QEMU question: upstreaming I2C device with unpublished datasheet

QEMU question: upstreaming I2C device with unpublished datasheet

[Shengtan Mao]

[-- Attachment #1: Type: text/plain, Size: 299 bytes --]

Hi everyone,
we are hoping to upstream a MAX I2C device to QEMU. The device's datasheet
is not public, and we are contacting the Maxim company to get their
permission. If Maxim is okay with upstreaming the device with an
unpublished datasheet, will this still be an issue with QEMU?

best,
Shengtan

[-- Attachment #2: Type: text/html, Size: 381 bytes --]

Re: QEMU question: upstreaming I2C device with unpublished datasheet

[Peter Maydell]

On Wed, 21 Jul 2021 at 22:06, Shengtan Mao <stmao@google.com> wrote:
>
> Hi everyone,
> we are hoping to upstream a MAX I2C device to QEMU. The device's datasheet is not public, and we are contacting the Maxim company to get their permission. If Maxim is okay with upstreaming the device with an unpublished datasheet, will this still be an issue with QEMU?

It's not inherently a blocker, but the difficulty is code review.
Without knowing how the device should behave, it's hard to tell
whether the model is doing the right thing. Ideally Maxim should
just publish their datasheet, so if you/your company has any
leverage with them you should push them to do that :-)

That said, we do have other device models in the tree which
don't have public datasheets, and as long as the device is not
too complicated we should be OK as long as you can also submit
enough test cases that we could refactor the device in future
without worrying too much that we're breaking it because we don't
know how it's supposed to behave.

thanks
-- PMM

Re: QEMU question: upstreaming I2C device with unpublished datasheet

[Corey Minyard]

On Tue, Jul 27, 2021 at 10:23:26AM +0100, Peter Maydell wrote:
> On Wed, 21 Jul 2021 at 22:06, Shengtan Mao <stmao@google.com> wrote:
> >
> > Hi everyone,
> > we are hoping to upstream a MAX I2C device to QEMU. The device's datasheet is not public, and we are contacting the Maxim company to get their permission. If Maxim is okay with upstreaming the device with an unpublished datasheet, will this still be an issue with QEMU?
> 
> It's not inherently a blocker, but the difficulty is code review.
> Without knowing how the device should behave, it's hard to tell
> whether the model is doing the right thing. Ideally Maxim should
> just publish their datasheet, so if you/your company has any
> leverage with them you should push them to do that :-)
> 
> That said, we do have other device models in the tree which
> don't have public datasheets, and as long as the device is not
> too complicated we should be OK as long as you can also submit
> enough test cases that we could refactor the device in future
> without worrying too much that we're breaking it because we don't
> know how it's supposed to behave.

Thanks Peter.

I will add that you need to be sure you have the legal right to publish
the software.  If you have a non-public datasheet, that probably means
you had to sign NDA or something like that to see it.  That may limit
what you can do.

-corey