stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Jason Gunthorpe <jgg@mellanox.com>,
	Jack Morgenstein <jackm@dev.mellanox.co.il>,
	Leon Romanovsky <leonro@mellanox.com>,
	Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Subject: [PATCH 4.4 14/43] IB/mlx4: Mark user MR as writable if actual virtual memory is writable
Date: Tue, 14 Aug 2018 19:17:50 +0200	[thread overview]
Message-ID: <20180814171518.046291292@linuxfoundation.org> (raw)
In-Reply-To: <20180814171517.014285600@linuxfoundation.org>

4.4-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Jack Morgenstein <jackm@dev.mellanox.co.il>

commit d8f9cc328c8888369880e2527e9186d745f2bbf6 upstream.

To allow rereg_user_mr to modify the MR from read-only to writable without
using get_user_pages again, we needed to define the initial MR as writable.
However, this was originally done unconditionally, without taking into
account the writability of the underlying virtual memory.

As a result, any attempt to register a read-only MR over read-only
virtual memory failed.

To fix this, do not add the writable flag bit when the user virtual memory
is not writable (e.g. const memory).

However, when the underlying memory is NOT writable (and we therefore
do not define the initial MR as writable), the IB core adds a
"force writable" flag to its user-pages request. If this succeeds,
the reg_user_mr caller gets a writable copy of the original pages.

If the user-space caller then does a rereg_user_mr operation to enable
writability, this will succeed. This should not be allowed, since
the original virtual memory was not writable.

Cc: <stable@vger.kernel.org>
Fixes: 9376932d0c26 ("IB/mlx4_ib: Add support for user MR re-registration")
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/infiniband/hw/mlx4/mr.c |   50 +++++++++++++++++++++++++++++++++-------
 1 file changed, 42 insertions(+), 8 deletions(-)

--- a/drivers/infiniband/hw/mlx4/mr.c
+++ b/drivers/infiniband/hw/mlx4/mr.c
@@ -130,6 +130,40 @@ out:
 	return err;
 }
 
+static struct ib_umem *mlx4_get_umem_mr(struct ib_ucontext *context, u64 start,
+					u64 length, u64 virt_addr,
+					int access_flags)
+{
+	/*
+	 * Force registering the memory as writable if the underlying pages
+	 * are writable.  This is so rereg can change the access permissions
+	 * from readable to writable without having to run through ib_umem_get
+	 * again
+	 */
+	if (!ib_access_writable(access_flags)) {
+		struct vm_area_struct *vma;
+
+		down_read(&current->mm->mmap_sem);
+		/*
+		 * FIXME: Ideally this would iterate over all the vmas that
+		 * cover the memory, but for now it requires a single vma to
+		 * entirely cover the MR to support RO mappings.
+		 */
+		vma = find_vma(current->mm, start);
+		if (vma && vma->vm_end >= start + length &&
+		    vma->vm_start <= start) {
+			if (vma->vm_flags & VM_WRITE)
+				access_flags |= IB_ACCESS_LOCAL_WRITE;
+		} else {
+			access_flags |= IB_ACCESS_LOCAL_WRITE;
+		}
+
+		up_read(&current->mm->mmap_sem);
+	}
+
+	return ib_umem_get(context, start, length, access_flags, 0);
+}
+
 struct ib_mr *mlx4_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 				  u64 virt_addr, int access_flags,
 				  struct ib_udata *udata)
@@ -144,10 +178,8 @@ struct ib_mr *mlx4_ib_reg_user_mr(struct
 	if (!mr)
 		return ERR_PTR(-ENOMEM);
 
-	/* Force registering the memory as writable. */
-	/* Used for memory re-registeration. HCA protects the access */
-	mr->umem = ib_umem_get(pd->uobject->context, start, length,
-			       access_flags | IB_ACCESS_LOCAL_WRITE, 0);
+	mr->umem = mlx4_get_umem_mr(pd->uobject->context, start, length,
+				    virt_addr, access_flags);
 	if (IS_ERR(mr->umem)) {
 		err = PTR_ERR(mr->umem);
 		goto err_free;
@@ -214,6 +246,9 @@ int mlx4_ib_rereg_user_mr(struct ib_mr *
 	}
 
 	if (flags & IB_MR_REREG_ACCESS) {
+		if (ib_access_writable(mr_access_flags) && !mmr->umem->writable)
+			return -EPERM;
+
 		err = mlx4_mr_hw_change_access(dev->dev, *pmpt_entry,
 					       convert_access(mr_access_flags));
 
@@ -227,10 +262,9 @@ int mlx4_ib_rereg_user_mr(struct ib_mr *
 
 		mlx4_mr_rereg_mem_cleanup(dev->dev, &mmr->mmr);
 		ib_umem_release(mmr->umem);
-		mmr->umem = ib_umem_get(mr->uobject->context, start, length,
-					mr_access_flags |
-					IB_ACCESS_LOCAL_WRITE,
-					0);
+		mmr->umem =
+			mlx4_get_umem_mr(mr->uobject->context, start, length,
+					 virt_addr, mr_access_flags);
 		if (IS_ERR(mmr->umem)) {
 			err = PTR_ERR(mmr->umem);
 			/* Prevent mlx4_ib_dereg_mr from free'ing invalid pointer */

  parent reply	other threads:[~2018-08-14 20:34 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-08-14 17:17 [PATCH 4.4 00/43] 4.4.148-stable review Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 01/43] ext4: fix check to prevent initializing reserved inodes Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 02/43] tpm: fix race condition in tpm_common_write() Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 03/43] ipv4+ipv6: Make INET*_ESP select CRYPTO_ECHAINIV Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 04/43] fork: unconditionally clear stack on fork Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 05/43] parisc: Enable CONFIG_MLONGCALLS by default Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 07/43] xen/netfront: dont cache skb_shinfo() Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 08/43] ACPI / LPSS: Add missing prv_offset setting for byt/cht PWM devices Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 09/43] scsi: sr: Avoid that opening a CD-ROM hangs with runtime power management enabled Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 10/43] root dentries need RCU-delayed freeing Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 11/43] fix mntput/mntput race Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 12/43] fix __legitimize_mnt()/mntput() race Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 13/43] IB/core: Make testing MR flags for writability a static inline function Greg Kroah-Hartman
2018-08-14 17:17 ` Greg Kroah-Hartman [this message]
2018-08-14 17:17 ` [PATCH 4.4 15/43] IB/ocrdma: fix out of bounds access to local buffer Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 16/43] ARM: dts: imx6sx: fix irq for pcie bridge Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 17/43] x86/paravirt: Fix spectre-v2 mitigations for paravirt guests Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 18/43] x86/speculation: Protect against userspace-userspace spectreRSB Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 19/43] kprobes/x86: Fix %p uses in error messages Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 20/43] x86/irqflags: Provide a declaration for native_save_fl Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 21/43] x86/speculation/l1tf: Increase 32bit PAE __PHYSICAL_PAGE_SHIFT Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 22/43] x86/mm: Move swap offset/type up in PTE to work around erratum Greg Kroah-Hartman
2018-08-14 17:17 ` [PATCH 4.4 23/43] x86/mm: Fix swap entry comment and macro Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 24/43] mm: x86: move _PAGE_SWP_SOFT_DIRTY from bit 7 to bit 1 Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 25/43] x86/speculation/l1tf: Change order of offset/type in swap entry Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 26/43] x86/speculation/l1tf: Protect swap entries against L1TF Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 27/43] x86/speculation/l1tf: Protect PROT_NONE PTEs against speculation Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 28/43] x86/speculation/l1tf: Make sure the first page is always reserved Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 29/43] x86/speculation/l1tf: Add sysfs reporting for l1tf Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 30/43] mm: Add vm_insert_pfn_prot() Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 31/43] mm: fix cache mode tracking in vm_insert_mixed() Greg Kroah-Hartman
2018-09-07 17:05   ` Ben Hutchings
2018-09-07 20:03     ` Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 32/43] x86/speculation/l1tf: Disallow non privileged high MMIO PROT_NONE mappings Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 33/43] x86/speculation/l1tf: Limit swap file size to MAX_PA/2 Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 34/43] x86/bugs: Move the l1tf function and define pr_fmt properly Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 35/43] x86/speculation/l1tf: Extend 64bit swap file size limit Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 36/43] x86/cpufeatures: Add detection of L1D cache flush support Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 37/43] x86/speculation/l1tf: Protect PAE swap entries against L1TF Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 38/43] x86/speculation/l1tf: Fix up pte->pfn conversion for PAE Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 39/43] x86/speculation/l1tf: Invert all not present mappings Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 40/43] x86/speculation/l1tf: Make pmd/pud_mknotpresent() invert Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 41/43] x86/mm/pat: Make set_memory_np() L1TF safe Greg Kroah-Hartman
2018-09-09 16:46   ` Ben Hutchings
2018-09-09 17:06     ` Guenter Roeck
2018-09-10  7:16       ` Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 42/43] x86/mm/kmmio: Make the tracer robust against L1TF Greg Kroah-Hartman
2018-08-14 17:18 ` [PATCH 4.4 43/43] x86/speculation/l1tf: Fix up CPU feature flags Greg Kroah-Hartman
2018-08-15  6:15 ` [PATCH 4.4 00/43] 4.4.148-stable review Greg Kroah-Hartman
2018-08-15 13:10 ` Guenter Roeck
2018-08-15 20:52 ` Dan Rue

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=20180814171518.046291292@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=jackm@dev.mellanox.co.il \
    --cc=jgg@mellanox.com \
    --cc=leonro@mellanox.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=sudipm.mukherjee@gmail.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).