Re: [Letux-kernel] [PATCH] ARM: omap2plus_defconfig: Fix selected panels after generic panel changes

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Oct 07, 2019 at 06:09:57PM +0200, Andreas Kemnade wrote:
> On Mon, 7 Oct 2019 17:29:20 +0300 Tomi Valkeinen wrote:
> > On 07/10/2019 07:40, Andreas Kemnade wrote:
> > 
> >>>> That's really weird, the driver name, the SPI device table and the OF
> >>>> device table are the same (except for the "omapdss," prefix that we
> >>>> don't add anymore in omapdss-boot-init.c). Would you be able to
> >>>> investigate what broke ?
> >>>>      
> >>> in earlier times:
> >>>
> >>> root@gta04:/sys/bus/spi/devices/spi4.0# cat modalias
> >>> spi:tpo,td028ttec1
> >>>
> >>> now in 5.4-rc1:
> >>> root@gta04:/sys/bus/spi/devices/spi4.0# cat modalias
> >>> spi:td028ttec1
> >>>
> >>> root@gta04:~# modinfo /lib/modules/5.4.0-rc1-letux+/kernel/drivers/gpu/drm/panel/panel-tpo-td028ttec1.ko
> >>> filename:       /lib/modules/5.4.0-rc1-letux+/kernel/drivers/gpu/drm/panel/panel-tpo-td028ttec1.ko
> >>> license:        GPL
> >>> description:    Toppoly TD028TTEC1 panel driver
> >>> author:         H. Nikolaus Schaller <hns@xxxxxxxxxxxxx>
> >>> srcversion:     6B3E224BCD3D76253CF361C
> >>> alias:          of:N*T*Ctoppoly,td028ttec1C*
> >>> alias:          of:N*T*Ctoppoly,td028ttec1
> >>> alias:          of:N*T*Ctpo,td028ttec1C*
> >>> alias:          of:N*T*Ctpo,td028ttec1
> >>> alias:          spi:toppoly,td028ttec1
> >>> alias:          spi:tpo,td028ttec1
> >>> depends:        drm
> >>> intree:         Y
> >>> name:           panel_tpo_td028ttec1
> >>> vermagic:       5.4.0-rc1-letux+ SMP preempt mod_unload ARMv7 p2v8
> >>>
> >>> That alias is not in the list.
> >>>  
> >> some more research:
> >> in former times
> >> of_modalias_node() has stripped the omapdss prefix, result: modalias=spi:tpo,td028ttec1
> >> now it strips the tpo prefix because there is no omapdss prefix anymore.  
> > 
> > I haven't studied this more yet, but yes, I can see of_modalias_node 
> > stripping the vendor prefix on purpose. I do wonder how this is supposed 
> > to work, it would make more sense to me to keep the vendor prefix.
> > 
> > Is the spi_device_id supposed to be without vendor prefix? With a quick 
> > grep, this seems to be the case.
>
> as it is done in drivers/spi/spi.c we have no influence from drm side.
> So we must live with it. Also checked several drivers, no vendor-prefix
> there.

I *think* SPI drivers are not supposed to include any prefix. The fix
would then simply be to strip the prefix from the SPI table (leaving it
there for the OF table of course). I've submitted a patch series to do
so.

-- 
Regards,

Laurent Pinchart



[Index of Archives]     [Linux Arm (vger)]     [ARM Kernel]     [ARM MSM]     [Linux Tegra]     [Linux WPAN Networking]     [Linux Wireless Networking]     [Maemo Users]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux