[RFC] hw/net/vmxnet3: allow VMXNET3_MAX_MTU itself as a value

[RFC] hw/net/vmxnet3: allow VMXNET3_MAX_MTU itself as a value

[Fiona Ebner]

Fixes: d05dcd94ae ("net: vmxnet3: validate configuration values during activate (CVE-2021-20203)")
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

I'm not familiar with this code, so really I'm asking: is the change
justified?

I tested the change and it seems to work, but I only have some rough
rationale for it, which is also why there's no commit message yet.

In the Linux kernel's net/core/dev.c, in dev_validate_mtu(), the upper
limit itself is a valid value:
    if (dev->max_mtu > 0 && new_mtu > dev->max_mtu) {
        NL_SET_ERR_MSG(extack, "mtu greater than device maximum");
        return -EINVAL;
    }
and AFAICT in the case of the vmxnet3 driver, max_mtu is set to
VMXNET3_MAX_MTU (as defined in the kernel, which is 9000, same as in
QEMU).

Reported by one of our users running into the failing assert():
https://forum.proxmox.com/threads/114011/#post-492916

 hw/net/vmxnet3.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/net/vmxnet3.c b/hw/net/vmxnet3.c
index 0b7acf7f89..a2037583bf 100644
--- a/hw/net/vmxnet3.c
+++ b/hw/net/vmxnet3.c
@@ -1441,7 +1441,7 @@ static void vmxnet3_activate_device(VMXNET3State *s)
     vmxnet3_setup_rx_filtering(s);
     /* Cache fields from shared memory */
     s->mtu = VMXNET3_READ_DRV_SHARED32(d, s->drv_shmem, devRead.misc.mtu);
-    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu < VMXNET3_MAX_MTU);
+    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu <= VMXNET3_MAX_MTU);
     VMW_CFPRN("MTU is %u", s->mtu);
 
     s->max_rx_frags =
-- 
2.30.2

Re: [RFC] hw/net/vmxnet3: allow VMXNET3_MAX_MTU itself as a value

[Jason Wang]

On Wed, Aug 24, 2022 at 7:17 PM Fiona Ebner <f.ebner@proxmox.com> wrote:
>
> Fixes: d05dcd94ae ("net: vmxnet3: validate configuration values during activate (CVE-2021-20203)")
> Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
> ---
>
> I'm not familiar with this code, so really I'm asking: is the change
> justified?

Patch looks good, but please re-submit with a formal one with
rationals via changelog.

Thanks

>
> I tested the change and it seems to work, but I only have some rough
> rationale for it, which is also why there's no commit message yet.
>
> In the Linux kernel's net/core/dev.c, in dev_validate_mtu(), the upper
> limit itself is a valid value:
>     if (dev->max_mtu > 0 && new_mtu > dev->max_mtu) {
>         NL_SET_ERR_MSG(extack, "mtu greater than device maximum");
>         return -EINVAL;
>     }
> and AFAICT in the case of the vmxnet3 driver, max_mtu is set to
> VMXNET3_MAX_MTU (as defined in the kernel, which is 9000, same as in
> QEMU).
>
> Reported by one of our users running into the failing assert():
> https://forum.proxmox.com/threads/114011/#post-492916
>
>  hw/net/vmxnet3.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/hw/net/vmxnet3.c b/hw/net/vmxnet3.c
> index 0b7acf7f89..a2037583bf 100644
> --- a/hw/net/vmxnet3.c
> +++ b/hw/net/vmxnet3.c
> @@ -1441,7 +1441,7 @@ static void vmxnet3_activate_device(VMXNET3State *s)
>      vmxnet3_setup_rx_filtering(s);
>      /* Cache fields from shared memory */
>      s->mtu = VMXNET3_READ_DRV_SHARED32(d, s->drv_shmem, devRead.misc.mtu);
> -    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu < VMXNET3_MAX_MTU);
> +    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu <= VMXNET3_MAX_MTU);
>      VMW_CFPRN("MTU is %u", s->mtu);
>
>      s->max_rx_frags =
> --
> 2.30.2
>
>

Re: [RFC] hw/net/vmxnet3: allow VMXNET3_MAX_MTU itself as a value

[P J P]

On Wednesday, 24 August, 2022, 04:46:21 pm IST, Fiona Ebner <f.ebner@proxmox.com> wrote:
>Reported by one of our users running into the failing assert():
>https://forum.proxmox.com/threads/114011/#post-492916
>
>- assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu < VMXNET3_MAX_MTU);
>+ assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu <= VMXNET3_MAX_MTU);
> VMW_CFPRN("MTU is %u", s->mtu);
>

* I wonder if setting s->mtu == buffer_upper_limit may lead to an out-of-bounds access issue?

* IIUC, VMXNET3_MAX_MTU OR s->mtu does not seem to be used to allocate and/or access packet buffer(s)
  so above check might work, but still it does not seem clean, ie. it may lead to some confusion.

* Nonetheless, Jason has acked it, so that's good.

Thank you.
---
  -P J P
http://feedmug.com