The new subtree model brings the subtrees up from the openbmc-machines
layer.

Change-Id: I58a03ae1be374bc79ae1438e65e888375d12d0c0
Signed-off-by: Dave Cobbley <david.j.cobbley@linux.intel.com>
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>

The new subtree paradigm has moved the meta-bsp layers to the top level

Change-Id: Ic4f64c39465be699bed3649e125e027d17daf938
Signed-off-by: Dave Cobbley <david.j.cobbley@linux.intel.com>
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>

As part of the move to subtrees, need to bring all the import layers
content to the top level.

Change-Id: I4a163d10898cbc6e11c27f776f60e1a470049d8f
Signed-off-by: Dave Cobbley <david.j.cobbley@linux.intel.com>
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>

This adds the initial recipe for the sdeventplus package.

Tested:
    Package builds successfully and the output looks correct.

Change-Id: I506955205dc9bcfe75a83b76442f4857c74b99bb
Signed-off-by: William A. Kennington III <wak@google.com>

Add mtd-utils upstream patch that fixes a regression on the
mtd-utils tools such as ubinfo.

Details of the issue which affects mtd-utils 2.0.1 and 2.0.2:
http://lists.infradead.org/pipermail/linux-mtd/2018-June/081562.html

Upstream-Status: Accepted [http://git.infradead.org/mtd-utils.git/commit/0f833ac73ad631248826386e2918d8571ecf0347

]

(From OE-Core rev: 41356d2c86d85b199962c3024f25361a709d9180)

Signed-off-by: Adriana Kobylak <anoo@linux.ibm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

(From Yocto rev: ffb7fd233ce23af06b54bee5b934f7d01271da97)

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: Ic6e244369f5e86d3ee2411bf1f7566a6637622e3

Change-Id: I10692dea155377dcd8558ae57e69948e6995ded0
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>

Change-Id: I849e03f2334a8602e05058669158083ad158104f
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

Change-Id: I1c709f716bec80bed0c4dc5b16fa70c809c8cc64
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

Change-Id: I3aa64d678759bcba72bfc093f88d07c5a6bb91ae
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

Change-Id: I6622c666a193d8c17b88027d88790fda88e33b96
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

This bumps the kernel to incldue a commit that switches to the ColdFire
FSI driver.

Benjamin Herrenschmidt (1):
      arm: dts: aspeed: Romulus can use coprocessor for FSI

A note to developers: from here on you must use an initrd that includes
the ast-cf-fsi.bin firmware, or FSI will not function. See commit
'openpower: Add ColdFire firmware to initrd'.

Change-Id: I701381de714b5f200fc796a09a055ff4f0d14633
Signed-off-by: Joel Stanley <joel@jms.id.au>

The kernel has changed how the character devices are allocated by the
fsi driver. All of the in-tree users of fsi have been updated to use
this new system.

The changes include an optional feature improve the layout of userspace
API. This is controlled by the CONFIG_FSI_NEW_DEV_NODE option, which we
do not set yet.

Benjamin Herrenschmidt (14):
      fsi: sbefifo: Fix inconsistent use of ffdc mutex
      fsi: Add support for device-tree provided chip IDs
      fsi: master-ast-cf: Fix build warnings on 64-bit platforms
      fsi: master-ast-cf: Mask unused bits in RTAG/RCRC
      fsi: master-ast-cf: Rename dump_trace() to avoid name collision
      fsi: Add new central chardev support
      fsi: sbefifo: Convert to use the new chardev
      fsi: scom: Convert to use the new chardev
      fsi: Add cfam char devices
      fsi: Prevent multiple concurrent rescans
      fsi: scom: Fix NULL dereference
      fsi: sbefifo: Bump max command length
      fsi: occ: Convert to use the new chardev
      arm: dts: aspeed: Change how power9 chips are represented on OpenPower

Gustavo A. R. Silva (1):
      fsi: master-ast-cf: Fix memory leak

Joel Stanley (1):
      fsi: sbefifo: Remove unnecessary mutex unlock

Change-Id: I1201a7e6317776ab6bb4f4307b95d388defd66ee
Signed-off-by: Joel Stanley <joel@jms.id.au>

Change-Id: I7ab76637af5a601b92f3fbcd31f5c39af49f6f0d
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

sensorNamePattern is missing the Romulus sensor yaml config.
Add it by sensor_yaml_config.py:

 sensor_yaml_config.py -i config.yaml -o output.yaml -f

Tested: Verify the sensorNamePattern is added and it is the same as
        Witherspoon's yaml config.

Change-Id: I4a21b4bb896ec058b5ac1e40fbda934516b62929
Signed-off-by: Lei YU <mine260309@gmail.com>

Add cpu core, dimm temperature sensors.
The yaml is generated from romulus.rpt by sensor_yaml_config.py

 sensor_yaml_config.py -i config.yaml -o output.yaml -e -r romulus.rpt

Tested: Verify that core/dimm temperature sensors are returned by
        ipmitool sensor command.

Change-Id: Ic564233d93d6a6297eff6368c14632a9e85258e1
Signed-off-by: Lei YU <mine260309@gmail.com>

Add entityID and entityInstance to sensor config yaml, with the help
script from
https://github.com/mine260309/openbmc-tools/blob/dev-mine/leiyu/obmc-utils/sensor_yaml_config.py



Tested: Verify it passes the build.

Change-Id: Ie9e1861e8c8e2919b88dabc387d8c1234486df51
Signed-off-by: Lei YU <mine260309@gmail.com>

Sort the sensor yaml config using PyYAML by loading and saving the yaml.
This is a preparation for further changes of sensor configs.

Tested: Verify the generated sensor-gen.cpp is technically the same as
        before.

Change-Id: I1315072b36236bb2eff01eb5d1507641e7978b42
Signed-off-by: Lei YU <mine260309@gmail.com>

bmc-fru-config.yaml defines the frus from BMC.
Romulus has 3 fans.

Tested: Verify that fan frus are shown with ipmi fru list command.

Change-Id: I9cd5a93c4711a0f336058617403c6560547a6121
Signed-off-by: Lei YU <mine260309@gmail.com>

In phosphor-ipmi-host, channel.yaml sets the info about number of
available Ethernet channels.
Romulus has eth0 only.

Tested: Verify the generated channel-gen.cpp reflects this yaml
config.

Change-Id: I902f6b0adb6066ad2c4f1654a4614bbf012e36a5
Signed-off-by: Lei YU <mine260309@gmail.com>

Serve the mostly pre-compressed static webui content via nginx. If nginx
can't find a static file, proxy the URI back to the rest server.

A follow-up commit will remove code that serves webui from the
phosphor-rest-server.

Tested:
- The webui, rest server and bmcweb seem to be working as before.
- curl -H "Accept-Encoding: gzip" https://<witherspoon>/foo.html
  is served compressed content.
- curl https://<witherspoon

>/foo.html is served uncompressed
  content.

Change-Id: Ice443f00a467f3558f28d697261df4f23278f420
Signed-off-by: Deepak Kodihalli <dkodihal@in.ibm.com>

Bump the Nuvoton Linux kernel version to 4.17.14. This includes a fix
for the security vulnerability in #3359.

Tested: Built and ran on the NPCM750 EVB.

Change-Id: Ied49e5486ea84408de16e5b081238acfa3d00ae6
Signed-off-by: Benjamin Fair <benjaminfair@google.com>

The OpenBMC build system doesn't currently generate a flash image usable
by NPCM7XX BMCs, and the OpenBMC Nuvoton kernel doesn't have an MTD
SPI-NOR driver.

The MTD image generation build step is failing for the NPCM750 eval
board, so disable it and instead generate a rootfs cpio suitable for
loading from u-boot.

Tested: Built for evb-npcm750 and was able to netboot the board

Change-Id: I8b6275979d78b7d5657e77d25bd74f933c740ea8
Signed-off-by: Benjamin Fair <benjaminfair@google.com>

This was moved to STATIC_HOSTING and the default for it (on)
is what the IBM layer wants

ref: https://gerrit.openbmc-project.xyz/#/c/openbmc/bmcweb/+/11905/



Change-Id: I4da0bdacc3a36c2760058f60de446e97817514f6
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

If the name of the volume is passed to ubirmvol to be removed,
ubirmvol searches and stores the information for the available
UBI volumes, such as id, type, name, etc. If a volume is
deleted by another process while this search is in process,
ubirmvol would fail, which is the case when removing a PNOR
version since 2 separate systemd service files remove the
read-only and read-write volumes.

Instead pass ubirmvol the volume id instead of the name, so
that the ubirmvol doesn't need to do any searches and instead
knows which volume to delete.

Closes openbmc/openbmc#2863

Tested:
Verified a Delete operation does not fail with a
ubirmvol error leaving the rw volume behind. Ex:
Aug 01 16:43:03 witherspoon systemd[1]: Starting Remove UBIFS volume pnor-rw-5ac20d5d...
Aug 01 16:43:04 witherspoon systemd[1]: Starting Remove UBIFS volume pnor-ro-5ac20d5d...
Aug 01 16:43:05 witherspoon obmc-flash-bios[1458]: libubi: error!: cannot read "/sys/class/ubi/ubi8_0/type"
Aug 01 16:43:05 witherspoon obmc-flash-bios[1458]:         error 19 (No such device)
Aug 01 16:43:05 witherspoon obmc-flash-bios[1458]: ubirmvol: error!: cannot find UBI volume "pnor-rw-5ac20d5d"
Aug 01 16:43:05 witherspoon obmc-flash-bios[1458]:           error 19 (No such device)

Change-Id: Ib9aaeef4bfc268f3dac29288a469256bc771e195
Signed-off-by: Adriana Kobylak <anoo@us.ibm.com>

This is used by the ColdFire FSI driver. All OpenPower machines will be
switched over to this driver, so it makes sense for it to be in the
common layer.

Change-Id: Iccd9e793169564c78a295ff3a44b7397ba67c091
Signed-off-by: Joel Stanley <joel@jms.id.au>

Alexander Amelkin (2):
      mtd: spi-nor: fix options for mx66l51235f
      mtd: spi-nor: fix spi register address for AST2400

Joel Stanley (1):
      fsi: master-ast-cf: Don't set FLAG_SWCLOCK on ast2400

Resolves: openbmc/openbmc#3358
Change-Id: Id2e7b9ba04f891c2a56bde1a6037e4c87ebf624c
Signed-off-by: Joel Stanley <joel@jms.id.au>

Once upon a time this service was added to initalise the host flash
before starting the host firmware, otherwise some systems would not
boot.

 > commit 14f965d8
 > Author: Andrew Geissler <andrewg@us.ibm.com>
 > Date:   Thu May 11 09:54:04 2017 -0500
 >
 >     PNOR init workaround for P8 based systems
 >
 >     The PNOR chip requires an init to be done to it prior
 >     to starting the host.  This is a temporary fix to this
 >     issue until we get the mboxd function ported to P8.

Ever since e7b0fd56 ("Switch userspace to use MTD access by
default", October 2017) running pflash hs not done this, as it
access the flash through the kernel mtd interface.

A machine could chose use mboxd and the mbox protocol on P8 but it is
not required to fix this issue.

Change-Id: I5c2400063bf776018ef60f2642fb2629c96af0a7
Resolves: openbmc/openbmc#1598
Signed-off-by: Joel Stanley <joel@jms.id.au>

Drop the unused meta-virtualization layer.

Change-Id: I14591d81f5d82cad7126921e5978a8707e2221d6
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>

Remove meta-virtualization from any bblayer.conf.samples.  It was
originally added for the python-pyyaml and python-gevent recipes
but better maintained versions of these recipes are provided by
meta-python.

Before this can be done the python-gevent bbappend in the phosphor
layer must be removed.  The version scope was wildcarded to any version
but the patches don't apply to python-gevent_1.2.2, which becomes the
default from meta-python when meta-virt is removed.  Drop the bbappend.

Change-Id: I6658a163460e3c8e63dc7371ffff803a7b6915b5
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>

Change-Id: I8473b4c93915d57a513384c38137186353c052bd
Signed-off-by: Alexander Hedges <ahedges@ethz.ch>

Change-Id: Icf5868a967bc55e5d45a8b58c80bfb6a1aaf96bd
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

Change-Id: Id24b1c26d0e91172285f745926d6b0ff538d37f0
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

Change-Id: Ia57c246400656cee580f8fd5f249a05ad44c2715
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

In obmc-init.sh, there is "clean-rwfs-filesystem" that clean the rwfs
during boot, but it backs up the files in whitelist, so it is not a
*real* factory reset.

Add a new "factory-reset" option, and when it exists, erase the rwfs
without restoring files in whitelist.

Tested: Set openbmconce=factory-reset in u-boot env, and verify the rwfs
        is erased during boot without restore;
        Set openbmconce=clean-rwfs-filesystem in u-boot env, and verify
        that rwfs is erased with whitelist restored.

Change-Id: Ifdc76da46fe273e9c942f30444173e51f8106fcc
Signed-off-by: Lei YU <mine260309@gmail.com>

Change-Id: Ibbb72012ce64021300413fa4de7f591eb8ae69a0
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

Change-Id: Ia9b517ccea96d32d65f38eeb38cd132a3023ab66
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>

This includes changes up to and including the 4.17.14 stable release.
Notably for OpenBMC there are a number of fixes for handling malformed
squashfs partitions.

Resolves: openbmc/openbmc#3359
Change-Id: Ifda0efa9ad4f701777d789dcc68f9dad34d121d3
Signed-off-by: Joel Stanley <joel@jms.id.au>

Moves Palmetto to use the kernel API, and the ColdFire master driver.
This change bumps the kernel (with device tree changes) and the skeleton
repository for a corresponding op-hostctl (host_control.exe) change.

Palmetto is the only system in the tree using this code. Any out of tree
systems (such as other P8s) must switch to using the kernel interface as
well.

Skeleton changes:

Joel Stanley (1):
      hostctl: Move to kernel FSI interface

Linux changes:

Benjamin Herrenschmidt (2):
      ARM: dts: aspeed: Add Power8 CFAM description for use by Palmetto
      ARM: dts: aspeed: Palmetto system can use coprocessor for FSI

Change-Id: I43aada22c9d7b21dc24bd5aae3c4dcb77bdaaa8f
Signed-off-by: Joel Stanley <joel@jms.id.au>

This includes the 'pnorboot' tool for configuring the LPC firmware
mapping.

This is used by Palmetto instead of mboxd, as Palmetto only requires the
LPC mapping component. The mbox component causes interrupts to be
continually triggered on the host, causing unnecessary CPU load.

As Palmetto is the only in-tree user of this tool, the service file hard
codes the PNOR size. If we merge machines in the future this should be
configured by the machine layer.

Resolves: openbmc/openbmc#3357
Tested: Booted host on Palmetto, checked interrupt count
Change-Id: I87daafa5008e978afd3a5fc386b347b014624911
Signed-off-by: Joel Stanley <joel@jms.id.au>

Change-Id: I3e4f730965d0c9ec8743c91b40891310fdad0890
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>