Re: [PATCH iwl-net] ice: fix bound parser hash offset before reading packet data
From: Simon Horman <horms@kernel.org>
Date: 2026-08-24 11:08:47
Also in:
intel-wired-lan
On Thu, Aug 20, 2026 at 01:43:14PM +0200, Aleksandr Loktionov wrote:
quoted hunk ↗ jump to hunk
ice_rt_ho_set() uses the HO register as the starting offset of an ICE_GPR_HV_SIZE-byte memcpy() out of rt->pkt_buf. Potentially HO can be advanced by user-controlled data reachable through ice_parse_raw_rss_pattern() -> ice_parser_run() -> ice_parser_rt_execute() -> ice_rt_gpr_set() -> ice_rt_ho_set(), i.e. a VF-supplied raw RSS pattern (virt/rss.c), with no bound against the size of pkt_buf. Clamp HO to the last offset from which ICE_GPR_HV_SIZE bytes can still be read out of pkt_buf, deriving the limit from sizeof(rt->pkt_buf) so it stays correct if the packet buffer layout changes. Fixes: 9a4c07aaa0f5 ("ice: add parser execution main loop") Cc: stable@vger.kernel.org Signed-off-by: Aleksandr Loktionov <redacted> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> --- drivers/net/ethernet/intel/ice/ice_parser_rt.c | 2 ++ 1 file changed, 2 insertions(+)diff --git a/drivers/net/ethernet/intel/ice/ice_parser_rt.c b/drivers/net/ethernet/intel/ice/ice_parser_rt.c index 3995d66..bfdb50b 100644 --- a/drivers/net/ethernet/intel/ice/ice_parser_rt.c +++ b/drivers/net/ethernet/intel/ice/ice_parser_rt.c@@ -10,6 +10,8 @@ static void ice_rt_tsr_set(struct ice_parser_rt *rt, u16 tsr) static void ice_rt_ho_set(struct ice_parser_rt *rt, u16 ho) { + /* keep the ICE_GPR_HV_SIZE-byte read below within pkt_buf */ + ho = min_t(u16, ho, sizeof(rt->pkt_buf) - ICE_GPR_HV_SIZE); rt->gpr[ICE_GPR_HO_IDX] = ho; memcpy(&rt->gpr[ICE_GPR_HV_IDX], &rt->pkt_buf[ho], ICE_GPR_HV_SIZE); }
Hi, There is an AI-generated review of this patch available at https://sashiko.dev/#/patchset/20260820114314.646344-1-aleksandr.loktionov%40intel.com The first item flagged there seems of concern and I am wondering if you could respond to it. Does this new clamp introduce an infinite loop? In ice_parser_rt_execute(), the loop termination condition is: if (rt->gpr[ICE_GPR_HO_IDX] >= rt->pkt_len) { break; } If a guest VM VF submits a raw RSS pattern with a pkt_len greater than 504 bytes, ice_parser_rt_pktbuf_set() saves that unclamped length directly into rt->pkt_len (e.g., 1024). With this patch, the HO register is capped at 488 (sizeof(rt->pkt_buf) - ICE_GPR_HV_SIZE). Since 488 will never be greater than or equal to 1024, can this cause the parser to loop indefinitely if the state machine processes a repetitive sequence?