mctp-demux-daemon: Fix cleanup path on socket_init() failure

Now that we have potentially set up packet capture, make sure to clean
up the packet capture resources if we get a failure during socket

mctp-demux-daemon: Fix cleanup path on socket_init() failure

Now that we have potentially set up packet capture, make sure to clean
up the packet capture resources if we get a failure during socket
initialisation.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I138cc24509bc6009ad0f3adda282e045a099984c

show more ...

mctp-demux-daemon: Add packet capture option

Optionally enable libpcap support in mctp-demux-daemon to capture
packets both from the Unix domain socket and binding interfaces.
Providing the two capt

mctp-demux-daemon: Add packet capture option

Optionally enable libpcap support in mctp-demux-daemon to capture
packets both from the Unix domain socket and binding interfaces.
Providing the two capture points allows for tracking down issues with
packets being dropped during binding initialisation.

As there's no formal linktype defined for MCTP or higher-level DMTF
protocols command-line switches provide the ability to specify one of
the private linktype values in the range 147-162.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I593d9e4be80c0198e643758f216e774169668a8c

show more ...

core: Add infrastructure for packet captures

A callback and user data member is defined to abstract out the
implementation details of capturing the packets. The captured packets
are in DSP0236 form,

core: Add infrastructure for packet captures

A callback and user data member is defined to abstract out the
implementation details of capturing the packets. The captured packets
are in DSP0236 form, i.e. no binding metadata is present.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Id8234d6e8f7856c70a0ca3fe5d4cee38ec11bd8c

show more ...

mctp-demux: Add signalfd support

This allows us to clean up resources when the process is terminated.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ia8ba75394490f0577bac7d860dbdb4f92e6

mctp-demux: Add signalfd support

This allows us to clean up resources when the process is terminated.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ia8ba75394490f0577bac7d860dbdb4f92e672d2d

show more ...

Add compiler flags to enable dev testing

This just adds compiler flags to CMakeLists.txt to be on par
with the autotools.

Tested: Manually invoked a developer build and verified

Signed-off-by: Sum

Add compiler flags to enable dev testing

This just adds compiler flags to CMakeLists.txt to be on par
with the autotools.

Tested: Manually invoked a developer build and verified

Signed-off-by: Sumanth Bhat <sumanth.bhat@linux.intel.com>
Change-Id: Ibb129c2055ad57af57b4bca4ac81590afd994f8d

show more ...

core: Handle exhaution of message contexts

If message contexts are exhausted, there is a possibility of
NULL pointer dereference. Close this attack surface with a
check on the message context creati

core: Handle exhaution of message contexts

If message contexts are exhausted, there is a possibility of
NULL pointer dereference. Close this attack surface with a
check on the message context creation status.

Signed-off-by: Sumanth Bhat <sumanth.bhat@linux.intel.com>
Change-Id: I208349b74a27a14b180c976bd78ec81a7991e43f

show more ...

core: Handle arithmetic overflow in adding new fragments

Large fragment sizes can cause arithmetic overflows and can
cause memory corruptions. However, this condition will not be hit
with a fragment

core: Handle arithmetic overflow in adding new fragments

Large fragment sizes can cause arithmetic overflows and can
cause memory corruptions. However, this condition will not be hit
with a fragment size check in place and with a sane memory allocator
in place.

Adding this check to ensure that we have defense in depth.

Signed-off-by: Sumanth Bhat <sumanth.bhat@linux.intel.com>
Change-Id: Iaa976ce636c1eb617afa75e852b39c50699ebeb2

show more ...

core: Handle MCTP fragment sizes

Message assembly can be terminated if case we receive a
middle/end packet of unexpected size. This provision is
provided in DSP0236 v1.3.1 section 8.8 incorrect tran

core: Handle MCTP fragment sizes

Message assembly can be terminated if case we receive a
middle/end packet of unexpected size. This provision is
provided in DSP0236 v1.3.1 section 8.8 incorrect transmission
unit.

Reception of middle packets whose size is not equal to
start packet and end packets whose size is not less than
or equal to the start packet causes message assembly
termination.

Signed-off-by: Sumanth Bhat <sumanth.bhat@linux.intel.com>
Change-Id: I6371ab9e22e2c8ece70a9480f224de3f1f2f184e

show more ...

mctp-demux-daemon: Use systemd socket activation

Take advantage of lazy initialisation.

However, this also allows (one) daemon providing services over MCTP
(pldmd) an opportunity to start prior to

mctp-demux-daemon: Use systemd socket activation

Take advantage of lazy initialisation.

However, this also allows (one) daemon providing services over MCTP
(pldmd) an opportunity to start prior to MCTP interfaces coming up. This
is a stunted way to provide capabilities that might be assumed by other
components (the host) without implementing some required messages from
the MCTP standard.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I4b1c21f4fd42f84e2c85a453570a74330fc70ecf

show more ...

systemd: Own service startup ordering with respect to pldmd

We're about to turn the dependencies around so mctp-demux-daemon can be
socket-activated by systemd. Move ownership of the start-up sequen

systemd: Own service startup ordering with respect to pldmd

We're about to turn the dependencies around so mctp-demux-daemon can be
socket-activated by systemd. Move ownership of the start-up sequencing
between mctp-demux-daemon and pldmd into libmctp so the switch to
socket-activation is transparent to pldmd.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ia094d7ce32eabccf86d6a53a4a74b7deea79488d

show more ...

libmctp: Add subdirectory for systemd metadata

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I0989331aff807928fc23a4b298ef97c491714093

core: Check Null pointer in mctp_bridge_busses()

Handle memory allocation failures in mctp_bridge_busses to
reduce the likelihood of a possible attack.

Change-Id: Ib8edc94ae5942dc1c0b2355aeb3642d07

core: Check Null pointer in mctp_bridge_busses()

Handle memory allocation failures in mctp_bridge_busses to
reduce the likelihood of a possible attack.

Change-Id: Ib8edc94ae5942dc1c0b2355aeb3642d079512537
Signed-off-by: Helen Huang <he.huang@linux.intel.com>

show more ...

configure: Ensure udev rules are installed in rules.d

udevdir is set to the udev root, e.g. /lib/udev, while rules need to
live in the rules.d subdirectory.

Signed-off-by: Andrew Jeffery <andrew@aj

configure: Ensure udev rules are installed in rules.d

udevdir is set to the udev root, e.g. /lib/udev, while rules need to
live in the rules.d subdirectory.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I9416eb7e5a99a4bf8391a4ba135c26f43a79822c

show more ...

astlpc: Ship udev rules to help with device and driver transition

Currently the OpenBMC kernel has a hacky device driver at
drivers/char/misc/mctp-lpc.c that only works against KCS channel 4 on
Aspe

astlpc: Ship udev rules to help with device and driver transition

Currently the OpenBMC kernel has a hacky device driver at
drivers/char/misc/mctp-lpc.c that only works against KCS channel 4 on
Aspeed BMCs. This hacky driver exposes an `mctp0` device node.

The intent is to replace it with one that is slightly less
hacky at drivers/char/ipmi/kcs_bmc_cdev_raw.c[1], though eventually this
too will be replaced by an in-kernel implementation of the entire LPC
MCTP binding.

kcs_bmc_dev_raw.c works against all KCS devices in both Aspeed and
Nuvoton BMC SoCs.

We need to move away from KCS channel 4 on the Aspeed BMCs as the way
the hardware initialises the status bits conflicts with the protocol
definition for the LPC MCTP binding[2].

Kill two birds with one stone by shipping udev rules that accomodate the
new device node names that come with kcs_bmc_dev_raw.c, and symlink them
upon creation to /dev/mctp0 to emulate the original mctp-lpc.c
interface.

[1] https://lore.kernel.org/openbmc/20210510064955.1704652-1-andrew@aj.id.au/T/#me02d9a0e6ce82c9e0602cd615ccdad66d242af87
[2] https://github.com/openbmc/libmctp/blob/eba19a3b122a39ef2b5dbda49b418a202f78a48d/docs/bindings/vendor-ibm-astlpc.md#kcs-status-register-layout

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I21446936cea6c5e4024ec6536ad5983dd360deb2

show more ...

astlpc: Introduce protocol v3 with integrity checks

v3 of the binding adds a CRC-32 value as a medium-specific trailer to
each packet passing over the binding interface.

The patch includes a naive

astlpc: Introduce protocol v3 with integrity checks

v3 of the binding adds a CRC-32 value as a medium-specific trailer to
each packet passing over the binding interface.

The patch includes a naive bit-shift implementation of CRC-32, we can
improve it later as necessary.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I93a95bccef30010d56e10e29b6d84554268ab7af

show more ...

libmctp: Split padding in pktbuf into header and trailer

The astlpc binding will shortly make use of a medium-specific trailer to
implement a CRC-32 integrity check over packet data. Allow for
mediu

libmctp: Split padding in pktbuf into header and trailer

The astlpc binding will shortly make use of a medium-specific trailer to
implement a CRC-32 integrity check over packet data. Allow for
medium-specific trailer data in the pktbuf in order to remove the need
for a bounce-buffer. The binding will trim the pktbuf to the payload
length before passing the instance up the stack.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ib215a552ff218b007f7f119c5235fbc75d9ca6cb

show more ...

astlpc: Make packet properties depend on protocol version

The astlpc binding will shortly have an implementation of the v3
protocol specification. v3 adjusts the medium-specific packet size to
inclu

astlpc: Make packet properties depend on protocol version

The astlpc binding will shortly have an implementation of the v3
protocol specification. v3 adjusts the medium-specific packet size to
include a CRC-32 in a packet trailer. Implementing v3 must not impact
the behaviour of earlier protocol versions, so provide an ops struct to
handle version-specific details.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Iab04defa76b57a592442189e6ade58f25ed1d9ae

show more ...

docs: bindings: Add a Makefile

Makes local preview of changes easier to manage.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I67be4b51664c1b31ab183d2e2f9c63faaf34b919

astlpc: Fix whitespace in ops struct

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: If64467bb99f480cddc0c315ef97069c0ab18d778

docs: vendor-ibm-astlpc: Define v3 of the protocol

v3 adds a CRC-32 integrity check across each packet exchanged via the
binding.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ie9b798e

docs: vendor-ibm-astlpc: Define v3 of the protocol

v3 adds a CRC-32 integrity check across each packet exchanged via the
binding.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ie9b798e1076e099525ebf7bcfe5d8ea81367ae3a

show more ...

docs: vendor-ibm-astlpc: Tweak description of header magic

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ied2a9d85c63d44bb227385f3ab981b44e8498d80

docs: vendor-ibm-astlpc: Style code snippets as C

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I95a31775cef4f54e37a94cde5185493d3427f586

docs: vendor-ibm-astlpc: Fix IDs in packet transmission sequences

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I974a1c72c57f899d297e7cb71e4dbec5ca325ea4

docs: vendor-ibm-astlpc: Increase prominence of endianness

Describe this before we get distracted by other details of the LPC FW
window.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I

docs: vendor-ibm-astlpc: Increase prominence of endianness

Describe this before we get distracted by other details of the LPC FW
window.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ibe2c17cdc60631a4402fd182be64d636c6835605

show more ...

docs: vendor-ibm-astlpc: Annotate {M,L}SB in Bit column

Tidy up alignment of the descriptions.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ie15ef81ed337bf183e6367e0f92547aa5bc2fd1a