Read the docs

Linux kernel nft_byteorder: broken 64-bit multi-register byteorder conversion

Priority LOW 22Published · Updated

Key facts

Attack vector
—
Privileges required
—
User interaction
—
Attack complexity
—
Impact type
—
Confidentiality · integrity · availability
—
Fix status
Partial Red Hat: 4 without a fix

No CVSS vector from any source. Fix status from CNA and Red Hat.

Summary

In the Linux kernel netfilter nft_byteorder, 64-bit byteorder conversion is broken when several registers are converted, because the source register array advances in 4-byte steps instead of 8. The multi-register feature is removed, since commit caf3ef7468f7 had already changed its semantics so the loop operates on bytes rather than registers. Pablo Neira Ayuso notes that nftables before 1.1.0 can generate incorrect byteorder conversions, and affected rulesets fail to load with this change and old userspace because of a 'len != size' check. Red Hat rates it Moderate.

Written with Claude Haiku 5.5 from the sources below · 2026-11-11

Description from the CNA

In the Linux kernel, the following vulnerability has been resolved: netfilter: nft_byteorder: remove multi-register support 64bit byteorder conversion is broken when several registers need to be converted because the source register array advances in steps for 4 bytes instead of 8: for (i = ... src64 = nft_reg_load64(&src[i]); ~~~~~ u32 *src nft_reg_store64(&dst64[i], Remove the multi-register support, it has other issues as well: Pablo points out that commit caf3ef7468f7 ("netfilter: nf_tables: prevent OOB access in nft_byteorder_eval") alters semantics: before the loop operated on registers, i.e. for ( ... ) dst32[i] = htons((u16)src32[i]) .. but after the patch it will operate on bytes, which makes this useless to convert e.g. concatenations, which store each compound in its own register. Multi-convert of u32 has one theoretical application: ct mark . meta mark . tcp dport @intervalset Because ct mark and meta mark are host byte order, use with intervals has to convert the byteorder for ct/meta mark value to network byte order (bigendian). nftables emits this: [ meta load mark => reg 1 ] [ byteorder reg 1 = hton(reg 1, 4, 4) ] [ ct load mark => reg 9 ] [ byteorder reg 9 = hton(reg 9, 4, 4) ] ... I.e. two separate calls. Theoretically it could be changed to do: [ meta load mark => reg 1 ] [ ct load mark => reg 9 ] [ byteorder reg 1 = htonl(reg 1, 4, 8) ] ... But then all it would take to change the set to meta mark . tcp dport . ct mark ... and we'd be back to two "byteorder" calls. IOW, support to convert a range of registers is both dysfunctional and dubious. Simplify this: remove the feature. Pablo Neira Ayuso points out that nftables before 1.1.0 can generate incorrect byteorder conversions, see 9fe58952c45a, "evaluate: skip byteorder conversion for selector smaller than 2 bytes" in nftables.git). Affected rulesets fail to load with this change and old userspace due to 'len != size' check.

Why it matters now

EPSS 0.00167 5th percentileScored 2026-11-16
Fix Partial Linux (Linux): 4.20, 5.5, 5.11 +7Red Hat: 4 without a fix

SSVC decision points

Not available: CISA has not published SSVC decision points for this record.

Affected products

VendorProductAffected versionsFixed inStatusSource
LinuxLinux4.19.307 to before 4.20; 5.4.269 to before 5.5; 5.10.210 to before 5.11; 5.15.140 to before 5.16; 6.1.64 to before 6.2; 6.5.13 to before 6.6; 6.6.3 to before 6.74.20, 5.5, 5.11, 5.16, 6.2, 6.6, 6.7Patch availableCNA
LinuxLinux6.7 and later6.12.115, 6.18.59, 7.1Patch availableCNA
  • Linux VendorLinux Affected 4.19.307 to before 4.20; 5.4.269 to before 5.5; 5.10.210 to before 5.11; 5.15.140 to before 5.16; 6.1.64 to before 6.2; 6.5.13 to before 6.6; 6.6.3 to before 6.7 Fixed in 4.20, 5.5, 5.11, 5.16, 6.2, 6.6, 6.7 Patch available CNA
  • Linux VendorLinux Affected 6.7 and later Fixed in 6.12.115, 6.18.59, 7.1 Patch available CNA

CPE names Threatist (inferred) · 1

Inferred from how NVD named this product before; not given by a source.

CPEVersionsRoleVendor
cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*——Linux
  • cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:* Versions — — ·Linux

Remediation

Install a kernel that includes the fix for your branch: 4.20, 5.5, 5.11, 5.16, 6.2, 6.6, 6.7, 6.12.115, 6.18.59 or 7.1. Use nftables 1.1.0 or later in userspace, because older versions can generate incorrect byteorder conversions.

Written with Claude Haiku 5.5 from the sources below · 2026-11-11

  • Fix · CNA Linux (Linux) Fixed in 4.20, 5.5, 5.11, 5.16, 6.2, 6.6, 6.7
  • Fix · CNA Linux (Linux) Fixed in 6.12.115, 6.18.59, 7.1

Vendor assessments

Red Hat

Red Hat rating: Moderate. Red Hat's page for this CVE →

ProductStatusPlan or fix
Red Hat Enterprise Linux 10AffectedNo fix yet: Fix deferred
Red Hat Enterprise Linux 9AffectedNo fix yet: Fix deferred
Red Hat OpenShift Container Platform 4AffectedNo fix yet: Fix deferred
Red Hat OpenShift Container Platform 5AffectedNo fix yet: Fix deferred
Not affected: 5 products
  • Red Hat Enterprise Linux 6
  • Red Hat Enterprise Linux 7
  • Red Hat Enterprise Linux 8
  • Red Hat Enterprise Linux 9
  • Red Hat OpenShift Container Platform 4

Timeline

  1. 2026-11-10 Threatist record created from CVE-2026-99285
  2. 2026-11-10 Alias added: GHSA-vmhm-3cg3-fm3j
  3. 2026-11-11 Affected versions changed · Red Hat
  4. 2026-11-11 CVSS score added · Red Hat

References

  • TI-2025-26358 Zero-length TLS records mishandled on rx_list in Linux kernel KEVCVSS 3.1CRITICAL 9.8
  • TI-2026-38352 Out-of-bounds write in Linux kernel ebtables SNAT ARP rewrite KEVCVSS 3.1HIGH 8.8
  • TI-2021-02452 Overlayfs flaw in Linux kernel lets attackers gain elevated privileges KEVCVSS 3.1HIGH 8.8
  • TI-2013-05264 Missing address checks in Linux kernel get_user and put_user on ARM KEVCVSS 3.1HIGH 8.8
  • TI-2022-00125 Heap overflow in Linux kernel lets local users escalate privileges KEVCVSS 3.1HIGH 8.4
  • TI-2013-01864 Wrong integer type in Linux kernel perf_swevent_init KEVCVSS 3.1HIGH 8.4
See all 21768 →

Sources