Re: [PATCH] iommu/qcom: Fix bogus detach logic

[Stephan Gerhold <stephan@gerhold.net>]

On Tue, Feb 18, 2020 at 06:12:41PM +0000, Robin Murphy wrote:
> Currently, the implementation of qcom_iommu_domain_free() is guaranteed
> to do one of two things: WARN() and leak everything, or dereference NULL
> and crash. That alone is terrible, but in fact the whole idea of trying
> to track the liveness of a domain via the qcom_domain->iommu pointer as
> a sanity check is full of fundamentally flawed assumptions. Make things
> robust and actually functional by not trying to be quite so clever.
> 
> Reported-by: Brian Masney <masneyb@onstation.org>
> Tested-by: Brian Masney <masneyb@onstation.org>
> Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
> Fixes: 0ae349a0f33f ("iommu/qcom: Add qcom_iommu")
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>

This fixes the warning reported by Naresh Kamboju [1] for me. Thank you!

Tested-by: Stephan Gerhold <stephan@gerhold.net>

[1]: https://lore.kernel.org/linux-arm-msm/CA+G9fYtScOpkLvx=__gP903uJ2v87RwZgkAuL6RpF9_DTDs9Zw@mail.gmail.com/

> ---
>  drivers/iommu/qcom_iommu.c | 28 ++++++++++++----------------
>  1 file changed, 12 insertions(+), 16 deletions(-)
> 
> diff --git a/drivers/iommu/qcom_iommu.c b/drivers/iommu/qcom_iommu.c
> index 39759db4f003..4328da0b0a9f 100644
> --- a/drivers/iommu/qcom_iommu.c
> +++ b/drivers/iommu/qcom_iommu.c
> @@ -344,21 +344,19 @@ static void qcom_iommu_domain_free(struct iommu_domain *domain)
>  {
>  	struct qcom_iommu_domain *qcom_domain = to_qcom_iommu_domain(domain);
>  
> -	if (WARN_ON(qcom_domain->iommu))    /* forgot to detach? */
> -		return;
> -
>  	iommu_put_dma_cookie(domain);
>  
> -	/* NOTE: unmap can be called after client device is powered off,
> -	 * for example, with GPUs or anything involving dma-buf.  So we
> -	 * cannot rely on the device_link.  Make sure the IOMMU is on to
> -	 * avoid unclocked accesses in the TLB inv path:
> -	 */
> -	pm_runtime_get_sync(qcom_domain->iommu->dev);
> -
> -	free_io_pgtable_ops(qcom_domain->pgtbl_ops);
> -
> -	pm_runtime_put_sync(qcom_domain->iommu->dev);
> +	if (qcom_domain->iommu) {
> +		/*
> +		 * NOTE: unmap can be called after client device is powered
> +		 * off, for example, with GPUs or anything involving dma-buf.
> +		 * So we cannot rely on the device_link.  Make sure the IOMMU
> +		 * is on to avoid unclocked accesses in the TLB inv path:
> +		 */
> +		pm_runtime_get_sync(qcom_domain->iommu->dev);
> +		free_io_pgtable_ops(qcom_domain->pgtbl_ops);
> +		pm_runtime_put_sync(qcom_domain->iommu->dev);
> +	}
>  
>  	kfree(qcom_domain);
>  }
> @@ -404,7 +402,7 @@ static void qcom_iommu_detach_dev(struct iommu_domain *domain, struct device *de
>  	struct qcom_iommu_domain *qcom_domain = to_qcom_iommu_domain(domain);
>  	unsigned i;
>  
> -	if (!qcom_domain->iommu)
> +	if (WARN_ON(!qcom_domain->iommu))
>  		return;
>  
>  	pm_runtime_get_sync(qcom_iommu->dev);
> @@ -417,8 +415,6 @@ static void qcom_iommu_detach_dev(struct iommu_domain *domain, struct device *de
>  		ctx->domain = NULL;
>  	}
>  	pm_runtime_put_sync(qcom_iommu->dev);
> -
> -	qcom_domain->iommu = NULL;
>  }
>  
>  static int qcom_iommu_map(struct iommu_domain *domain, unsigned long iova,
> -- 
> 2.23.0.dirty
>