amd kdump failure with iommu=nopt

Joerg Roedel joro at 8bytes.org
Thu May 14 15:36:24 UTC 2020


Hi Jerry,

On Wed, May 13, 2020 at 08:18:38PM -0700, Jerry Snitselaar wrote:
> We've seen kdump failures with recent kernels (5.5, 5.6, 5.7-rc1) on
> amd systems when iommu is enabled in translation mode. In the cases so
> far there has been mpt3sas involved, but I'm also seeing io page
> faults for ahci right before mpt3sas has an io page fault:

Thanks for the report!

>> It was bisected to:
> 
> commit be62dbf554c5b50718a54a359372c148cd9975c7
> Author: Tom Murphy <murphyt7 at tcd.ie>
> Date:   Sun Sep 8 09:56:41 2019 -0700
> 
>     iommu/amd: Convert AMD iommu driver to the dma-iommu api

This commit also removes the deferred attach of the device to its new
domain. Does the attached diff fix the problem for you?


	Joerg

diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
index 4050569188be..40df255b6c7a 100644
--- a/drivers/iommu/iommu.c
+++ b/drivers/iommu/iommu.c
@@ -1889,23 +1889,37 @@ void iommu_domain_free(struct iommu_domain *domain)
 }
 EXPORT_SYMBOL_GPL(iommu_domain_free);
 
-static int __iommu_attach_device(struct iommu_domain *domain,
-				 struct device *dev)
+static bool __iommu_is_attach_deferred(struct iommu_domain *domain,
+				       struct device *dev)
 {
-	int ret;
-	if ((domain->ops->is_attach_deferred != NULL) &&
-	    domain->ops->is_attach_deferred(domain, dev))
-		return 0;
+	if (!domain->ops->is_attach_deferred)
+		return false;
+
+	return domain->ops->is_attach_deferred(domain, dev);
+}
 
+static int __iommu_attach_device_no_defer(struct iommu_domain *domain,
+					  struct device *dev)
+{
 	if (unlikely(domain->ops->attach_dev == NULL))
 		return -ENODEV;
 
 	ret = domain->ops->attach_dev(domain, dev);
 	if (!ret)
 		trace_attach_device_to_domain(dev);
+
 	return ret;
 }
 
+static int __iommu_attach_device(struct iommu_domain *domain,
+				 struct device *dev)
+{
+	if (__iommu_is_attach_deferred(domain, dev))
+		return 0;
+
+	return __iommu_attach_device_no_defer(domain, dev);
+}
+
 int iommu_attach_device(struct iommu_domain *domain, struct device *dev)
 {
 	struct iommu_group *group;
@@ -2023,7 +2037,12 @@ EXPORT_SYMBOL_GPL(iommu_get_domain_for_dev);
  */
 struct iommu_domain *iommu_get_dma_domain(struct device *dev)
 {
-	return dev->iommu_group->default_domain;
+	struct iommu_domain *domain = dev->iommu_group->default_domain;
+
+	if (__iommu_is_attach_deferred(domain, dev))
+		__iommu_attach_device_no_defer(domain, dev);
+
+	return domain;
 }
 
 /*


More information about the iommu mailing list