forge/linux-pinenote
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

iommu/amd: Don't call the inv_ctx_cb when pasid is not set up

Browse source 
On the error path of amd_iommu_bind_pasid() we call
mmu_notifier_unregister() for cleanup. This calls
mn_release() which calls the users inv_ctx_cb function if
one is available. Since the pasid is not set up yet there is
nothing the user can to tear down in this call-back. So
don't call inv_ctx_cb on the error path of
amd_iommu_unbind_pasid() and make life of the users simpler.

Signed-off-by: Joerg Roedel <jroedel@suse.de>
Tested-by: Oded Gabbay <Oded.Gabbay@amd.com>
This commit is contained in:
Joerg Roedel 2014-07-09 15:43:11 +02:00
commit d9e1611e73
1 changed files with 12 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

13
drivers/iommu/amd_iommu_v2.c
View file

@ -52,7 +52,8 @@ struct pasid_state {
struct pri_queue pri[PRI_QUEUE_SIZE]; /* PRI tag states */ struct pri_queue pri[PRI_QUEUE_SIZE]; /* PRI tag states */
struct device_state *device_state; /* Link to our device_state */ struct device_state *device_state; /* Link to our device_state */
int pasid; /* PASID index */ int pasid; /* PASID index */
bool invalid; /* Used during teardown */ bool invalid; /* Used during setup and
teardown of the pasid */
spinlock_t lock; /* Protect pri_queues and spinlock_t lock; /* Protect pri_queues and
mmu_notifer_count */ mmu_notifer_count */
wait_queue_head_t wq; /* To wait for count == 0 */ wait_queue_head_t wq; /* To wait for count == 0 */
@ -473,13 +474,15 @@ static void mn_release(struct mmu_notifier *mn, struct mm_struct *mm)
{ {
struct pasid_state *pasid_state; struct pasid_state *pasid_state;
struct device_state *dev_state; struct device_state *dev_state;
bool run_inv_ctx_cb;
might_sleep(); might_sleep();
pasid_state = mn_to_state(mn); pasid_state = mn_to_state(mn);
dev_state = pasid_state->device_state; dev_state = pasid_state->device_state;
run_inv_ctx_cb = !pasid_state->invalid;
if (pasid_state->device_state->inv_ctx_cb) if (run_inv_ctx_cb && pasid_state->device_state->inv_ctx_cb)
dev_state->inv_ctx_cb(dev_state->pdev, pasid_state->pasid); dev_state->inv_ctx_cb(dev_state->pdev, pasid_state->pasid);
unbind_pasid(pasid_state); unbind_pasid(pasid_state);
@ -674,7 +677,8 @@ int amd_iommu_bind_pasid(struct pci_dev *pdev, int pasid,
pasid_state->mm = mm; pasid_state->mm = mm;
pasid_state->device_state = dev_state; pasid_state->device_state = dev_state;
pasid_state->pasid = pasid; pasid_state->pasid = pasid;
pasid_state->invalid = false; pasid_state->invalid = true; /* Mark as valid only if we are
done with setting up the pasid */
pasid_state->mn.ops = &iommu_mn; pasid_state->mn.ops = &iommu_mn;
if (pasid_state->mm == NULL) if (pasid_state->mm == NULL)
@ -691,6 +695,9 @@ int amd_iommu_bind_pasid(struct pci_dev *pdev, int pasid,
if (ret) if (ret)
goto out_clear_state; goto out_clear_state;
/* Now we are ready to handle faults */
pasid_state->invalid = false;
/* /*
* Drop the reference to the mm_struct here. We rely on the * Drop the reference to the mm_struct here. We rely on the
* mmu_notifier release call-back to inform us when the mm * mmu_notifier release call-back to inform us when the mm