Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1624892 > unrolled thread

[PATCH v5 0/3] vfio/type1: Synchronous locked page accounting

Back to article view | Back to linux.kernel

Contents

  [PATCH v5 0/3] vfio/type1: Synchronous locked page accounting Alex Williamson <alex.williamson@redhat.com> - 2017-04-18 00:40 +0200
    [PATCH v5 2/3] vfio/type1: Prune vfio_pin_page_external() Alex Williamson <alex.williamson@redhat.com> - 2017-04-18 00:40 +0200
    Re: [PATCH v5 0/3] vfio/type1: Synchronous locked page accounting Peter Xu <peterx@redhat.com> - 2017-04-18 06:30 +0200
      Re: [PATCH v5 0/3] vfio/type1: Synchronous locked page accounting Kirti Wankhede <kwankhede@nvidia.com> - 2017-04-18 21:00 +0200

#1624892 — [PATCH v5 0/3] vfio/type1: Synchronous locked page accounting

v5:
  - patch 1/ Use bool* to cleanup vfio_lock_acct() callers; sorry
             we cannot re-test CAP_IPC_LOCK for all callers
  - patch 2/ Re-add pr_warn, add Kirti's R-b
  - patch 3/ NEW, analyzing impact of vfio_lock_acct() testing
             CAP_IPC_LOCK for all callers revealed a long hanging
             optimization

Thanks for the reviews, keep 'em coming,
Alex

---

Alex Williamson (3):
      vfio/type1: Remove locked page accounting workqueue
      vfio/type1: Prune vfio_pin_page_external()
      vfio/type1: Reduce repetitive calls in vfio_pin_pages_remote()


 drivers/vfio/vfio_iommu_type1.c |  150 +++++++++++++++++----------------------
 1 file changed, 64 insertions(+), 86 deletions(-)

[toc] | [next] | [standalone]

#1624893 — [PATCH v5 2/3] vfio/type1: Prune vfio_pin_page_external()

With vfio_lock_acct() testing the locked memory limit under mmap_sem,
it's redundant to do it here for a single page.  We can also reorder
our tests such that we can avoid testing for reserved pages if we're
not doing accounting and let vfio_lock_acct() test the process
CAP_IPC_LOCK.  Finally, this function oddly returns 1 on success.
Update to return zero on success, -errno on error.  Since the function
only pins a single page, there's no need to return the number of pages
pinned.

N.B. vfio_pin_pages_remote() can pin a large contiguous range of pages
before calling vfio_lock_acct().  If we were to similarly remove the
extra test there, a user could temporarily pin far more pages than
they're allowed.

Suggested-by: Kirti Wankhede <kwankhede@nvidia.com>
Suggested-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Kirti Wankhede <kwankhede@nvidia.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
---
 drivers/vfio/vfio_iommu_type1.c |   35 ++++++++---------------------------
 1 file changed, 8 insertions(+), 27 deletions(-)

diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
index a8a079ba9477..372e4f626138 100644
--- a/drivers/vfio/vfio_iommu_type1.c
+++ b/drivers/vfio/vfio_iommu_type1.c
@@ -482,43 +482,26 @@ static long vfio_unpin_pages_remote(struct vfio_dma *dma, dma_addr_t iova,
 static int vfio_pin_page_external(struct vfio_dma *dma, unsigned long vaddr,
 				  unsigned long *pfn_base, bool do_accounting)
 {
-	unsigned long limit;
-	bool lock_cap = has_capability(dma->task, CAP_IPC_LOCK);
 	struct mm_struct *mm;
 	int ret;
-	bool rsvd;
 
 	mm = get_task_mm(dma->task);
 	if (!mm)
 		return -ENODEV;
 
 	ret = vaddr_get_pfn(mm, vaddr, dma->prot, pfn_base);
-	if (ret)
-		goto pin_page_exit;
-
-	rsvd = is_invalid_reserved_pfn(*pfn_base);
-	limit = task_rlimit(dma->task, RLIMIT_MEMLOCK) >> PAGE_SHIFT;
-
-	if (!rsvd && !lock_cap && mm->locked_vm + 1 > limit) {
-		put_pfn(*pfn_base, dma->prot);
-		pr_warn("%s: Task %s (%d) RLIMIT_MEMLOCK (%ld) exceeded\n",
-			__func__, dma->task->comm, task_pid_nr(dma->task),
-			limit << PAGE_SHIFT);
-		ret = -ENOMEM;
-		goto pin_page_exit;
-	}
-
-	if (!rsvd && do_accounting) {
-		ret = vfio_lock_acct(dma->task, 1, &lock_cap);
+	if (!ret && do_accounting && !is_invalid_reserved_pfn(*pfn_base)) {
+		ret = vfio_lock_acct(dma->task, 1, NULL);
 		if (ret) {
 			put_pfn(*pfn_base, dma->prot);
-			goto pin_page_exit;
+			if (ret == -ENOMEM)
+				pr_warn("%s: Task %s (%d) RLIMIT_MEMLOCK "
+					"(%ld) exceeded\n", __func__,
+					dma->task->comm, task_pid_nr(dma->task),
+					task_rlimit(dma->task, RLIMIT_MEMLOCK));
 		}
 	}
 
-	ret = 1;
-
-pin_page_exit:
 	mmput(mm);
 	return ret;
 }
@@ -598,10 +581,8 @@ static int vfio_iommu_type1_pin_pages(void *iommu_data,
 		remote_vaddr = dma->vaddr + iova - dma->iova;
 		ret = vfio_pin_page_external(dma, remote_vaddr, &phys_pfn[i],
 					     do_accounting);
-		if (ret <= 0) {
-			WARN_ON(!ret);
+		if (ret)
 			goto pin_unwind;
-		}
 
 		ret = vfio_add_to_pfn_list(dma, iova, phys_pfn[i]);
 		if (ret) {

[toc] | [prev] | [next] | [standalone]

#1625012

On Mon, Apr 17, 2017 at 04:37:38PM -0600, Alex Williamson wrote:
> v5:
>   - patch 1/ Use bool* to cleanup vfio_lock_acct() callers; sorry
>              we cannot re-test CAP_IPC_LOCK for all callers
>   - patch 2/ Re-add pr_warn, add Kirti's R-b
>   - patch 3/ NEW, analyzing impact of vfio_lock_acct() testing
>              CAP_IPC_LOCK for all callers revealed a long hanging
>              optimization
> 
> Thanks for the reviews, keep 'em coming,

All patches looks good to me.

Reviewed-by: Peter Xu <peterx@redhat.com>

Thanks!

-- 
Peter Xu

[toc] | [prev] | [next] | [standalone]

#1625542

On 4/18/2017 9:53 AM, Peter Xu wrote:
> On Mon, Apr 17, 2017 at 04:37:38PM -0600, Alex Williamson wrote:
>> v5:
>>   - patch 1/ Use bool* to cleanup vfio_lock_acct() callers; sorry
>>              we cannot re-test CAP_IPC_LOCK for all callers
>>   - patch 2/ Re-add pr_warn, add Kirti's R-b
>>   - patch 3/ NEW, analyzing impact of vfio_lock_acct() testing
>>              CAP_IPC_LOCK for all callers revealed a long hanging
>>              optimization
>>
>> Thanks for the reviews, keep 'em coming,
> 
> All patches looks good to me.
> 
> Reviewed-by: Peter Xu <peterx@redhat.com>
> 
> Thanks!
> 

All three looks good to me too.

Reviewed-by: Kirti Wankhede <kwankhede@nvidia.com>

Thanks,
Kirti

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web