RE: [PATCH 4/4] dmaengine/idxd: Re-enable kernel workqueue under DMA API

[Tian, Kevin]

> From: Jacob Pan <jacob.jun.pan@xxxxxxxxxxxxxxx>
> Sent: Thursday, March 2, 2023 9:00 AM
> 
>  static int idxd_enable_system_pasid(struct idxd_device *idxd)
>  {
> -	return -EOPNOTSUPP;
> +	struct pci_dev *pdev = idxd->pdev;
> +	struct device *dev = &pdev->dev;
> +	struct iommu_domain *domain;
> +	union gencfg_reg gencfg;
> +	ioasid_t pasid;
> +	int ret;
> +
> +	domain = iommu_get_domain_for_dev(dev);
> +	if (!domain || domain->type == IOMMU_DOMAIN_BLOCKED)
> +		return -EPERM;

what about UNMANAGED?

> +
> +	pasid = iommu_sva_reserve_pasid(1, dev->iommu->max_pasids);
> +	if (pasid == IOMMU_PASID_INVALID)
> +		return -ENOSPC;

as commented in last patch we can just pass a device pointer to a
general allocation interface.

> +
> +	ret = iommu_attach_device_pasid(domain, dev, pasid);
> +	if (ret) {
> +		dev_err(dev, "failed to attach device pasid %d, domain
> type %d",
> +			pasid, domain->type);
> +		iommu_sva_unreserve_pasid(pasid);
> +		return ret;
> +	}
> +
> +	/* Since we set user privilege for kernel DMA, enable completion IRQ
> */
> +	gencfg.bits = ioread32(idxd->reg_base + IDXD_GENCFG_OFFSET);
> +	gencfg.user_int_en = 1;
> +	iowrite32(gencfg.bits, idxd->reg_base + IDXD_GENCFG_OFFSET);
> +	idxd->pasid = pasid;

Why does user privilege requires a completion interrupt?

Or instead it's more due to doing kernel DMA itself then we certainly
don't want to poll in the kernel?