With commit
5b6272e0efd5 ("PCI: qcom: Add OPP support to scale
performance"), OPP was used to control the interconnect and power domains
if the platform supported OPP. Also to maintain the backward compatibility
with platforms not supporting OPP but just ICC, the above mentioned commit
assumed that if ICC was not available on the platform, it would resort to
OPP.
Unfortunately, some old platforms don't support either ICC or OPP. On those
platforms, resorting to OPP in the absence of ICC throws below errors from
OPP core during suspend and resume:
qcom-pcie
1c08000.pcie: dev_pm_opp_set_opp: device opp doesn't exist
qcom-pcie
1c08000.pcie: _find_key: OPP table not found (-19)
Also, it doesn't make sense to invoke the OPP APIs when OPP is not
supported by the platform at all.
Add a "use_pm_opp" flag to identify whether OPP is supported and use it to
control invoking the OPP APIs.
Fixes: 5b6272e0efd5 ("PCI: qcom: Add OPP support to scale performance")
Link: https://lore.kernel.org/linux-pci/20240722131128.32470-1-manivannan.sadhasivam@linaro.org
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Krzysztof WilczyĆski <kwilczynski@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mayank Rana <quic_mrana@quicinc.com>
const struct qcom_pcie_cfg *cfg;
struct dentry *debugfs;
bool suspended;
+ bool use_pm_opp;
};
#define to_qcom_pcie(x) dev_get_drvdata((x)->dev)
dev_err(pci->dev, "Failed to set bandwidth for PCIe-MEM interconnect path: %d\n",
ret);
}
- } else {
+ } else if (pcie->use_pm_opp) {
freq_mbps = pcie_dev_speed_mbps(pcie_link_speed[speed]);
if (freq_mbps < 0)
return;
max_freq);
goto err_pm_runtime_put;
}
+
+ pcie->use_pm_opp = true;
} else {
/* Skip ICC init if OPP is supported as it is handled by OPP */
ret = qcom_pcie_icc_init(pcie);
if (ret)
dev_err(dev, "Failed to disable CPU-PCIe interconnect path: %d\n", ret);
- if (!pcie->icc_mem)
+ if (pcie->use_pm_opp)
dev_pm_opp_set_opp(pcie->pci->dev, NULL);
}
return ret;