Re: [PATCH 1/5] dt-bindings: power: bq27xxx: document bq27z746

Hi Krzysztof, 

Thanks for the comment!

On Mon, 2026-08-17 at 09:50 +0200, Krzysztof Kozlowski wrote:
> On Tue, Aug 11, 2026 at 01:47:43PM +0200, Henrik Grimler wrote:
> > Add compatible for Texas Instruments bq27z746 fuel gauge, which is
> > similar to its sibling bqz27561, but with more protection
> > functionality.
> 
> "More" means superset? Driver code says it is compatible, so this has to
> be explained 

I will update the commit message to explain the differences. Main
difference is that bq27z746 has integrated protection logic, and can
control external MOSFETs to stop charging/discharging if an issue is
detected (like overcurrent or overvoltage), or to keep the rest of the
system fully powered off in shipping mode. bq27z561 does not have such
protection circuitry, so best it can do if it detects issues is to send
an interrupt to some other component.

The standard register map is almost identical between the devices.
bq27z561 has 6 extra registers at the end that bq27z746 lacks though,
which I missed when comparing the TRMs the first time (so bq27z561 map
is a superset of bq27z746's). Neither of these extra registers are
supported/handled by the Linux driver through. I will update all the
commit messages for v2 to clarify.

There are also additional "ManufacturerAccess" registers where there
are some more differences between bq27z561 and bq27z746 mostly related
to the above mentioned MOSFET control, but the current Linux driver
does not read or handle any "ManufacturerAccess" register. One of my
colleagues has sent a patch to add support for reading PROP_MODEL_NAME
from these additional registers for bq27z561 though [1]. This model
name register is the same for bq27z561 and bq27z746.

In summary the chips have differences, but I expect it will be fine for
the driver to use same regs and props for bq27z746 and bq27z561.

> (explicitly requested in writing bindings btw).

I will read writing-bindings and writing-schema more carefully.

> Best regards,
> Krzysztof

[1]
https://lore.kernel.org/linux-pm/[email protected]/

Best regards,
Henrik Grimler