resolve stricter warnings

In order to convert this repository to Meson, we need to make it
compile under `warning_level=3`. Fix a number of warning classes
across the repository or disable them.

S

resolve stricter warnings

In order to convert this repository to Meson, we need to make it
compile under `warning_level=3`. Fix a number of warning classes
across the repository or disable them.

Some fixes are:

* Add missing header files.
* Fully initialize structs as necessary.
* Add `__attribute__((unused))` on parameters as necessary.
* Fix comparisons between signed and unsigned.
* Fix printf specifiers as necessary.
* Avoid case-fallthrough.
* Remove if conditions which are always true.

Some warnings would require extensive code changes, due to their
pervasive use, and so are disabled at a per-file level:
* `-Wpointer-arith`
* `-Wunused-result`

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: If8992b9108f12b39f796ed090ba29868c9f3c627

show more ...

mboxd: Don't require mbox

With the DBus interface to the daemon in place it's possible to
implement any hardware interface in terms of the DBus interface. This is
the strategy used to support the IP

mboxd: Don't require mbox

With the DBus interface to the daemon in place it's possible to
implement any hardware interface in terms of the DBus interface. This is
the strategy used to support the IPMI HIOMAP transport. Further, the use
of MBOX is deprecated due to security concerns. We want to drop the
driver from the kernel, which failed to get traction upstream, so make
support optional.

As a consequence, switch the default transport to DBus.

Change-Id: I9f16ca053ce48943dce59b83ca991ec5494580d8
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

mboxd: Add backend DBus interface and commandline options

Also implement a backend commandline option to mboxctl: `mboxctl
--backend ...`, to allow easy run-time switching of the backend from the
co

mboxd: Add backend DBus interface and commandline options

Also implement a backend commandline option to mboxctl: `mboxctl
--backend ...`, to allow easy run-time switching of the backend from the
commandline.

Switching between VPNOR and file backends via mboxctl was tested on
Witherspoon, and MTD and file backends on Romulus.

Change-Id: Iaf0e27ecf1d5cdd9e3a31729fb179096bbc37408
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

mboxd: Remove flash API compatibility shim

The flash API compatibility was kept to reduce the line noise in the
previous backend patch. Remove the compatibility layer now and convert
the remaining c

mboxd: Remove flash API compatibility shim

The flash API compatibility was kept to reduce the line noise in the
previous backend patch. Remove the compatibility layer now and convert
the remaining call-sites.

Change-Id: I4b6e54f4463059a7804918add81e7572db7b7c21
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

mboxd: Add a backend abstraction layer to mboxd.

Introduce a backend abstraction, enabling multiple implementations to be
compiled in at once. This change formally abstracts the two existing
backend

mboxd: Add a backend abstraction layer to mboxd.

Introduce a backend abstraction, enabling multiple implementations to be
compiled in at once. This change formally abstracts the two existing
backends, mtd and vpnor.

With the backend abstraction in place, subsequent backends are easier to
implement.

This change is based of Evan's work and he retains authorship credit. I
(AJ) have reworked the patch to pass the vpnor tests, refactored some
parts to enable broader use of const structures and others to clarify
the initialisation sequences.

Due to the existing lack of abstraction the patch has unfortunately
wide-ranging impacts. I've whittled it down as much as I consider
reasonable.

Change-Id: I29984a36dae4ea86ec00b853d2a756f0b9afb3ec
Signed-off-by: Evan Lojewski <github@meklort.com>
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

flash: Introduce flash_validate()

Clean up the protocol_negotiate_version() mess. The existing approach
came about due to viewing the vpnor implementation as an edge case in
its own right. The code

flash: Introduce flash_validate()

Clean up the protocol_negotiate_version() mess. The existing approach
came about due to viewing the vpnor implementation as an edge case in
its own right. The code becomes much neater if we consider all backends
as equal and afford them the callbacks necessary for correct behaviour.

Change-Id: Ifaeee9da459818cf22b2f137ddc5b8d0356b9be9
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

test/flash_copy: Configure MTD device to avoid undefined behaviour

Otherwise we observe invalid memory accesses due to uninitialised
variables.

Change-Id: I8b9063ccc9a25b225a562ebe120f2a99a28788ca

test/flash_copy: Configure MTD device to avoid undefined behaviour

Otherwise we observe invalid memory accesses due to uninitialised
variables.

Change-Id: I8b9063ccc9a25b225a562ebe120f2a99a28788ca
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

test/mbox: Assert that the flash fd is sane

Change-Id: I8f08db1b0ae19197c3fb8c4053deab900154a125
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

test/mbox: Use assertions for error handling

No need to try to exit gracefully in the test cases.

Change-Id: Id558c5201c08bdb0b34859cb3af1a0efa1a2809b
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

test: Add windows_equally_evictable

Tests for the condition where the minimum age is less than any window's
current age, in which case we can get a NULL dereference if the
windows_reset_all() implem

test: Add windows_equally_evictable

Tests for the condition where the minimum age is less than any window's
current age, in which case we can get a NULL dereference if the
windows_reset_all() implementation is broken.

windows.c:409:23: runtime error: member access within null pointer of type 'struct window_context'
ASAN:DEADLYSIGNAL
=================================================================
==31400==ERROR: AddressSanitizer: SEGV on unknown address 0x00000004 (pc 0x0002b658 bp 0x74c00270 sp 0x7eb7c678 T0)
==31400==The signal is caused by a WRITE memory access.
==31400==Hint: address points to the zero page.
#0 0x2b657 in window_reset windows.c:410
#1 0x2cc9b in windows_create_map windows.c:572
#2 0x1f3f3 in protocol_v1_create_window protocol.c:167
#3 0x2121b in protocol_v2_create_window protocol.c:417
#4 0x24cd7f in generic_vpnor_create_window vpnor/protocol.cpp:51
#5 0x24d053 in protocol_v2_vpnor_create_window vpnor/protocol.cpp:63
#6 0x2663b in mbox_handle_create_window transport_mbox.c:282
#7 0x276db in handle_mbox_req transport_mbox.c:568
#8 0x276db in transport_mbox_dispatch transport_mbox.c:649
#9 0x17fcb in poll_loop mboxd.c:185
#10 0x17fcb in main mboxd.c:423
#11 0x46b68517 in __libc_start_main (/lib/libc.so.6+0x46b68517)

AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV windows.c:410 in window_reset
==31400==ABORTING

Change-Id: I8161e2ea17953e196d4bb3ca90d19e44ec10c86d
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

daemon: Remove the concept of triggering or suppressing events

Rather, trigger them (or not) as necessary in the relevant code-paths.
This ensures that any call to one of protocol_events_{set,clear}

daemon: Remove the concept of triggering or suppressing events

Rather, trigger them (or not) as necessary in the relevant code-paths.
This ensures that any call to one of protocol_events_{set,clear}()
actually has a consequence that we can set about dealing with in the
transport layer.

Change-Id: If64733fa53ed9def0da8330c99cbe48327bab934
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

transport: Switch transports as required on GET_MBOX_INFO

Also flush the event state out via the new transport. This must be done
after the command has completed for the mbox transport.

Change-Id:

transport: Switch transports as required on GET_MBOX_INFO

Also flush the event state out via the new transport. This must be done
after the command has completed for the mbox transport.

Change-Id: I251fb949ae67a477288d0d57a1a6afe5b2af5f8f
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

mboxd: Refactor and rename mbox.h to mboxd.h

Refine the purpose of the header file to represent what's required for
the daemon itself, not its constituent pieces. Rather, split those
definitions out

mboxd: Refactor and rename mbox.h to mboxd.h

Refine the purpose of the header file to represent what's required for
the daemon itself, not its constituent pieces. Rather, split those
definitions out to their respective header files and include them as
necessary.

Finally the header file is renamed to better reflect its purpose.

Change-Id: I48c409f57d96c844589cd865b24f197477dfe87c
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

transport: mbox: Rename *init_mbox_dev to *transport_mbox_init

Change-Id: I212277ed2462089b905546ef048308d24a5489d0
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

transport: mbox: Rename dispatch_mbox to transport_mbox_dispatch

Change-Id: I7c5457099e90ea4ed498fd1898b4504f051d13fc
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

protocol: Provide abstraction over event notification

How this works will be transport-dependent. Move the event notification
helpers into the protocol abstraction and call-back through the
register

protocol: Provide abstraction over event notification

How this works will be transport-dependent. Move the event notification
helpers into the protocol abstraction and call-back through the
registered flush handler as necessary.

Change-Id: I29e3a9a9785b92de46a2b2750257fb7f8480a184
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

vpnor: Hijack protocol rather than transport

By hijacking the transport the changes in behaviour were limited to the
mailbox interface. Now that we have a DBus interface as well this would
lead to i

vpnor: Hijack protocol rather than transport

By hijacking the transport the changes in behaviour were limited to the
mailbox interface. Now that we have a DBus interface as well this would
lead to inconsistent behaviour dependent on the transport.

Instead of hooking the transport, push the hook down to the protocol
level where we will achieve consistent behaviour across all transports.

Change-Id: I437866a6dbda107149336c15a00ee1aa058f5875
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

show more ...

protocol: Add get_info

Change-Id: Ie3338714813bb65f5d37fcd046dd5bebc0ba21f0
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

mboxd: Rename mboxd_msg to transport_mbox

Change-Id: I0f4e206648c02a98fa938a7e582470c8d90904f1
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

windows: Rename init_windows to windows_init

Change-Id: Icdee1694b230d3c03075b840ec31450a236ab9e1
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

mboxd: Rename windows files

Change-Id: I44e110906fd0f751f8f34b175f971e65896a7301
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

lpc: Rename init_lpc_dev to lpc_dev_init

Change-Id: I1b3d8f4b561fda13dee8c9c0460aa463d6200e1d
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

mboxd: Rename lpc source files

Change-Id: I289618c90b2df8911e362397b7eab99403c86c05
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

flash: Rename write_flash to flash_write

Change-Id: I4b41aac597299c93369d8998b03a7d3e2a128742
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>

flash: Rename erase_flash to flash_erase

Change-Id: I417319aadb2fdbfa2d688bb26b5515d2f9265994
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>