From: Przemek Kitszel <przemyslaw.kitszel@intel.com>
To: Jacob Keller <jacob.e.keller@intel.com>,
Paul Menzel <pmenzel@molgen.mpg.de>,
Intel Wired LAN <intel-wired-lan@lists.osuosl.org>
Cc: Paul Greenwalt <paul.greenwalt@intel.com>
Subject: Re: [Intel-wired-lan] [PATCH iwl-net v3] ice: avoid infinite loop if NVM has invalid TLV length
Date: Mon, 20 May 2024 10:33:05 +0200 [thread overview]
Message-ID: <2da2aa2d-5d65-4250-b39f-6b09eb07239d@intel.com> (raw)
In-Reply-To: <20240517-iwl-net-2024-05-16-fix-nvm-tlv-issue-v3-1-f46c53cfb67f@intel.com>
On 5/18/24 01:22, Jacob Keller wrote:
> The ice_get_pfa_module_tlv() function iterates over the TLVs in the
> Preserved Fields Area (PFA) of the NVM. This is used to access data such as
> the Part Board Assembly identifier.
>
> Some NVMs in the wild have been found with incorrect TLV lengths including
> at least one which reports a TLV length of 0xFFFF. When trying to read the
> PBA from such an NVM, the driver will compute a new offset for the next_tlv
> which is lower, due to overflowing the 16-bit next_tlv variable.
>
> In the best case, the driver will incorrectly interpret values until it
> finds one which has an offset greater than the PFA area without
> overflowing. In the worst case, the values in the NVM result in an infinite
> loop as the misinterpreted lengths result in checking offsets which are
> valid within the PFA, and which ultimately point in an infinite loop.
>
> Fix this by using check_add_overflow when calculating the NVM offsets, and
> bailing if we ever overflow. Additionally, use check_add_overflow when
> calculating the initial maximum PFA size.
>
> This ensures that we bail immediately on encountering any TLV who's length
> would have caused the naive addition to overflow, rather than entering an
> infinite loop or otherwise misinterpreting NVM values.
>
> Fixes: e961b679fb0b ("ice: add board identifier info to devlink .info_get")
> Co-developed-by: Paul Greenwalt <paul.greenwalt@intel.com>
> Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
> ---
> Changes in v3:
> - Fix missing {
> - Fix missing pfa_ptr variable to dev_warn()
> - Add Fixes tag
> - Link to v2: https://lore.kernel.org/r/20240517-iwl-net-2024-05-16-fix-nvm-tlv-issue-v2-1-fdee184ece86@intel.com
>
> Changes in v2:
> - Use check_add_overflow instead of increasing the variables to u32
> - Upgrade log messages to dev_warn()
> - Link to v1: https://lore.kernel.org/r/20240516-iwl-net-2024-05-16-fix-nvm-tlv-issue-v1-1-ecbb6a75961e@intel.com
> ---
> drivers/net/ethernet/intel/ice/ice_nvm.c | 19 +++++++++++++++----
> 1 file changed, 15 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/ice/ice_nvm.c b/drivers/net/ethernet/intel/ice/ice_nvm.c
> index 84eab92dc03c..be731b83d667 100644
> --- a/drivers/net/ethernet/intel/ice/ice_nvm.c
> +++ b/drivers/net/ethernet/intel/ice/ice_nvm.c
> @@ -440,8 +440,7 @@ int
> ice_get_pfa_module_tlv(struct ice_hw *hw, u16 *module_tlv, u16 *module_tlv_len,
> u16 module_type)
> {
> - u16 pfa_len, pfa_ptr;
> - u16 next_tlv;
> + u16 pfa_len, pfa_ptr, next_tlv, max_pfa;
> int status;
>
> status = ice_read_sr_word(hw, ICE_SR_PFA_PTR, &pfa_ptr);
> @@ -454,11 +453,18 @@ ice_get_pfa_module_tlv(struct ice_hw *hw, u16 *module_tlv, u16 *module_tlv_len,
> ice_debug(hw, ICE_DBG_INIT, "Failed to read PFA length.\n");
> return status;
> }
> +
> + if (check_add_overflow(pfa_ptr, pfa_len, &max_pfa)) {
> + dev_warn(ice_hw_to_dev(hw), "PFA starts at offset %u. PFA length of %u causes 16-bit arithmetic overflow.\n",
> + pfa_ptr, pfa_len);
> + return -EINVAL;
> + }
> +
> /* Starting with first TLV after PFA length, iterate through the list
> * of TLVs to find the requested one.
> */
> next_tlv = pfa_ptr + 1;
> - while (next_tlv < pfa_ptr + pfa_len) {
> + while (next_tlv < max_pfa) {
> u16 tlv_sub_module_type;
> u16 tlv_len;
>
> @@ -485,7 +491,12 @@ ice_get_pfa_module_tlv(struct ice_hw *hw, u16 *module_tlv, u16 *module_tlv_len,
> /* Check next TLV, i.e. current TLV pointer + length + 2 words
> * (for current TLV's type and length)
> */
> - next_tlv = next_tlv + tlv_len + 2;
> + if (check_add_overflow(next_tlv, 2, &next_tlv) ||
> + check_add_overflow(next_tlv, tlv_len, &next_tlv)) {
> + dev_warn(ice_hw_to_dev(hw), "Failed to locate PFA TLV module of type %u due to arithmetic overflow. The PFA length is %u. The last TLV has length of %u\n",
> + module_type, pfa_len, tlv_len);
> + return -EINVAL;
> + }
> }
> /* Module does not exist */
> return -ENOENT;
>
> ---
> base-commit: 83e93942796db58652288f0391ac00072401816f
> change-id: 20240516-iwl-net-2024-05-16-fix-nvm-tlv-issue-99ebb2c55c52
>
> Best regards,
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
next prev parent reply other threads:[~2024-05-20 8:33 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-17 23:22 [Intel-wired-lan] [PATCH iwl-net v3] ice: avoid infinite loop if NVM has invalid TLV length Jacob Keller
2024-05-20 8:33 ` Przemek Kitszel [this message]
2024-05-22 18:33 ` Jacob Keller
2024-05-24 20:31 ` Jacob Keller
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=2da2aa2d-5d65-4250-b39f-6b09eb07239d@intel.com \
--to=przemyslaw.kitszel@intel.com \
--cc=intel-wired-lan@lists.osuosl.org \
--cc=jacob.e.keller@intel.com \
--cc=paul.greenwalt@intel.com \
--cc=pmenzel@molgen.mpg.de \
/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