CVE-2026-53134

*** Pendiente de traducción *** In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> netfilter: nft_fib: fix stale stack leak via the OIFNAME register<br /> <br /> For NFT_FIB_RESULT_OIFNAME the destination register is declared with<br /> len = IFNAMSIZ (four 32-bit registers), but on the lookup-fail,<br /> RTN_LOCAL and oif-mismatch paths nft_fib{4,6}_eval() only writes one<br /> register via "*dest = 0". The remaining three registers are left as<br /> whatever was on the stack in nft_do_chain()&amp;#39;s struct nft_regs, and a<br /> downstream expression that loads the register span can leak that<br /> uninitialised kernel stack to userspace.<br /> <br /> The NFTA_FIB_F_PRESENT existence check has the same shape: it is only<br /> meaningful for NFT_FIB_RESULT_OIF, yet it was accepted for any result type<br /> while the eval stores a single byte via nft_reg_store8(), leaving the rest<br /> of the declared span stale.<br /> <br /> Fix both:<br /> <br /> - replace the bare "*dest = 0" in the eval with nft_fib_store_result(),<br /> which strscpy_pad()s the whole IFNAMSIZ for OIFNAME (and is already<br /> used on the other early-return path), and<br /> <br /> - restrict NFTA_FIB_F_PRESENT to NFT_FIB_RESULT_OIF and declare its<br /> destination as a single u8, so the marked span matches the one byte<br /> the eval writes.