Re: [PATCH v3 2/2] mmc: sdhci-cadence: add CQE support

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] In `sdhci_cdns_resume()`, if `cqhci_resume()` were to fail, the error path fails to roll back the already-resumed SDHCI host.
--

--- Patch [2]: [PATCH v3 2/2] mmc: sdhci-cadence: add CQE support ---
commit 3bdb7a0ff9ae82cc040dd069fa74bd0366328e92
Author: Rohan Joshi <rohan1sj@cadence.com>

mmc: sdhci-cadence: add CQE support

This commit adds optional Command Queue Engine (CQE) support to the Cadence
SDHCI host controller driver. The presence of CQE is determined by an
additional host capability register and integrated into the host setup.

> diff --git a/drivers/mmc/host/sdhci-cadence.c b/drivers/mmc/host/sdhci-cadence.c
> --- a/drivers/mmc/host/sdhci-cadence.c
> +++ b/drivers/mmc/host/sdhci-cadence.c

[ ... ]

> @@ -630,6 +739,13 @@ static int sdhci_cdns_resume(struct device *dev)
>  	ret = sdhci_resume_host(host);
>  	if (ret)
>  		goto disable_clk;
>  
> +	/* Resume CQE if enabled */
> +	if (priv->cqe_support) {
> +		ret = cqhci_resume(host->mmc);
> +		if (ret)
> +			goto disable_clk;
> +	}
> +
>  	return 0;
>  
>  disable_clk:
>  	clk_disable_unprepare(pltfm_host->clk);
>  
>  	return ret;
>  }

[Severity: Low]
If cqhci_resume() fails, the code jumps to disable_clk, which disables the
clock but leaves the SDHCI host in a resumed state since sdhci_resume_host()
was already called successfully. Is it intentional to skip rolling back the
host resume here?

While cqhci_resume() unconditionally returns 0 today, making this error
path structurally impossible to execute at the moment, should this perhaps
call sdhci_suspend_host() on error to keep the power management state
balanced in case cqhci_resume() behavior changes in the future?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260722-cdns_sdhci_cqe-support-v3-0-4412856a4488@cadence.com?part=2