blackbird-hostboot/src/include/kernel, branch 07-25-2019

Add current istep into TI SRC

2019-12-18T17:41:02+00:00

Word4 of the SRC is defined to be the 'last progress code' for FSP SRCs. For Hostboot TIs, we currently leave that word zero. This change will add the same data that we put into the scratch register 5 into word4 so that we will have the failing istep for any TI we encounter. Change-Id: Iaf5ec835d45b1ea3a6ced20b5b7f7d07216c548e Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/88813 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Ilya Smirnov Reviewed-by: Matt Derksen Reviewed-by: Nicholas E Bofferding

Automatically include config.h

2019-12-06T16:28:47+00:00

Rather than having to remember to include config.h anywhere we reference a CONFIG variable (and usually forgetting), this adds it to the default compiler flags so that it gets included in every source file we build. Change-Id: I53622ab4d46c55d942e98cae6ec03049fd5b3d08 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/87475 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Zachary Clark Reviewed-by: Roland Veloz Reviewed-by: Christian R Geddes Reviewed-by: Nicholas E Bofferding

Fix deadlock in ECC error shutdown path

2019-04-01T21:59:46+00:00

There is a scenario wherein if a PNOR ECC UE occurs when attempting to service a page fault, the kernel will deadlock when it attempts to kill the task that triggered the fault. This is due to the kill routine being called while holding the VMM spinlock and then collecting a backtrace, which also performs a nested acquire of the same spinlock. This fix inhibits invoking the backtrace if the kill routine holds the VMM spinlock. Change-Id: I75bf1f248740a08fd485379d88e146096edf65a9 CQ: SW461429 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/75264 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Ilya Smirnov Reviewed-by: Michael Baiocchi Reviewed-by: Matthew Raybuck Reviewed-by: William G. Hoffa

Tweak page eviction parameters to handle low mem better

2019-03-18T18:27:44+00:00

We are memory constrained in some configurations during 10.2 because we are forced to pin large chunks of memory to hold a few data partitions. In some instances we have run out of free memory completely. The fact that we don't hit this all the time indicates it is purely based on timing and the random nature of page allocations. A few changes were made to better handle these kinds of low memory situations: - Any time we dip into the kernel's private stash of free pages we will force memory periodics to run. This should free up some pages so that the kernel's stash can remain full. - The size of the kernel's private page stash is increased from 4 to 10. We need this larger number due to the increased layers involved in paging code in/out with secureboot enabled. - The LRU algorithm was modified to keep pages around longer. A page will not be evicted under normal circumstances until we go 10 refresh cycles without an access (up from 3). This should help balance out the increased number of forced periodics that could now come into play. Change-Id: I289d866270db9d70cac275d8c98ab185caaade05 CQ: SW458623 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/73189 Reviewed-by: Nicholas E. Bofferding Reviewed-by: Matt Derksen 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

Add more agressive memory allocation calls

2019-02-12T22:24:46+00:00

If we hit a situation where we can't allocate a page of memory, this will force a coalesce (defrag) a few times and then eventually trigger other memory reclamation actions. Also tweaked a few spots in the kernel to enhance debug: - add more stops to look at errors (HB_BREAK_ON_ERROR) - add more backtrace calls - add a new debug flag to count the extra coalesce calls Change-Id: Ibac7079a44a12dc61e41304de4c4ae518c206d13 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/71653 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Reviewed-by: Matt Derksen Reviewed-by: Nicholas E. Bofferding Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Daniel M. Crowell

Add retry to slave core wakeup path

2019-01-25T16:07:27+00:00

We are still seeing some very intermittent errors in the slave core wakeup path. It still seems like we may have a timing issue. Until we figure out exactly what is going on, I am adding a retry mechanism that should get the core to report in correctly. The retry is done by issuing an additional doorbell message to the core that didn't report in. Change-Id: Ib87e5d58e079674d1eebb44c10d0252a35ea0519 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/70761 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Nicholas E. Bofferding Reviewed-by: Dean Sanner Reviewed-by: William G. Hoffa Reviewed-by: Daniel M. Crowell

Close windows where Hostboot HRMOR is not available to FSP

2018-10-12T21:13:39+00:00

There are certain small time windows where the core scratch register, which holds the Hostboot load address for FSP to read in order to dump Hostboot, is cleared during winkle. This change caches the Hostboot load address in a processor attribute that FSP can read/use in the event it happens to request a dump in the window where the core scratch register is not valid. Change-Id: I776df5e64f4d8ad382392662d08ee349661ba78a CQ: SW444212 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/66664 Reviewed-by: Michael Baiocchi Reviewed-by: Ilya Smirnov 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

Support HB running in SMF

2018-09-24T17:17:16+00:00

Support SMF for P9N/P9C. Lots of minor tweaks to make this work, but the biggest is to run userspace in problem state This is needed because for SMF Hostboot will need to run in S=1, HV=0,PR=1 (and kernel in S=1, HV=1, PR=0) This commit makes P9 HB userpsace run in HV=0 PR=1 and kernel in HV=1, PR=0. Change-Id: Ia4771df5e8858c6b7ae54b0746e62b283afb4bc4 RTC: 197243 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/50530 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Dean Sanner Reviewed-by: Nicholas E. Bofferding Reviewed-by: Christian R. Geddes Reviewed-by: Daniel M. Crowell

Make processing of hrmor value in MemStateInfo consistent

2018-08-01T18:34:25+00:00

There was some confusion on whether the value in core scratch 1 that told us where Host HRMOR lives was in MB or just Bytes. Going forward this value will be in bytes. CQ: SW440425 Change-Id: I5e7e5efdd584c3a8c050091b32d5443dff3cd417 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/63627 Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Tested-by: FSP CI Jenkins Tested-by: Jenkins OP HW Reviewed-by: Nicholas E. Bofferding Reviewed-by: Daniel M. Crowell

Restore Timebase on Master Core Threads 1-3 after Sleep/Winkle

2018-07-24T14:39:23+00:00

Change-Id: I329dd64345f2474cb0dad628ccc2244d85be86c2 CQ: SW429364 Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/63147 Reviewed-by: ILYA SMIRNOV Tested-by: Jenkins Server Tested-by: Jenkins OP Build CI Reviewed-by: Prachi Gupta Reviewed-by: Christian R. Geddes Tested-by: Jenkins OP HW Tested-by: FSP CI Jenkins Reviewed-by: Daniel M. Crowell