summaryrefslogtreecommitdiffstats
path: root/procedures
Commit message (Collapse)AuthorAgeFilesLines
* cleanupPCIE: Catch file_error::Open exceptionAnthony Wilson2019-04-011-13/+22
| | | | | | | | | | | | | | | | | The bus that is read may only show up at certain times, such as during poweron. It is not a fatal error condition if the bus is not available. So we will let it get logged to the journal, but not let it get committed in this case. Tested: It was an intermittent failure, but ran poweron/poweroff a few times and didn't see the error pop up again. Resolves: openbmc/openbmc#3510 Change-Id: I47629e5f27fb847aa0094f0757b1988f8e645ebc Signed-off-by: Anthony Wilson <wilsonan@us.ibm.com>
* Don't create errors in cleanupPCIEMatt Spinler2019-02-051-1/+10
| | | | | | | | | | This code always just runs in the power off path, and at times it may be called when power is already off. In that case the FSI access would fail since the processor won't have power, so just catch the error and continue on. Change-Id: Ic02b17875763b0540edaaec47ee19846f305db72 Signed-off-by: Matt Spinler <spinler@us.ibm.com>
* In startHost, don't set bit 2 in scratch reg 8Matt Spinler2018-11-261-3/+0
| | | | | | | | | | This was keeping hostboot from using some default scratch register settings. Tested: Tested by HW team Change-Id: I1290acbabfa785e67a6038a3aadb7ececd9e1f7a Signed-off-by: Matt Spinler <spinler@us.ibm.com>
* p9: cleanupPcie: Run procedure on all CPUsAnthony Wilson2018-11-081-3/+5
| | | | | | | | Procedure should be run on all of the master and slave CPUs. Change-Id: I323519b059cfbe14ccd6145e22b1bdae4fd3d476 Signed-off-by: Anthony Wilson <wilsonan@us.ibm.com>
* p9: cleanupPcie: Correct line endingsAnthony Wilson2018-11-071-51/+51
| | | | | Change-Id: Ia8c516ca1a4ceca46a88c9de72cd765e3b2884c2 Signed-off-by: Anthony Wilson <wilsonan@us.ibm.com>
* add .clang-formatPatrick Venture2018-11-027-109/+99
| | | | | | | Add .clang-format for automatic style. Change-Id: I6d240009370179b5b8f1f646b0476a059ec6aa85 Signed-off-by: Patrick Venture <venture@google.com>
* s/copyright char/(C)/g to make compiler happyPatrick Venture2018-11-026-8/+7
| | | | | | | Compiler complaining about the copyright characters. Change-Id: I6730a50a668899a33e9dfcc4accb67ee207a3f08 Signed-off-by: Patrick Venture <venture@google.com>
* Add cleanupPcie to openpower-proc-controlAnthony Wilson2018-10-261-0/+50
| | | | | | | | | | | | | | This disables the drivers and receiver in the PCIE root control 1 register as follows: BIT NAME VALUE 19 TP_RI_DC_B 0b0 20 TP_DI1_DC_B 0b0 21 TP_DI2_DC_B 0b0 This should be run on a power off. Change-Id: I6e027260f78a3fc451a45832f6f9bcf9afc8c3b9 Signed-off-by: Anthony Wilson <wilsonan@us.ibm.com>
* Log istep info on watchdog timeoutspashabk-in2017-10-251-0/+150
| | | | | | | | 1. SBE istep info (cfam 2809) 2. HB istep info (cfam 283C) Change-Id: I67632c58fe1148b980791dc017e5c099107caae5 Signed-off-by: Shakeeb Pasha <shakeebbk@in.ibm.com>
* Add FSI scan procedureMatt Spinler2017-10-041-0/+133
| | | | | | | | | | | | | | | | | This procedure will perform both the FSI master and FSI hub scans. It replaces doing the scans by writing to sysfs using the echo command so that real errors can be created on failures. Currently the only way to tell if the master scan fails is by checking for missing sysfs files, and this code will do that by checking for the hub scan file. Change-Id: Id4b91592b8c8b9a5fc9f1a56f4d89e142a6c6b74 Signed-off-by: Matt Spinler <spinler@us.ibm.com>
* Use all phosphor-dbus-interfaces errorsMatt Spinler2017-10-041-1/+2
| | | | | | | | | | Remove the local error definitions and use the ones in phosphor-dbus-interfaces instead as the ones there suit our purposes and it simplifies this repository to do so. Change-Id: I18428c496914153270ecb181e7193acd0e386e97 Signed-off-by: Matt Spinler <spinler@us.ibm.com>
* Boot from sbe side 0 when boot count is greater than 0Andrew Geissler2017-08-181-1/+1
| | | | | | | | | | | | | | | Recent code changes now decrement the boot count at the start of every boot. The sbe side select design is to boot from side 0 twice, then side 1 on the last attempt. Now that the boot count is decremented at the start of the boot (and the total boot count is 3), the software needs to use side 0 when the boot count is greater than or equal to 1. Resolves openbmc/openbmc#2169 Change-Id: Ic08bed34d58de3d40b742ce52ca83627976d8fee Signed-off-by: Andrew Geissler <andrewg@us.ibm.com>
* Don't run vcs workaround on non-dd10 chipsAndrew Geissler2017-06-091-0/+12
| | | | | | | | | | | | | | | | The VCS workaround is only required (and only make sense) on P9 DD1.0 based systems. This code does a check of the chip level and does not run the workaround if the chip is not DD1.0. The code will continue to run some of the other workarounds because although not required for non-dd10 chips, they do no harm. Resolves openbmc/openbmc#1695 Change-Id: I1409ca359ccff7b78a186211e4cd447cd753eda7 Signed-off-by: Andrew Geissler <andrewg@us.ibm.com>
* Set SBE seeprom boot sideAndrew Geissler2017-06-011-1/+20
| | | | | | | | | | | | | | The default side for the SBE to boot from is side 0, which is indicated by a 0 in bit 17 of cfam 0x2808. When the boot count goes to 1 (the last before giving up), the start_host logic will switch over to side 1 for the SBE to boot from. Resolves openbmc/openbmc#1467 Change-Id: I61aa22939baa4cde38c8716429b6ca55f7c850bd Signed-off-by: Andrew Geissler <andrewg@us.ibm.com>
* Replacing runtime exceptions with elogDhruvaraj Subhashchandran2017-05-171-2/+9
| | | | | | | Resolves openbmc/openbmc#1323 Change-Id: Ia93d1f0036341c26fdc6c5e3133d4a05346090be Signed-off-by: Dhruvaraj Subhashchandran <dhruvaraj@in.ibm.com>
* P9 CFAM register override procedureMichael Tritz2017-04-142-2/+70
| | | | | | | | | This new procedure will allow users to provide P9 CFAM register overrides before the P9 SBE is started. Change-Id: If3658f9a6ede9a3682c4ed7888e9acb328c67709 Signed-off-by: Michael Tritz <mtritz@us.ibm.com>
* New procedure to set synchronous FSI clock modeMatt Spinler2017-03-301-0/+48
| | | | | | | | | | This procedure is required when switching to external cronus mode which needs synchronous FSI clock mode set as opposed to the asynchronous mode required by the Open FSI driver. Change-Id: I509d113bd5cdadedb64b0275628f95572c21b0a3 Signed-off-by: Matt Spinler <spinler@us.ibm.com>
* Register proceduresMatt Spinler2017-03-142-0/+5
| | | | | | | | | | For a procedure to be available to run, it needs to have a call to a REGISTER_PROCEDURE macro. This macro wraps a call to a Register class that adds the procedure to the list along with the name to call it. Change-Id: I20d02e8f004c1c726228469465ae89b60ee52d66 Signed-off-by: Matt Spinler <spinler@us.ibm.com>
* Move procedures into separate files.Matt Spinler2017-03-142-0/+141
Procedures will now be in procedures/<chip>/. The next commit will have the procedures register themselves so they won't be in a hardcoded list. Change-Id: I0ff90afe2b51aaff25b5c25bf87070c820323e30 Signed-off-by: Matt Spinler <spinler@us.ibm.com>
OpenPOWER on IntegriCloud