linux-kernel - Re: [GIT PULL] Ressurect the atomisp staging driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200503120714.28f62360@coco.lan>
Date:   Sun, 3 May 2020 12:07:14 +0200
From:   Mauro Carvalho Chehab <mchehab@...nel.org>
To:     Francescodario Cuzzocrea <francescodario.cuzzocrea@...l.polimi.it>
Cc:     "mchehab+huawei@...nel.org" <mchehab+huawei@...nel.org>,
        "linux-media@...r.kernel.org" <linux-media@...r.kernel.org>,
        "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "gnomes@...rguk.ukuu.org.uk" <gnomes@...rguk.ukuu.org.uk>,
        "alan@...ux.intel.com" <alan@...ux.intel.com>
Subject: Re: [GIT PULL] Ressurect the atomisp staging driver

Hi Francescodario,

Em Sun, 3 May 2020 09:19:55 +0000
Francescodario Cuzzocrea <francescodario.cuzzocrea@...l.polimi.it> escreveu:

> Hi ! 
> 
> Thanks for this, it is very appreciated !! 
> 
> I have a bay trail 2in1 (namely, the acer aspire swtich 10 sw5-012,
> with an OV2722 module, so supported by the driver) on top of which I
> run android-x86-q so I would be able to help a bit with testing the
> driver.
> 
> I tried to pull in your changes on top of my 5.7-rc2 here : 
> https://gitlab.com/lineageos-x86/android_kernel_common/-/commits/5.7-media
> 
> but I did not understood how to correctly enable the driver.
> According
> to Kconfig help, there is still the hardcoded switch between ISP2401
> and ISP2400, so for my Bay Trail tablet I should set :
> 
> #
> CONFIG_VIDEO_ATOMISP_ISP2401 is not set
> 
> but settings this breaks compilation with : 
> 
> ERROR: modpost:
> "ia_css_configure_sc"[drivers/staging/media/atomisp/atomisp.ko]
> undefined!

I'll double check and address this issue. 

> 
> Looking at commit history, I noticed that the driver should be able to
> recognize at runtime the ISP version,

Not yet. There are simply too many places there with checks for ISP2400
or ISP2401. I got rid of most of them, but there are 4 header files
that contain ISP-version specific information, whose are used by several
parts of the driver. Getting rid of it would require more work.

> so what I tried next was to set
> CONFIG_VIDEO_ATOMISP_ISP2401=y, and the compilation went fine.
> I was able to boot the kernel (and I also added both
> shisp_2400b0_v21.bin and shisp_2401a0_v21.bin from an android tablet
> dump).
> 
> The kernel booted fine, the ov2722 module is loaded, but according to
> dmesg it is not able to initialize the hardware : 
> https://pastebin.com/VBMD1bau

(transcribing the relevant messages from pastebin below)

> ov2722 i2c-INT33FB: 00: gmin: initializing atomisp module subdev data.PMIC ID 1
> ov2722 i2c-INT33FB: 00: supply V1P2A not found, using dummy regulator
> ov2722 i2c-INT33FB: 00: supply VPROG4B not found, using dummy regulator
> ov2722 i2c-INT33FB: 00: supply Regulator1p8v not found, using dummy regulator
> ov2722 i2c-INT33FB: 00: supply Regulator2p8v not found, using dummy regulator
> ov2722 i2c-INT33FB: 00: unable to set PMC rate 1
> ov2722 i2c-INT33FB: 00: camera pdata: port: 0 lanes: 1 order: 00000000
> ov2722 i2c-INT33FB: 00: read from offset 0x300a error -121
> ov2722 i2c-INT33FB: 00: sensor_id_high = 0xffff
> ov2722 i2c-INT33FB: 00: ov2722_detect err s_config.
> ov2722 i2c-INT33FB: 00: sensor power-gating failed

That's the same that happened with the tests we've done so far:
it is not finding the regulators. I'm currently seeking for a solution
to that.

Usually, regulator data comes via OpenFirmware/Device Tree.

However, ACPI BIOS store this on a different way. Some (missing) code
would need to get the relevant parameters from the BIOS and setup
the regulators.

Btw, could you please send me a tarball with the output of your
BIOS ACPI tables?

You can get it with:

	$ sudo acpidump |bzip2 > acpi.log.bz2

Thanks,
Mauro