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

docs: vendor-ibm-astlpc: Fix packet exchange order of operations

We read data from the KCS channels to confirm the command type before
reading the packet data from the FW window.

Signed-off-by: And

docs: vendor-ibm-astlpc: Fix packet exchange order of operations

We read data from the KCS channels to confirm the command type before
reading the packet data from the FW window.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I19378b20c91183b5403f046cb7ab22d83c257e40

show more ...

docs: vendor-ibm-astlpc: Add definitions for KCS registers

These were explained in the document body but deserve their own
definitions given their importance to the protocol.

Signed-off-by: Andrew

docs: vendor-ibm-astlpc: Add definitions for KCS registers

These were explained in the document body but deserve their own
definitions given their importance to the protocol.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I2cdf60dda74efba95e6e78bd4a53aa9558308d75

show more ...

docs: vendor-ibm-astlpc: Use numbered list for strictly ordered ops

I'm not sure how we used a bullet list for a section titled "the
following sequence occurs".

Signed-off-by: Andrew Jeffery <andre

docs: vendor-ibm-astlpc: Use numbered list for strictly ordered ops

I'm not sure how we used a bullet list for a section titled "the
following sequence occurs".

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I28dbe579de62a77726c617dd7aad11239a23f2ff

show more ...

docs: vendor-ibm-astlpc: Reorder sections to improve flow

The new order tries to keep the concepts flowing from least-to-most
detailed.

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

docs: vendor-ibm-astlpc: Reorder sections to improve flow

The new order tries to keep the concepts flowing from least-to-most
detailed.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ib076c78f79aaedcb4b6c5c0e2606f8c2cbdd17a9

show more ...

docs: vendor-ibm-astlpc: Reduce emphasis for definitions

The aim is to be eye-catching but not over-bearing.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I4cde5f858ab6e26754c8b9d48df3

docs: vendor-ibm-astlpc: Reduce emphasis for definitions

The aim is to be eye-catching but not over-bearing.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I4cde5f858ab6e26754c8b9d48df3a96b07bd4545

show more ...

docs: vendor-ibm-astlpc: References as a numbered list

Make it easier to distinguish what text is which document at a glance.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ifdd69d9e3af

docs: vendor-ibm-astlpc: References as a numbered list

Make it easier to distinguish what text is which document at a glance.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ifdd69d9e3af3706c044e374984337609f9d70dde

show more ...

docs: vendor-ibm-astlpc: Improve readability with explicit line breaks

Separate the reference document URLs from the document names.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I0f7e

docs: vendor-ibm-astlpc: Improve readability with explicit line breaks

Separate the reference document URLs from the document names.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I0f7eff977f01c67f2757814ede81da07f7cbba43

show more ...

docs: vendor-ibm-astlpc: Remove intimidating header wording

Given the state of the rest of the document, stop trying to be super
formal.

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

docs: vendor-ibm-astlpc: Remove intimidating header wording

Given the state of the rest of the document, stop trying to be super
formal.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ibdff4bdbc927cfcb77a691344748ac384c039641

show more ...

docs: Rename astlpc binding to include vendor

It helps to identify who is driving development of the binding.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ifb4508701e2a511632ba50dcee0

docs: Rename astlpc binding to include vendor

It helps to identify who is driving development of the binding.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ifb4508701e2a511632ba50dcee027dc01a247fa4

show more ...

core: Make pkt_size and pkt_pad size_t

Fix bad signedness comparison:

```
core.c: In function ‘mctp_binding_set_tx_enabled’:
core.c:630:25: error: comparison of integer expressions of different sig

core: Make pkt_size and pkt_pad size_t

Fix bad signedness comparison:

```
core.c: In function ‘mctp_binding_set_tx_enabled’:
core.c:630:25: error: comparison of integer expressions of different signedness: ‘int’ and ‘long unsigned int’ [-Werror=sign-compare]
630 | if (binding->pkt_size < MCTP_PACKET_SIZE(MCTP_BTU)) {
| ^
cc1: all warnings being treated as errors
```

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Idb91b16efa43526ea8e86fff666d0151f2987247

show more ...

core: Fail bridge initialisation if a binding fails to start

The start() callback can return an error, so make sure we propagate this
to the caller of mctp_bridge_busses().

mctp_destroy() will clea

core: Fail bridge initialisation if a binding fails to start

The start() callback can return an error, so make sure we propagate this
to the caller of mctp_bridge_busses().

mctp_destroy() will clean up the memory allocated for the bridged
busses. Cleaning the bus memory up before returning from
mctp_bridge_busses() may leave a started binding in an unsafe state
state (as its bus pointer will be either invalid or NULL), so we'll
avoid doing that for now.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I84803ced41d106ac96e2bf94c90feb5eea10579f

show more ...

mctp-demux-daemon: Warn about failed messages

It's possible that messages might not be sent if they are routed to a
binding before that binding has completed its protocol initialisation.

Signed-off

mctp-demux-daemon: Warn about failed messages

It's possible that messages might not be sent if they are routed to a
binding before that binding has completed its protocol initialisation.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I1ddf6817fdef974b61dcb3ce0f166af94b4b0f16

show more ...

tests: astlpc: Exercise mctp_message_tx() before channel init

Attempting to send messages before channel init should intentionally
fail, not fail by OOMing the machine.

Signed-off-by: Andrew Jeffer

tests: astlpc: Exercise mctp_message_tx() before channel init

Attempting to send messages before channel init should intentionally
fail, not fail by OOMing the machine.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: If912f488d89f14e66e7bb18dbf3c929648117827

show more ...

astlpc: Initialise with at least a BTU-sized packet

The astlpc binding's MTU negotiation takes place during processing of
the `Channel Init` command, after which both sides update `pkt_size`
accordi

astlpc: Initialise with at least a BTU-sized packet

The astlpc binding's MTU negotiation takes place during processing of
the `Channel Init` command, after which both sides update `pkt_size`
according to the negotiated MTU. Despite this, let's not set a
potentially invalid MTU at binding construction.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ic6d3dba38102b81ff9b7c09865f359d31ee829f7

show more ...

astlpc: Warn on unexpected commands before binding has started

We should receive neither `Tx Begin` nor `Rx Complete` prior to
sending or receiving `Channel Init`. If we do, warn about it and don't

astlpc: Warn on unexpected commands before binding has started

We should receive neither `Tx Begin` nor `Rx Complete` prior to
sending or receiving `Channel Init`. If we do, warn about it and don't
proceed to process the unexpected message.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ib154436305afe291f2cc71bbb861edd8362e42e3

show more ...

core: Validate binding MTU before allowing transmission

Don't accept MTU values like 0, which will cause the packetisation loop
to fail in unacceptable ways (OOM).

Signed-off-by: Andrew Jeffery <an

core: Validate binding MTU before allowing transmission

Don't accept MTU values like 0, which will cause the packetisation loop
to fail in unacceptable ways (OOM).

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I00d563cd91f718e695ed28d2b9978fc38566c4ae

show more ...

core: Drop messages sent before binding enables transmission

Some bindings (astlpc) have a handshake that must occur before packets
can be sent across the bus, but after the binding instance has bee

core: Drop messages sent before binding enables transmission

Some bindings (astlpc) have a handshake that must occur before packets
can be sent across the bus, but after the binding instance has been
constructed. Similarly, some bindings (astlpc) require coordination of
resources between endpoints as part of regular operation, which
motivates the existence of the associated `tx_enabled` state.

If a binding invokes mctp_binding_set_tx_enabled() with a `true`
argument it implies that all required handshaking has been completed
(if it had not, we still would not be able to send packets).

Switch `tx_enabled` from a boolean to a ternary state machine to capture
the `constructed` state of a binding (before required handshaking has
completed). Further, error out if mctp_message_tx_on_bus() is invoked
with a binding in the `constructed` state, as it's possible the
information required to correctly packetise the provided message is
invalid.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: If4baf07ab5fe7edd7f522fac29e32fc5c2a6105d

show more ...

core: Introduce mctp_bus_destroy()

Tearing down a binding needs to free all packets that have been queued
but not yet transmitted.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ia4bff3

core: Introduce mctp_bus_destroy()

Tearing down a binding needs to free all packets that have been queued
but not yet transmitted.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: Ia4bff3243f80e1088c521b09a512584383deb1ad

show more ...

core: Set n_busses appropriately on failure to start binding

We were cleaning up the allocated memory but failing to reset the number
of allocated busses to zero. This makes it harder to correctly c

core: Set n_busses appropriately on failure to start binding

We were cleaning up the allocated memory but failing to reset the number
of allocated busses to zero. This makes it harder to correctly clean up
the MCTP context later in the lifecycle.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I392ef6d6f7ffa7af7bb33d1aedfc4968f1749f7b

show more ...

core: Remove Rube Goldberg sequence point in packet body assignment

The code isn't wrong* - if it were we certainly would have noticed - but
why it's *right* isn't something we want to be reasoning

core: Remove Rube Goldberg sequence point in packet body assignment

The code isn't wrong* - if it were we certainly would have noticed - but
why it's *right* isn't something we want to be reasoning about.

Convert the sequence point to a plain statement to make the ordering and
value state requirements obvious to the reader.

* the sequence point will order the write to p before mctp_rx() is
invoked, giving p the correct value at the time of invocation, but its
value after the return of mctp_rx() is undefined.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I68a1e35ba6a9b8daf89c4d5782037ed623ebbeb5

show more ...

astlpc: Fix version test before assignment

The bug leads to the wrong packet size being used for the first
negotiation.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I1419c5905b2815359

astlpc: Fix version test before assignment

The bug leads to the wrong packet size being used for the first
negotiation.

Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Change-Id: I1419c5905b28153594304d8fb5511c5c4782fe72

show more ...