[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20240625092332.GE2642@thinkpad>
Date: Tue, 25 Jun 2024 14:53:32 +0530
From: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
To: Slark Xiao <slark_xiao@....com>
Cc: Jeffrey Hugo <quic_jhugo@...cinc.com>,
Loic Poulain <loic.poulain@...aro.org>, ryazanov.s.a@...il.com,
johannes@...solutions.net, netdev@...r.kernel.org,
mhi@...ts.linux.dev, linux-arm-msm@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: Re: Re: Re: [PATCH v2 1/2] bus: mhi: host: Import mux_id item
On Tue, Jun 25, 2024 at 04:28:25PM +0800, Slark Xiao wrote:
[...]
> >Hmm, sadly we shouldn't have used the same controller config for all these
> >devices across different product families. I didn't really paid attention to the
> >device name which is supposed to be unique (that's my bad).
> >
> >For instance, because of the controller config reuse, your SDX62 modem would
> >print:
> >
> >"MHI PCI device found: foxconn-sdx65"
> >
> >which clearly is misleading the users...
> >
> >I've submitted a patch that uses unique product name across the product families
> >[1]. Please take a look. After this patch, you can use the modem name to
> >differentiate in client drivers.
> >
> >- Mani
> >
> >[1] https://lore.kernel.org/mhi/20240625074148.7412-1-manivannan.sadhasivam@linaro.org/
> >
> >--
> >மணிவண்ணன் சதாசிவம்
> For same chip platform, I don't think it's necessary to separate into different parts.
> Like t99w368 and DW5932e, all things are same except the 'name'. For previous
> mux_id settings, we would like to add it for sdx72/sdx75 platforms, but shall
> no difference on T99W515 and DW5934e.
> Otherwise, we must to update both mhi and wwan side if we have a new foxconn
> SDX72 device support since the name is different with foxconn-t99w515 or
> foxconn-dw5934e.
>
Name is an important factor for an end user. Because, even though both products
are same in functionality, they are marketed as different products. So the users
should be provided with the actual product name, not baseline.
Even though it requires an update to the pci_generic driver, it ought to happen
for correctness.
- Mani
--
மணிவண்ணன் சதாசிவம்
Powered by blists - more mailing lists