Re: [PATCH] usb: typec: ucsi: introduce read_explicit operation

[Heikki Krogerus <heikki.krogerus@linux.intel.com>]

Hi Sam,

On Thu, Mar 30, 2023 at 01:48:18AM +0200, Samuel Čavoj wrote:
> Hi Heikki,
> 
> On 2023-03-18 03:04, Samuel Čavoj wrote:
> ...
> > > Unfortunately nothing seems to work... I'm attaching the DMI quirk
> > > patch here. Can you test it?
> > 
> > I'll definitely try it out, I hope sometime next week!
> > 
> > > I'm not sure if the DMI_PRODUCT_NAME is just "ZenBook" so you may
> > > need to fix that in the patch!!
> > > 
> > > You can get the correct value by running dmidecode. This should work:
> > > 
> > >         % dmidecode -s system-product-name
> 
> This returns "ZenBook UX325UA_UM325UA", so the DMI_MATCH would work.
> However my DMI_SYS_VENDOR is "ASUSTeK COMPUTER INC.", uppercase.
> 
> All in all, the patch works after some modifications which I'm
> attaching below. In summary:
> 
>  - The DMI_MATCH SYS_VENDOR was changed to uppercase
>  - The DMI_MATCH PRODUCT_NAME was changed to be more specific, although
>    I'm not sure what the best value is here.
>  - The conditional in ucsi_zenbook_read was fixed.
>  - ua->cmd cannot be reset to 0 after read because the reset
>    procedure repeatedly calls read without performing
>    any other commands. I don't think this should cause any problems
>  - ucsi_acpi_notify needs to call the quirked variant
>    as well, so I put an indirect call there.
> 
> Otherwise maybe ucsi_acpi_async_write should only set cmd if the offset
> is UCSI_CONTROL.

Thanks!

> I'm occasionally getting some other errors later on, but I think
> those might be specific to a certain cheap USB-C hub I have. They
> don't occur with it unplugged.

Okay... Did you see those errors with your original patch?

Br,

-- 
heikki