{"identifier":"CVE-2026-99285","cve":{"id":"CVE-2026-99285","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nnetfilter: nft_byteorder: remove multi-register support\n\n64bit byteorder conversion is broken when several registers need to be\nconverted because the source register array advances in steps for 4 bytes\ninstead of 8:\n\n  for (i = ...\n      src64 = nft_reg_load64(&src[i]);\n                             ~~~~~ u32 *src\n      nft_reg_store64(&dst64[i],\n\nRemove the multi-register support, it has other issues as well:\n\nPablo points out that commit\ncaf3ef7468f7 (\"netfilter: nf_tables: prevent OOB access in nft_byteorder_eval\")\nalters semantics: before the loop operated on registers, i.e.\n for ( ... )\n   dst32[i] = htons((u16)src32[i])\n\n .. but after the patch it will operate on bytes, which makes this\n useless to convert e.g. concatenations, which store each compound\n in its own register.\n\nMulti-convert of u32 has one theoretical application:\n\nct mark . meta mark . tcp dport @intervalset\n\nBecause ct mark and meta mark are host byte order, use with\nintervals has to convert the byteorder for ct/meta mark value\nto network byte order (bigendian).\n\nnftables emits this:\n [ meta load mark => reg 1 ]\n [ byteorder reg 1 = hton(reg 1, 4, 4) ]\n [ ct load mark => reg 9 ]\n [ byteorder reg 9 = hton(reg 9, 4, 4) ]\n ...\n\nI.e. two separate calls.  Theoretically it could be changed to do:\n [ meta load mark => reg 1 ]\n [ ct load mark => reg 9 ]\n [ byteorder reg 1 = htonl(reg 1, 4, 8) ]\n ...\n\nBut then all it would take to change the set to\nmeta mark . tcp dport . ct mark\n\n... and we'd be back to two \"byteorder\" calls. IOW, support to\nconvert a range of registers is both dysfunctional and dubious.\n\nSimplify this: remove the feature.\n\nPablo Neira Ayuso points out that nftables before 1.1.0 can generate\nincorrect byteorder conversions, see 9fe58952c45a,\n\"evaluate: skip byteorder conversion for selector smaller than 2 bytes\"\nin nftables.git).  Affected rulesets fail to load with this change and\nold userspace due to 'len != size' check.","published":"2026-11-10T18:17:49.147000+00:00","modified":"2026-11-10T18:17:49.147000+00:00","status":"Received","score":null,"severity":"UNSCORED","vector":"","cvss_version":"","affected":[],"weaknesses":[],"references":[{"url":"https://git.kernel.org/stable/c/3c5fb3a6c8118d15e60295e76fb5b879021eef43","source":"416baaa9-dc9f-4396-8d5f-8c081fb06d67"},{"url":"https://git.kernel.org/stable/c/bb061d3de41707415269be75ebf700efb03ec212","source":"416baaa9-dc9f-4396-8d5f-8c081fb06d67"},{"url":"https://git.kernel.org/stable/c/dc29c0478b8564feb9371e715b5a0016eb8903fd","source":"416baaa9-dc9f-4396-8d5f-8c081fb06d67"}],"kev":null,"kev_checked":true,"metrics":{},"configurations":[],"source":"nvd","retrieved_at":"2026-11-10T20:37:37.323251+00:00","exploited":false,"kev_stale":false,"products":[],"priority":{"label":"Assess exposure and vendor guidance","reasons":["No higher-priority signal is available in the current stored evidence."]},"epss":{"probability":0.00167,"percentile":0.05988,"date":"2026-11-18","stale":false},"version_assessment":null},"news":[],"campaigns":[],"domains":[{"domain":"git.kernel.org","source_url":"https://git.kernel.org/stable/c/3c5fb3a6c8118d15e60295e76fb5b879021eef43","investigate_url":"https://investigate.aeglixmind.com/#git.kernel.org"}],"arsenal_url":"https://arsenal.aeglixmind.com/?cve=CVE-2026-99285","catalog_url":"https://cves.aeglixmind.com/cve/CVE-2026-99285","research_url":"https://investigate.aeglixmind.com/research/cve/CVE-2026-99285","coverage":{"news_candidates_checked":0,"campaign_candidates_checked":0,"bounded":false,"notice":"Exact CVE mentions in stored public excerpts and reviewed campaign content only. A mention is not proof of exploitation, attribution or tool suitability. Missing matches do not establish absence."}}