All of lore.kernel.org
 help / color / mirror / Atom feed
From: Leon Romanovsky <leon@kernel.org>
To: Doug Ledford <dledford@redhat.com>, Jason Gunthorpe <jgg@mellanox.com>
Cc: Leon Romanovsky <leonro@mellanox.com>,
	RDMA mailing list <linux-rdma@vger.kernel.org>,
	Andrea Arcangeli <aarcange@redhat.com>,
	Feras Daoud <ferasda@mellanox.com>,
	Haggai Eran <haggaie@mellanox.com>,
	Jason Gunthorpe <jgg@ziepe.ca>,
	Saeed Mahameed <saeedm@mellanox.com>,
	linux-netdev <netdev@vger.kernel.org>
Subject: [PATCH rdma-next 4/6] RDMA/ucontext: Fix regression with disassociate
Date: Tue, 16 Apr 2019 14:07:28 +0300	[thread overview]
Message-ID: <20190416110730.32230-5-leon@kernel.org> (raw)
In-Reply-To: <20190416110730.32230-1-leon@kernel.org>

From: Jason Gunthorpe <jgg@mellanox.com>

When this code was consolidated the intention was that the VMA would
become backed by anonymous zero pages after the zap_vma_pte - however this
very subtly relied on setting the vm_ops = NULL and clearing the VM_SHARED
bits to transform the VMA into an anonymous VMA. Since the vm_ops was
removed this broke.

Now userspace gets a SIGBUS if it touches the vma after disassociation.

Instead of converting the VMA to anonymous provide a fault handler that
puts a zero'd page into the VMA when user-space touches it after
disassociation.

Cc: stable@vger.kernel.org
Suggested-by: Andrea Arcangeli <aarcange@redhat.com>
Fixes: 5f9794dc94f5 ("RDMA/ucontext: Add a core API for mmaping driver IO memory")
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
 drivers/infiniband/core/uverbs.h      |  1 +
 drivers/infiniband/core/uverbs_main.c | 51 +++++++++++++++++++++++++--
 2 files changed, 49 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h
index 0fc71ad42490..d2c29868172c 100644
--- a/drivers/infiniband/core/uverbs.h
+++ b/drivers/infiniband/core/uverbs.h
@@ -160,6 +160,7 @@ struct ib_uverbs_file {
 
 	struct mutex umap_lock;
 	struct list_head umaps;
+	struct page *disassociate_page;
 
 	struct idr		idr;
 	/* spinlock protects write access to idr */
diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index 3ef6474cd201..4a7cf5fddaee 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -208,6 +208,9 @@ void ib_uverbs_release_file(struct kref *ref)
 		kref_put(&file->async_file->ref,
 			 ib_uverbs_release_async_event_file);
 	put_device(&file->device->dev);
+
+	if (file->disassociate_page)
+		__free_pages(file->disassociate_page, 0);
 	kfree(file);
 }
 
@@ -877,9 +880,50 @@ static void rdma_umap_close(struct vm_area_struct *vma)
 	kfree(priv);
 }
 
+/*
+ * Once the zap_vma_ptes has been called touches to the VMA will come here and
+ * we return a dummy writable zero page for all the pfns.
+ */
+static vm_fault_t rdma_umap_fault(struct vm_fault *vmf)
+{
+	struct ib_uverbs_file *ufile = vmf->vma->vm_file->private_data;
+	struct rdma_umap_priv *priv = vmf->vma->vm_private_data;
+	vm_fault_t ret = 0;
+
+	if (!priv)
+		return VM_FAULT_SIGBUS;
+
+	/* Read only pages can just use the system zero page. */
+	if (!(vmf->vma->vm_flags & (VM_WRITE | VM_MAYWRITE))) {
+		vmf->page = ZERO_PAGE(vmf->vm_start);
+		get_page(vmf->page);
+		return 0;
+	}
+
+	mutex_lock(&ufile->umap_lock);
+	if (!ufile->disassociate_page)
+		ufile->disassociate_page =
+			alloc_pages(vmf->gfp_mask | __GFP_ZERO, 0);
+
+	if (ufile->disassociate_page) {
+		/*
+		 * This VMA is forced to always be shared so this doesn't have
+		 * to worry about COW.
+		 */
+		vmf->page = ufile->disassociate_page;
+		get_page(vmf->page);
+	} else {
+		ret = VM_FAULT_SIGBUS;
+	}
+	mutex_unlock(&ufile->umap_lock);
+
+	return ret;
+}
+
 static const struct vm_operations_struct rdma_umap_ops = {
 	.open = rdma_umap_open,
 	.close = rdma_umap_close,
+	.fault = rdma_umap_fault,
 };
 
 static struct rdma_umap_priv *rdma_user_mmap_pre(struct ib_ucontext *ucontext,
@@ -889,6 +933,8 @@ static struct rdma_umap_priv *rdma_user_mmap_pre(struct ib_ucontext *ucontext,
 	struct ib_uverbs_file *ufile = ucontext->ufile;
 	struct rdma_umap_priv *priv;
 
+	if (!(vma->vm_flags & VM_SHARED))
+		return ERR_PTR(-EINVAL);
 	if (vma->vm_flags & VM_EXEC)
 		return ERR_PTR(-EINVAL);
 	vma->vm_flags &= ~VM_MAYEXEC;
@@ -996,7 +1042,7 @@ void uverbs_user_mmap_disassociate(struct ib_uverbs_file *ufile)
 		 * at a time to get the lock ordering right. Typically there
 		 * will only be one mm, so no big deal.
 		 */
-		down_write(&mm->mmap_sem);
+		down_read(&mm->mmap_sem);
 		mutex_lock(&ufile->umap_lock);
 		list_for_each_entry_safe (priv, next_priv, &ufile->umaps,
 					  list) {
@@ -1008,10 +1054,9 @@ void uverbs_user_mmap_disassociate(struct ib_uverbs_file *ufile)
 
 			zap_vma_ptes(vma, vma->vm_start,
 				     vma->vm_end - vma->vm_start);
-			vma->vm_flags &= ~(VM_SHARED | VM_MAYSHARE);
 		}
 		mutex_unlock(&ufile->umap_lock);
-		up_write(&mm->mmap_sem);
+		up_read(&mm->mmap_sem);
 		mmput(mm);
 	}
 }
-- 
2.20.1

  parent reply	other threads:[~2019-04-16 11:07 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-16 11:07 [PATCH rdma-next 0/6] BAR mappings fixes in RDMA Leon Romanovsky
2019-04-16 11:07 ` Leon Romanovsky
2019-04-16 11:07 ` [PATCH rdma-next 1/6] RDMA/mlx5: Do not allow the user to write to the clock page Leon Romanovsky
2019-04-16 11:07 ` [PATCH rdma-next 2/6] RDMA/mlx5: Use rdma_user_map_io for mapping BAR pages Leon Romanovsky
2019-04-16 11:07 ` [PATCH rdma-next 3/6] RDMA/ucontext: Do not allow BAR mappings to be executable Leon Romanovsky
2019-04-17 19:05   ` Ruhl, Michael J
2019-04-18  5:58     ` Jason Gunthorpe
2019-04-18  6:30       ` Kees Cook
2019-04-18  7:01         ` Jason Gunthorpe
2019-04-18  7:23           ` Kees Cook
2019-04-18  7:23             ` Kees Cook
2019-04-22 12:51             ` Jason Gunthorpe
2019-04-16 11:07 ` Leon Romanovsky [this message]
2019-04-16 11:07 ` [PATCH mlx5-next 5/6] RDMA/mlx5: Use get_zeroed_page() for clock_info Leon Romanovsky
2019-04-16 11:07 ` [PATCH rdma-next 6/6] RDMA: Remove rdma_user_mmap_page Leon Romanovsky
2019-04-24 19:24 ` [PATCH rdma-next 0/6] BAR mappings fixes in RDMA Jason Gunthorpe

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190416110730.32230-5-leon@kernel.org \
    --to=leon@kernel.org \
    --cc=aarcange@redhat.com \
    --cc=dledford@redhat.com \
    --cc=ferasda@mellanox.com \
    --cc=haggaie@mellanox.com \
    --cc=jgg@mellanox.com \
    --cc=jgg@ziepe.ca \
    --cc=leonro@mellanox.com \
    --cc=linux-rdma@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=saeedm@mellanox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.