CVE-2026-93270

*** Pendiente de traducción *** In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> bpf: Disallow interpreter fallback for BPF_ADDR_PERCPU insn<br /> <br /> The BPF_MOV64_PERCPU_REG insn requires JIT to emit native code to for<br /> &amp;#39;dst_reg = src_reg + &amp;#39;.<br /> <br /> However, the interpreter ignores the &amp;#39;off&amp;#39; at its ALU64_MOV_X label.<br /> The &amp;#39;off&amp;#39; indicates the insn is BPF_MOV64_PERCPU_REG insn. Then, when<br /> the interpreter loads memory from the register, it will hit a page<br /> fault.<br /> <br /> [ 2.545572] BUG: unable to handle page fault for address: ffffffffacaaf034<br /> [ 2.546485] #PF: supervisor read access in kernel mode<br /> [ 2.547167] #PF: error_code(0x0000) - not-present page<br /> [ 2.547850] PGD 134e63067 P4D 134e63067 PUD 134e64063 PMD 10021c063 PTE 800ffffeca550062<br /> [ 2.548912] Oops: Oops: 0000 [#1] SMP PTI<br /> <br /> Set jit_required as true in order to disallow interpreter fallback in<br /> core.c::__bpf_prog_select_runtime(), if any BPF_ADDR_PERCPU insn is<br /> patched to the prog.<br /> <br /> BTW, rename the helper bpf_map_supports_cpu_flags() to<br /> bpf_map_is_percpu_map().