Re: [PATCH v9 0/6] Enable cameras on Dell Latitude 5285 2-in-1

From: Matthias Eisenacher

Date: Sat Nov 14 2026 - 10:52:46 EST


Hi Thierry, Vitalii,

Thanks for the quick and helpful reply, and congratulations on getting
the 5285 series merged.

I should be upfront: I'm a user rather than a kernel developer, and my
5290 is a daily-use machine on a distribution kernel (7.0), so I can't
easily run and test a 7.4-rc mainline kernel on it at the moment. I
also only use the front camera.

Vitalii, since your tree already has 5290 board data covering both
cameras, you are probably in a much better position to send the two
follow-up patches Thierry suggested. My report above at least confirms
that the 5290 shows the same INT3446 resource conflict and needs board
data, and that the front camera works once both are addressed.

If DMI data, an acpidump or other information from my machine would
help, I'm happy to provide it.

Thanks again,
Matthias Eisenacher

Am 14.11.26 um 04:03 schrieb Thierry Chatard:
Hi Matthias,

Thanks a lot for testing on the 5290 and for the detailed report - great
that the front camera already comes up.

One thing worth flagging: the 5285 series has since been merged upstream
(mainline as of 7.4-rc1 - the mfd quirks via Lee's tree, the int3472 board
data and the media bits via the media tree). So there's no open series to
extend anymore; 5290 support would be a small follow-up on top of mainline
rather than an addition to this thread.

From your report the 5290 needs the same two pieces the 5285 did, so I
suggest two short patches:

1) mfd: intel-lpss: add Dell Latitude 5290 to the resource conflict quirk
- just another DMI entry in intel_lpss_quirk_dmi[]. This replaces the
acpi_enforce_resources=lax stand-in you used, so please test with the
real quirk entry rather than the boot parameter. Goes via Lee's tree.
Template: the 7275 commit that did exactly this,
495d5f41a6161dffafe61ac458459c62d440344a.

2) platform/x86: int3472: tps68470: add board data for Dell Latitude 5290
- a DMI entry in int3472_tps68470_board_data_table plus the board data.
Template: the 5285 board-data commit e605a094ff7a.

Two suggestions before sending:

- Please coordinate with Vitalii Tytskyi (CC'd) - his 5290 tree already
has 5290 board data and covers the rear camera (OV8858), so it is
better to reuse/validate that than to reuse the 7212 data. That also
gets the rear camera tested, which I think is still open on your side.

- Since you have the hardware, it is best that the patches come from you
(and/or Vitalii) with a Tested-by on the actual 5290 - board-data and
quirk patches really want to be tested on the machine, and I cannot do
that for the 5290.

I'm happy to review and give a Reviewed-by on the mfd quirk.

Thanks again,
Thierry

On Fri, Nov 13, 2026 04:04 PM, Matthias Eisenacher
<meisena2@xxxxxxxxxxxx> wrote:
Hi Thierry,

Thanks for this work. A short report from a Dell Latitude 5290 2-in-1
(same platform as the 5285), in case it helps to extend this series
to the 5290.

I did not apply this series. Instead I tested the minimal changes
needed for the front camera on the 5290:

- Linux Mint 22 (Ubuntu 24.04 base), kernel 7.0.0-43-generic (HWE)
- Added a DMI entry for "Latitude 5290 2-in-1" to
int3472_tps68470_board_data_table, reusing
dell_7212_tps68470_board_data
- Booted with acpi_enforce_resources=lax as a stand-in for your
intel-lpss quirk. Without it I get the same I2C4 conflict as on
the 5285:

ACPI Warning: SystemMemory range 0x00000000FE028000-0x00000000FE0281FF
conflicts with OpRegion 0x00000000FE028000-0x00000000FE028207
(\_SB.PCI0.GEXP.BAR0)
intel-lpss INT3446:00: probe with driver intel-lpss failed with error -16

With both changes the front camera (OV5670/INT3479) works: cam -l
lists "Internal front camera (\_SB_.PCI0.LNK0)", and I get video via
libcamerasrc/GStreamer and in Zoom (via libcamerify). The rear camera
was not tested.

DMI: sys_vendor "Dell Inc.", product_name "Latitude 5290 2-in-1",
board 09WJPW.

So the 5290 seems to need the same intel-lpss quirk plus board data.
Vitalii's tree has 5290 board data; would it make sense to add the
5290 DMI matches to this series or to a follow-up? I'm happy to test
patches on the 5290.

Thanks,
Matthias