Hello,
This patchset adds virtio specification for GPIO devices. It supports basic GPIO line operations, along with optional interrupts on them (in a separate patch).
This is an *alternative implementation* of the virtio-gpio specification, a different version of it was earlier posted by Enrico [1].
I took back V4 of the specification I posted earlier (on June 17th), to allow Enrico to come up with something that is robust and works for everyone (as he started this thing last year), but it didn't go as expected. I couldn't see any of my review comments being incorporated (or any intentions of them getting in ever) in the two versions Enrico posted and a month passed since then.
And so I am trying to present an alternative approach here to solve the problem, which I believe is more clear and robust. I am open to suggestions to improve it further.
I will let the virtio/gpio maintainers decide on its fate.
Key differences from Enrico's approach [1]:
- config structure is rearranged to remove everything apart from number of gpios and size of the gpio_names_field.
- Supports freeing of a GPIO line after use, we call them activate/deactivate now, which suits their purpose better.
- Interrupt implementation handled with feature bit 0. Either the interrupts are fully supported or not at all.
- All non-interrupt traffic happens on a single virtqueue, requestq. Interrupt traffic goes over eventq.
- Doesn't add any ordering restrictions on the device, it can respond earlier to the second request, while still processing the first one.
- Clearly state that two requests can't be initiated for the same line by device or driver.
Version history of the specification I sent:
V5 -> V6: - All non-interrupt traffic happens on a single virtqueue, requestq. Interrupt traffic goes over eventq now. - Many fields dropped from the config structure. - Separate message type to get gpio-names, added more description about how the names should be. - Much clearer message flows, both non-irq messages and irq-messages. - Parallelism supported for all type of messages, for different GPIO pins. - All references to POSIX errors dropped, just reply pass or fail. - request/free renamed to activate/deactivate, as that's what we will end up doing there, activate or deactivate a GPIO line. - General purpose IO as I/O or Input/Output instead. - Hopefully I didn't miss any review comments.
V4 -> V5: - Split into two patches, irq stuff in patch 2. - Use separate virtqueue for sending data from device/driver. - Allow parallel processing of requests for different GPIOs, only one request at a time for the same GPIO line. - Same goes for interrupt side, only one interrupt request per GPIO line. - Improved formatting in general. - Add new sections explaining message flow sequence.
V3 -> V4: - The GPIO line names must be unique within a device. - The gpio_names[0] field is dropped and gpio_names_offset field is added in place of the 2 bytes of padding. - New interrupts must not be initiated by the device, without a response for the previous one.
V2 -> V3: - Unused char in name string should be marked 0 now. - s/host/device/ and s/guest/driver/ - Added a new feature for IRQ mode, VIRTIO_GPIO_F_IRQ. - A new feature should be added for Version information if required later on.
V1 -> V2: - gpio_names_size is 32 bit. - gpio field is 16 bit. - padding added 16 bit. - Added packed attribute to few structures - Add the missing 'type' field to the request - Dropped to _nodata request/responses to simplify a bit, updated related text.
-- Viresh
[1] https://lists.oasis-open.org/archives/virtio-dev/202106/msg00030.html
Viresh Kumar (2): virtio-gpio: Add the device specification virtio-gpio: Add support for interrupts
conformance.tex | 30 ++- content.tex | 1 + virtio-gpio.tex | 608 ++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 635 insertions(+), 4 deletions(-) create mode 100644 virtio-gpio.tex