Re: atomisp kernel driver(s)

[Andy Shevchenko <andy.shevchenko@gmail.com>]

On Tue, May 12, 2020 at 5:56 PM Mauro Carvalho Chehab
<mchehab@kernel.org> wrote:
> Em Tue, 12 May 2020 14:21:49 +0300
> Andy Shevchenko <andy.shevchenko@gmail.com> escreveu:
> > On Tue, May 12, 2020 at 2:12 PM Andy Shevchenko
> > <andy.shevchenko@gmail.com> wrote:
> > > On Tue, May 12, 2020 at 1:21 PM Mauro Carvalho Chehab
> > > <mchehab@kernel.org> wrote:
> > > > Em Sat, 2 May 2020 20:33:14 +0300
> > > > Andy Shevchenko <andy.shevchenko@gmail.com> escreveu:
> > > > > On Sat, May 2, 2020 at 8:04 PM Mauro Carvalho Chehab <mchehab@kernel.org> wrote:
> > > > > > Em Sat, 2 May 2020 19:08:36 +0300
> > > > > > Andy Shevchenko <andy.shevchenko@gmail.com> escreveu:
> >
> > ...
> >
> > > Btw, https://git.linuxtv.org/mchehab/experimental.git/commit/?h=atomisp_v2.1&id=65608aa8d34ea274600ab2cf6f0cf54ee86d8fd1
> > > is incorrect approach. Look closer what PCI does in case when
> > > pcim_enable_device() has been called.
> >
> > This has resource leaks
> > https://git.linuxtv.org/mchehab/experimental.git/commit/?h=atomisp_v2.1&id=88b03de5d705f5f46a896dbd21ef9472030bb8d3
> >
> > Easier just to acpi_handle_info(ACPI_HANDLE(...), ...);
> >
> > We don't enumerate them w/o ACPI IIRC.
>
> Well, at least while in staging, it would be good to know more details
> about the probing issues on different hardware that people might be
> testing it.
>
> avoiding to use a get_device() here is not possible, because the
> prints will use the adev to return some info we may need:
>
>         ov2680 i2c-OVTI2680:00: ov2680_probe: ACPI detected it on bus ID=CAM1, HID=OVTI2680

Why do you need get_device()? Are you expecting device may have gone?
acpi_handle_*() / dev_*() are not enough?

> This, together with the DMI product ID on this message:
>
>         atomisp-isp2 0000:00:03.0: Detected Cherrytrail version 54 (ISP2401) on T101HA
>
> may help to avoid users to run acpixtract/iasl/dmidecode when adding
> new entries to the dmi match tables. Not 100%, since BIOSes may
> still use different names for the sensor data.

> Trying to open a video devnode returns error and produce those logs:
>
>         [ 3262.961623] atomisp-isp2 0000:00:03.0: open device ATOMISP ISP CAPTURE output
>         [ 3262.971350] sh_css_hrt_system_is_idle() 44: warning: SP not idle
>         [ 3262.971364] sh_css_hrt_system_is_idle() 49: warning: ISP not idle
>         [ 3262.971712] sh_css_hrt_system_is_idle() 56: warning: FIFO channel 12 is not empty
>         [ 3262.972079] sh_css_hrt_system_is_idle() 56: warning: FIFO channel 26 is not empty
>         [ 3262.972188] atomisp-isp2 0000:00:03.0: css init failed --- bad firmware?
>         [ 3262.974982] atomisp-isp2 0000:00:03.0: atomisp_mrfld_pre_power_down: error in iunit interrupt. status reg=0xffffffff

In my case when I tried with 2015-04 / 2015-05 firmware for ISP2400 I
got to no crashes, but no IRQs, presumably due to absence of PMIC
integration.

-- 
With Best Regards,
Andy Shevchenko