey V. Swenson Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Daniel M. Crowell

Support thread local storage

2019-03-11T22:07:52+00:00

Generally adds support to declare variables as thread_local - Add support in HBRT start assembly to skip adjusting TLS relocations - Add support in linker to generate tagged TLS entries - Update linker to process TLS relocations correctly - Update TLS code to ignore top half of module ID - Update module images to hold a "module ID" - Update custome linker to update module ID during binary link - Update TLS code to track TLS sections via module ID Change-Id: I1589550d7787beb08827ca24a728397dedf0373b RTC: 147599 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/71709 Reviewed-by: Ilya Smirnov Reviewed-by: Michael Baiocchi Tested-by: Jenkins Server Tested-by: FSP CI Jenkins Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Reviewed-by: Daniel M. Crowell

Attribute support for recursive mutexes

2018-11-12T15:07:12+00:00

Added a new recursive mutex attribute, new test cases, and initializers for recursive mutex attributes. Change-Id: I49e6bc4fc2fd84e5c46e5d4c5ec125270347bde5 RTC: 196793 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/67939 Reviewed-by: Nicholas E. Bofferding Reviewed-by: Ilya Smirnov Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: FSP CI Jenkins Tested-by: Jenkins OP HW Reviewed-by: Daniel M. Crowell

assert: Include file name in assert output

2018-10-02T19:36:19+00:00

Printing the line number without an associated file is almost meaningless; add the file name to improve debug ergonomics. Change-Id: I454876ccc567cb58066362caced7efbaf92cfe62 Signed-off-by: Andrew Jeffery Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/66117 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Corey V. Swenson Reviewed-by: Daniel M. Crowell

HBRT Reserved Mem Trace Buffer implementation.

2018-06-19T21:24:01+00:00

-Hostboot master Drawer will create TRACEBUF section in the HB Rsvd Mem and HBRT, while booting, gets the TRACEBUF section details and initializes the circual buffer, if available. -If a valid data is present in the buffer, HBRT will create an info log ERRL, otherwise initializes the buffer. -Traces from all components are serialized and stored in the single TRACEBUF in the same format (fsp-binary Trace). When buffer is full, it deletes the oldest entry/entries to store the new ones. -ERRL->collectTrace() works similar to other buffers. -A new test case is added to dump the traces from TRACEBUF. Change-Id: I4ce943231a2ba30e3a13ca34d1c40ff68464a994 RTC:188726 RTC:191302 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/56450 Reviewed-by: Prachi Gupta Reviewed-by: Daniel M. Crowell Tested-by: Daniel M. Crowell

Check for pending OCC messages on HBRT start in OpenPOWER

2018-04-17T18:15:53+00:00

The OCC will set an interrupt for the Host to look for a pending message/action. If opal-prd is not running at the time that this interrupt happens, HTMGT will never see the message. Therefore we need to have HTMGT explicitly check for anything pending any time HBRT starts up. Change-Id: Iaae075ae6828a9de5a7bd5afe55c541b0076f2d2 RTC: 187912 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/57175 Reviewed-by: Christopher J. Cain Reviewed-by: Matt Derksen Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Daniel M. Crowell

Initialize FIRDATA section and ErrlManager just incase BMC resets

2018-03-07T17:54:24+00:00

PRD needs to read the PNOR to tell OCC where to write in the pnor. This will cache the PNOR section so PRD will not fail the read call. If an error is created with pnor inaccessible, errl manager could get stuck unless it was initialized first. Change-Id: Ia7212bf99e947a1016423b37aff6f13ddfc7190a CQ:SW412798 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/54990 Reviewed-by: Martin Gloff Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Reviewed-by: Roland Veloz Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins CI-Ready: Daniel M. Crowell Reviewed-by: Daniel M. Crowell

Initialize SBE message communication on HBRT start

2018-02-07T22:40:16+00:00

Need to clear first 2 bits in CFAM register 0x283B. This basically resets our SBE communication stream just incase HBRT restarts after processing SBE interrupt. Clearing these bits tells the sender to resend the command. Change-Id: I1a1c5b80cc99414649d38c06b93428c87299f242 RTC:186334 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/53225 Reviewed-by: Martin Gloff Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Reviewed-by: Brian E. Bakke Reviewed-by: Roland Veloz Tested-by: FSP CI Jenkins Reviewed-by: Daniel M. Crowell