On 5/5/25 07:26, Varadarajan Narayanan wrote:
On Fri, May 02, 2025 at 02:32:52PM +0200, Casey Connolly wrote:
Hi Varadarajan,
You have almost entirely ignored my feedback on the previous revision.
Sorry. Since you had said "For now I'm fine with this new
sysreset-qcom.c driver", in [1], I assumed you were ok with just
addressing the technical issues alone. Will add a new op to sysreset_ops
and post a new version.
Ah right, yeah my wording was not super clear there, my apologies. I'm
ok with the new driver but it should only handle the edl case, the
existing psci driver should be used for everything else (hence why it's
necessary to iterate over all sysreset devices first passing in the
arguments so the new sysreset-qcom driver would be able to do it's thing
before the generic psci driver.
Thanks and regards,>
Thanks
Varada
1 -
https://lore.kernel.org/u-boot/1e0e9a72-40a8-4ce3-a390-edef1d6a0...@linaro.org/
If there's some technical disagreement or you don't understand my feedback
then please let me know (in private if you prefer).
This approach is fundamentally wrong, full of layering violations and
totally unsustainable (imagine if every vendor started polluting core code
with their custom stuff).
These abstractions need to be improved to add this feature properly and
there's no getting around it.
Kind regards,
Casey (she/they)
[ . . . ]
--
Casey (she/they)
