RE: [PATCH v3 7/8] staging: fsl-mc: update TODO and README for restool driver

From: Stuart Yoder
Date: Tue Feb 09 2016 - 13:29:22 EST




> -----Original Message-----
> From: Lijun Pan [mailto:Lijun.Pan@xxxxxxxxxxxxx]
> Sent: Monday, February 08, 2016 5:40 PM
> To: gregkh@xxxxxxxxxxxxxxxxxxx; arnd@xxxxxxxx; devel@xxxxxxxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx
> Cc: stuart.yoder@xxxxxxxxxxxxx; itai.katz@xxxxxxxxxxxxx; german.rivera@xxxxxxxxxxxxx;
> leoli@xxxxxxxxxxxxx; scottwood@xxxxxxxxxxxxx; agraf@xxxxxxx; bhamciu1@xxxxxxxxxxxxx;
> R89243@xxxxxxxxxxxxx; bhupesh.sharma@xxxxxxxxxxxxx; nir.erez@xxxxxxxxxxxxx;
> richard.schmitt@xxxxxxxxxxxxx; dan.carpenter@xxxxxxxxxx; Lijun.Pan2000@xxxxxxxxx; Lijun
> Pan <Lijun.Pan@xxxxxxxxxxxxx>
> Subject: [PATCH v3 7/8] staging: fsl-mc: update TODO and README for restool driver
>
> Add more introduction of restool driver and state why
> restool driver is needed in helping moving fsl-mc bus
> out of staging tree.
>
> Signed-off-by: Lijun Pan <Lijun.Pan@xxxxxxxxxxxxx>
> ---
> drivers/staging/fsl-mc/README.txt | 11 ++++++++++-
> drivers/staging/fsl-mc/TODO | 18 ++++++++++++++++--
> 2 files changed, 26 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/staging/fsl-mc/README.txt b/drivers/staging/fsl-mc/README.txt
> index 8214102..e9ec507 100644
> --- a/drivers/staging/fsl-mc/README.txt
> +++ b/drivers/staging/fsl-mc/README.txt
> @@ -130,7 +130,16 @@ the objects involved in creating a network interfaces.
> via a config file passed to the MC when firmware starts
> it. There is also a Linux user space tool called "restool"
> that can be used to create/destroy containers and objects
> - dynamically.
> + dynamically. The kernel side restool driver communicates with
> + user space restool via ioctl. Restool relies on allocator driver
> + to allocate dpmcp resources, enumerates fsl-mc bus to find root dprc
> + objects of interest. When the user space restool program sends a request
> + to restool driver to create a dp* objects in MC firmware, an interrupt
> + will be triggered by MC firmware and the dprc driver's interrupt handler
> + shall process the interrupt (synchronizing the objects in MC firmware and
> + objects in Linux kernel). Though small, restool driver helps verify all
> + the functionality of fsl-mc bus, dprc driver, allocator driver,
> + and MC flib API.
>
> -DPAA2 Objects for an Ethernet Network Interface
>
> diff --git a/drivers/staging/fsl-mc/TODO b/drivers/staging/fsl-mc/TODO
> index 5065821..4892eb6 100644
> --- a/drivers/staging/fsl-mc/TODO
> +++ b/drivers/staging/fsl-mc/TODO
> @@ -5,10 +5,24 @@
> fsl-mc bus out of staging.
>
> * Decide if multiple root fsl-mc buses will be supported per Linux instance,
> - and if so add support for this.
> + and if so add support for this. No matter fsl-mc bus support multiple root
> + dprc or not, restool driver is designed to support multiple root if fsl-mc
> + bus is ready some day later. If there is only one root dprc, restool driver
> + works fine.
> +
> +* Add at least one driver utilizing fsl-mc bus. Restool driver is a very
> + small and simple driver, which interacts with fsl-mc bus, dprc driver,
> + allocator driver. Restool relies on allocator driver to allocate
> + dpmcp resources, enumerates fsl-mc bus to find root dprc objects of interest.
> + When the user space restool program sends a request to restool driver to
> + create a dp* objects in MC firmware, an interrupt will be triggered by
> + MC firmware and the dprc driver's interrupt handler shall process the
> + interrupt. Though small, restool driver helps verify all the functionality
> + of fsl-mc bus, dprc driver, allocator driver, and MC flib API. Restool
> + driver helps in moving fsl-mc bus out of staging branch.

I don't agree with expanding the TODO list. The discussion with Alex Graf
and Greg last year left us with the target of getting a object driver into
staging that uses the fsl-mc bus in a normal way, which is what the current
TODO list reflects. Now that MSI support is in we are closer to being able
to do that.

The restool driver is a special case that I think can wait.

Stuart