]> git.ipfire.org Git - thirdparty/kernel/linux.git/commitdiff
arm64: dts: qcom: msm8998: set qcom,no-msa-ready-indicator for wifi
authorMarc Gonzalez <mgonzalez@freebox.fr>
Mon, 29 Apr 2024 14:07:27 +0000 (16:07 +0200)
committerBjorn Andersson <andersson@kernel.org>
Tue, 28 May 2024 21:13:18 +0000 (16:13 -0500)
The ath10k driver waits for an "MSA_READY" indicator
to complete initialization. If the indicator is not
received, then the device remains unusable.

cf. ath10k_qmi_driver_event_work()

Several msm8998-based devices are affected by this issue.
Oddly, it seems safe to NOT wait for the indicator, and
proceed immediately when QMI_EVENT_SERVER_ARRIVE.

Jeff Johnson wrote:

  The feedback I received was "it might be ok to change all ath10k qmi
  to skip waiting for msa_ready", and it was pointed out that ath11k
  (and ath12k) do not wait for it.

  However with so many deployed devices, "might be ok" isn't a strong
  argument for changing the default behavior.

cf. also
https://wiki.postmarketos.org/wiki/Qualcomm_Snapdragon_835_(MSM8998)#WLAN

Signed-off-by: Marc Gonzalez <mgonzalez@freebox.fr>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Link: https://lore.kernel.org/r/0914f96e-fcfd-4088-924a-fc1991bce75f@freebox.fr
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
arch/arm64/boot/dts/qcom/msm8998.dtsi

index 254c12d7373c9a754528be870c4fcfda0ab8e4c4..580b7cabf757cd11a175194c34dc857d18d96a78 100644 (file)
                        iommus = <&anoc2_smmu 0x1900>,
                                 <&anoc2_smmu 0x1901>;
                        qcom,snoc-host-cap-8bit-quirk;
+                       qcom,no-msa-ready-indicator;
                };
        };
 };