提交 741669c7 编写于 作者: J Joerg Roedel

iommu/amd: Convert IOMMUv2 state_table into state_list

The state_table consumes 512kb of memory and is only sparsly
populated. Convert it into a list to save memory. There
should be no measurable performance impact.
Signed-off-by: NJoerg Roedel <jroedel@suse.de>
Tested-by: NJay Cornwall <Jay.Cornwall@amd.com>
上级 b87d2d7c
...@@ -56,6 +56,8 @@ struct pasid_state { ...@@ -56,6 +56,8 @@ struct pasid_state {
}; };
struct device_state { struct device_state {
struct list_head list;
u16 devid;
atomic_t count; atomic_t count;
struct pci_dev *pdev; struct pci_dev *pdev;
struct pasid_state **states; struct pasid_state **states;
...@@ -81,7 +83,7 @@ struct fault { ...@@ -81,7 +83,7 @@ struct fault {
u16 flags; u16 flags;
}; };
static struct device_state **state_table; static LIST_HEAD(state_list);
static spinlock_t state_lock; static spinlock_t state_lock;
/* List and lock for all pasid_states */ /* List and lock for all pasid_states */
...@@ -113,7 +115,14 @@ static u16 device_id(struct pci_dev *pdev) ...@@ -113,7 +115,14 @@ static u16 device_id(struct pci_dev *pdev)
static struct device_state *__get_device_state(u16 devid) static struct device_state *__get_device_state(u16 devid)
{ {
return state_table[devid]; struct device_state *dev_state;
list_for_each_entry(dev_state, &state_list, list) {
if (dev_state->devid == devid)
return dev_state;
}
return NULL;
} }
static struct device_state *get_device_state(u16 devid) static struct device_state *get_device_state(u16 devid)
...@@ -778,7 +787,8 @@ int amd_iommu_init_device(struct pci_dev *pdev, int pasids) ...@@ -778,7 +787,8 @@ int amd_iommu_init_device(struct pci_dev *pdev, int pasids)
spin_lock_init(&dev_state->lock); spin_lock_init(&dev_state->lock);
init_waitqueue_head(&dev_state->wq); init_waitqueue_head(&dev_state->wq);
dev_state->pdev = pdev; dev_state->pdev = pdev;
dev_state->devid = devid;
tmp = pasids; tmp = pasids;
for (dev_state->pasid_levels = 0; (tmp - 1) & ~0x1ff; tmp >>= 9) for (dev_state->pasid_levels = 0; (tmp - 1) & ~0x1ff; tmp >>= 9)
...@@ -808,13 +818,13 @@ int amd_iommu_init_device(struct pci_dev *pdev, int pasids) ...@@ -808,13 +818,13 @@ int amd_iommu_init_device(struct pci_dev *pdev, int pasids)
spin_lock_irqsave(&state_lock, flags); spin_lock_irqsave(&state_lock, flags);
if (state_table[devid] != NULL) { if (__get_device_state(devid) != NULL) {
spin_unlock_irqrestore(&state_lock, flags); spin_unlock_irqrestore(&state_lock, flags);
ret = -EBUSY; ret = -EBUSY;
goto out_free_domain; goto out_free_domain;
} }
state_table[devid] = dev_state; list_add_tail(&dev_state->list, &state_list);
spin_unlock_irqrestore(&state_lock, flags); spin_unlock_irqrestore(&state_lock, flags);
...@@ -852,7 +862,7 @@ void amd_iommu_free_device(struct pci_dev *pdev) ...@@ -852,7 +862,7 @@ void amd_iommu_free_device(struct pci_dev *pdev)
return; return;
} }
state_table[devid] = NULL; list_del(&dev_state->list);
spin_unlock_irqrestore(&state_lock, flags); spin_unlock_irqrestore(&state_lock, flags);
...@@ -927,7 +937,6 @@ EXPORT_SYMBOL(amd_iommu_set_invalidate_ctx_cb); ...@@ -927,7 +937,6 @@ EXPORT_SYMBOL(amd_iommu_set_invalidate_ctx_cb);
static int __init amd_iommu_v2_init(void) static int __init amd_iommu_v2_init(void)
{ {
size_t state_table_size;
int ret; int ret;
pr_info("AMD IOMMUv2 driver by Joerg Roedel <joerg.roedel@amd.com>\n"); pr_info("AMD IOMMUv2 driver by Joerg Roedel <joerg.roedel@amd.com>\n");
...@@ -943,16 +952,10 @@ static int __init amd_iommu_v2_init(void) ...@@ -943,16 +952,10 @@ static int __init amd_iommu_v2_init(void)
spin_lock_init(&state_lock); spin_lock_init(&state_lock);
state_table_size = MAX_DEVICES * sizeof(struct device_state *);
state_table = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO,
get_order(state_table_size));
if (state_table == NULL)
return -ENOMEM;
ret = -ENOMEM; ret = -ENOMEM;
iommu_wq = create_workqueue("amd_iommu_v2"); iommu_wq = create_workqueue("amd_iommu_v2");
if (iommu_wq == NULL) if (iommu_wq == NULL)
goto out_free; goto out;
ret = -ENOMEM; ret = -ENOMEM;
empty_page_table = (u64 *)get_zeroed_page(GFP_KERNEL); empty_page_table = (u64 *)get_zeroed_page(GFP_KERNEL);
...@@ -967,16 +970,13 @@ static int __init amd_iommu_v2_init(void) ...@@ -967,16 +970,13 @@ static int __init amd_iommu_v2_init(void)
out_destroy_wq: out_destroy_wq:
destroy_workqueue(iommu_wq); destroy_workqueue(iommu_wq);
out_free: out:
free_pages((unsigned long)state_table, get_order(state_table_size));
return ret; return ret;
} }
static void __exit amd_iommu_v2_exit(void) static void __exit amd_iommu_v2_exit(void)
{ {
struct device_state *dev_state; struct device_state *dev_state;
size_t state_table_size;
int i; int i;
if (!amd_iommu_v2_supported()) if (!amd_iommu_v2_supported())
...@@ -1005,9 +1005,6 @@ static void __exit amd_iommu_v2_exit(void) ...@@ -1005,9 +1005,6 @@ static void __exit amd_iommu_v2_exit(void)
destroy_workqueue(iommu_wq); destroy_workqueue(iommu_wq);
state_table_size = MAX_DEVICES * sizeof(struct device_state *);
free_pages((unsigned long)state_table, get_order(state_table_size));
free_page((unsigned long)empty_page_table); free_page((unsigned long)empty_page_table);
} }
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册