diff mbox series

[v2] Stop vhost-user sending uninitialized mmap_offsets

Message ID 1592650156-25845-1-git-send-email-raphael.norwitz@nutanix.com (mailing list archive)
State New, archived
Headers show
Series [v2] Stop vhost-user sending uninitialized mmap_offsets | expand

Commit Message

Raphael Norwitz June 22, 2020, 11:50 p.m. UTC
Prior to this change, the vhost_user_fill_msg_region function filled out
all elements of the VhostUserMemoryRegion struct except the mmap_offset.

This function is often called on uninitialized structs, which are then
copied into VHOST_USER_SET_MEM_TABLE and VHOST_USER_ADD/REM_MEM_REG
messages. In some cases, where the mmap_offset was not needed, it was
left uninitialized, causing QEMU to send the backend uninitialized data,
which Coverity flagged as a series of issues.

This change augments the vhost_user_fill_msg_region API, adding a
mmap_offset paramenter, forcing the caller to initialize mmap_offset.

Fixes: ece99091c2d0aeb23734289a50ef2ff4e0a08929
Fixes: f1aeb14b0809e313c74244d838645ed25e85ea63
Reported-by: Coverity (CIDs 1429802, 1429803 and 1429804)
Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
---
 hw/virtio/vhost-user.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

Comments

Peter Maydell June 23, 2020, 8:58 a.m. UTC | #1
On Tue, 23 Jun 2020 at 00:50, Raphael Norwitz
<raphael.norwitz@nutanix.com> wrote:
>
> Prior to this change, the vhost_user_fill_msg_region function filled out
> all elements of the VhostUserMemoryRegion struct except the mmap_offset.
>
> This function is often called on uninitialized structs, which are then
> copied into VHOST_USER_SET_MEM_TABLE and VHOST_USER_ADD/REM_MEM_REG
> messages. In some cases, where the mmap_offset was not needed, it was
> left uninitialized, causing QEMU to send the backend uninitialized data,
> which Coverity flagged as a series of issues.
>
> This change augments the vhost_user_fill_msg_region API, adding a
> mmap_offset paramenter, forcing the caller to initialize mmap_offset.
>
> Fixes: ece99091c2d0aeb23734289a50ef2ff4e0a08929
> Fixes: f1aeb14b0809e313c74244d838645ed25e85ea63
> Reported-by: Coverity (CIDs 1429802, 1429803 and 1429804)
> Suggested-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
> ---


Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM
Michael S. Tsirkin June 23, 2020, 9:07 a.m. UTC | #2
On Tue, Jun 23, 2020 at 09:58:23AM +0100, Peter Maydell wrote:
> On Tue, 23 Jun 2020 at 00:50, Raphael Norwitz
> <raphael.norwitz@nutanix.com> wrote:
> >
> > Prior to this change, the vhost_user_fill_msg_region function filled out
> > all elements of the VhostUserMemoryRegion struct except the mmap_offset.
> >
> > This function is often called on uninitialized structs, which are then
> > copied into VHOST_USER_SET_MEM_TABLE and VHOST_USER_ADD/REM_MEM_REG
> > messages. In some cases, where the mmap_offset was not needed, it was
> > left uninitialized, causing QEMU to send the backend uninitialized data,
> > which Coverity flagged as a series of issues.
> >
> > This change augments the vhost_user_fill_msg_region API, adding a
> > mmap_offset paramenter, forcing the caller to initialize mmap_offset.
> >
> > Fixes: ece99091c2d0aeb23734289a50ef2ff4e0a08929
> > Fixes: f1aeb14b0809e313c74244d838645ed25e85ea63
> > Reported-by: Coverity (CIDs 1429802, 1429803 and 1429804)
> > Suggested-by: Peter Maydell <peter.maydell@linaro.org>
> > Signed-off-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
> > ---
> 
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> 
> thanks
> -- PMM

Queued, thanks!
Stefan Hajnoczi June 23, 2020, 3:06 p.m. UTC | #3
On Mon, Jun 22, 2020 at 11:50:44PM +0000, Raphael Norwitz wrote:
> Prior to this change, the vhost_user_fill_msg_region function filled out
> all elements of the VhostUserMemoryRegion struct except the mmap_offset.
> 
> This function is often called on uninitialized structs, which are then
> copied into VHOST_USER_SET_MEM_TABLE and VHOST_USER_ADD/REM_MEM_REG
> messages. In some cases, where the mmap_offset was not needed, it was
> left uninitialized, causing QEMU to send the backend uninitialized data,
> which Coverity flagged as a series of issues.
> 
> This change augments the vhost_user_fill_msg_region API, adding a
> mmap_offset paramenter, forcing the caller to initialize mmap_offset.
> 
> Fixes: ece99091c2d0aeb23734289a50ef2ff4e0a08929
> Fixes: f1aeb14b0809e313c74244d838645ed25e85ea63
> Reported-by: Coverity (CIDs 1429802, 1429803 and 1429804)
> Suggested-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
> ---
>  hw/virtio/vhost-user.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
diff mbox series

Patch

diff --git a/hw/virtio/vhost-user.c b/hw/virtio/vhost-user.c
index 4d6cd4e..3123121 100644
--- a/hw/virtio/vhost-user.c
+++ b/hw/virtio/vhost-user.c
@@ -460,12 +460,14 @@  static MemoryRegion *vhost_user_get_mr_data(uint64_t addr, ram_addr_t *offset,
 }
 
 static void vhost_user_fill_msg_region(VhostUserMemoryRegion *dst,
-                                       struct vhost_memory_region *src)
+                                       struct vhost_memory_region *src,
+                                       uint64_t mmap_offset)
 {
     assert(src != NULL && dst != NULL);
     dst->userspace_addr = src->userspace_addr;
     dst->memory_size = src->memory_size;
     dst->guest_phys_addr = src->guest_phys_addr;
+    dst->mmap_offset = mmap_offset;
 }
 
 static int vhost_user_fill_set_mem_table_msg(struct vhost_user *u,
@@ -500,9 +502,8 @@  static int vhost_user_fill_set_mem_table_msg(struct vhost_user *u,
                 error_report("Failed preparing vhost-user memory table msg");
                 return -1;
             }
-            vhost_user_fill_msg_region(&region_buffer, reg);
+            vhost_user_fill_msg_region(&region_buffer, reg, offset);
             msg->payload.memory.regions[*fd_num] = region_buffer;
-            msg->payload.memory.regions[*fd_num].mmap_offset = offset;
             fds[(*fd_num)++] = fd;
         } else if (track_ramblocks) {
             u->region_rb_offset[i] = 0;
@@ -649,7 +650,7 @@  static int send_remove_regions(struct vhost_dev *dev,
 
         if (fd > 0) {
             msg->hdr.request = VHOST_USER_REM_MEM_REG;
-            vhost_user_fill_msg_region(&region_buffer, shadow_reg);
+            vhost_user_fill_msg_region(&region_buffer, shadow_reg, 0);
             msg->payload.mem_reg.region = region_buffer;
 
             if (vhost_user_write(dev, msg, &fd, 1) < 0) {
@@ -709,9 +710,8 @@  static int send_add_regions(struct vhost_dev *dev,
                 u->region_rb[reg_idx] = mr->ram_block;
             }
             msg->hdr.request = VHOST_USER_ADD_MEM_REG;
-            vhost_user_fill_msg_region(&region_buffer, reg);
+            vhost_user_fill_msg_region(&region_buffer, reg, offset);
             msg->payload.mem_reg.region = region_buffer;
-            msg->payload.mem_reg.region.mmap_offset = offset;
 
             if (vhost_user_write(dev, msg, &fd, 1) < 0) {
                 return -1;