Re: [PATCH v8] PCI: exynos: Handle PCIe PHY initialization errors

From: Krzysztof Kozlowski

Date: Thu Nov 19 2026 - 07:10:31 EST


On 31/10/2026 08:35, Sreeraj S Kurup wrote:
> Check the return values of phy_init() and phy_power_on() to avoid
> continuing PCIe host initialization when PHY setup fails.
>
> Track PHY initialization, PHY power, and regulator enable state
> across probe, remove, suspend, and resume to ensure cleanup is
> performed only when appropriate and to prevent unbalanced
> reference counts.
>
> Assert core reset on probe failure after PHY initialization has
> been reached, and handle cleanup errors during teardown and
> suspend.
>
> Also check the return value of exynos_pcie_host_init() during
> resume and abort the resume sequence if PHY initialization fails.
>
> Signed-off-by: Sreeraj S Kurup <sreekuttan2156239@xxxxxxxxx>
> ---
> v8:
> - Fix state desynchronization in exynos_pcie_phy_exit() by returning
> early if phy_power_off() fails, preventing calls to phy_exit() on a
> still-powered PHY
> - Prevent PM core desynchronization and orphaned controller states
> during suspend-abort sequences in exynos_pcie_suspend_noirq() by
> logging teardown errors and returning 0 instead of aborting system
> suspend
> v7:

How many of these are you going to send (even the same day...) till you
get it right? The process is not random punching review system till you
get it past it. Sending random code till bot stops to complain only
proves that you do not understand what you sent.

I have zero confidence that you understand what you are doing and all
this looks like LLM slop sent to us. Why would we want LLM slop?

Best regards,
Krzysztof