From a5f7a5bb3b7f28ba7e4fa246775b29a0e5537255 Mon Sep 17 00:00:00 2001 From: Jinke Han Date: Tue, 8 Sep 2026 15:37:42 +0800 Subject: x86/kprobes: Fix crash when probing CS CALL instructions When using eBPF to probe CS CALL instructions within a function, a crash can be triggered. The eBPF tool probes offset 257 of the __hrtimer_run_queues() function: <__hrtimer_run_queues+249>: nopl 0x0(%rax,%rax,1) <__hrtimer_run_queues+254>: mov %r14,%rdi <__hrtimer_run_queues+257>: cs call <__x86_indirect_thunk_r12> <__hrtimer_run_queues+263>: mov %eax,%r12d <__hrtimer_run_queues+266>: xchg %ax,%ax <__hrtimer_run_queues+268>: mov %r13,%rdi Which triggers this crash: BUG: unable to handle page fault for address: 00000000000f41c9 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 0 P4D 0 Oops: 0002 [#1] SMP NOPTI CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P RIP: 0010:__hrtimer_run_queues+0x106/0x230 Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is at the 6th byte of the above CS CALL instruction. Since the CS CALL instruction occupies 6 bytes, the exception occurred in the middle of that call instruction. The root cause is that when using eBPF tools to probe in the middle of a function, a kprobe with INT3 is used as the underlying implementation. During single-step emulation of the original CALL instruction, int3_emulate_call() assumes that the probed CALL instruction is 5 bytes long. However, the actual CS-prefixed CALL instruction occupies 6 bytes, so it constructs an incorrect exception return address. When the CPU returns from the kprobe handler, the next instruction to be executed is at the address of the last byte of that CS CALL instruction. Coincidentally, starting from that address, the CPU fetches and decodes a completely different instruction, which ultimately triggers a kernel crash. Fix the issue by using the actual instruction length obtained from the instruction decoder when constructing the exception return address, rather than relying on the hardcoded CALL_INSN_SIZE macro. [ mingo: Refined the changelog ] Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step") Suggested-by: Masami Hiramatsu (Google) Signed-off-by: Jinke Han Signed-off-by: Ingo Molnar Reviewed-by: Masami Hiramatsu (Google) Acked-by: Yafang Shao Acked-by: Borislav Petkov Cc: Peter Zijlstra Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000 --- drivers/net/phy/linkmode.c | 95 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 95 insertions(+) create mode 100644 drivers/net/phy/linkmode.c (limited to 'drivers/net/phy/linkmode.c') diff --git a/drivers/net/phy/linkmode.c b/drivers/net/phy/linkmode.c new file mode 100644 index 000000000..f60560fe3 --- /dev/null +++ b/drivers/net/phy/linkmode.c @@ -0,0 +1,95 @@ +// SPDX-License-Identifier: GPL-2.0+ +#include + +/** + * linkmode_resolve_pause - resolve the allowable pause modes + * @local_adv: local advertisement in ethtool format + * @partner_adv: partner advertisement in ethtool format + * @tx_pause: pointer to bool to indicate whether transmit pause should be + * enabled. + * @rx_pause: pointer to bool to indicate whether receive pause should be + * enabled. + * + * Flow control is resolved according to our and the link partners + * advertisements using the following drawn from the 802.3 specs: + * Local device Link partner + * Pause AsymDir Pause AsymDir Result + * 0 X 0 X Disabled + * 0 1 1 0 Disabled + * 0 1 1 1 TX + * 1 0 0 X Disabled + * 1 X 1 X TX+RX + * 1 1 0 1 RX + */ +void linkmode_resolve_pause(const unsigned long *local_adv, + const unsigned long *partner_adv, + bool *tx_pause, bool *rx_pause) +{ + __ETHTOOL_DECLARE_LINK_MODE_MASK(m); + + linkmode_and(m, local_adv, partner_adv); + if (linkmode_test_bit(ETHTOOL_LINK_MODE_Pause_BIT, m)) { + *tx_pause = true; + *rx_pause = true; + } else if (linkmode_test_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, m)) { + *tx_pause = linkmode_test_bit(ETHTOOL_LINK_MODE_Pause_BIT, + partner_adv); + *rx_pause = linkmode_test_bit(ETHTOOL_LINK_MODE_Pause_BIT, + local_adv); + } else { + *tx_pause = false; + *rx_pause = false; + } +} +EXPORT_SYMBOL_GPL(linkmode_resolve_pause); + +/** + * linkmode_set_pause - set the pause mode advertisement + * @advertisement: advertisement in ethtool format + * @tx: boolean from ethtool struct ethtool_pauseparam tx_pause member + * @rx: boolean from ethtool struct ethtool_pauseparam rx_pause member + * + * Configure the advertised Pause and Asym_Pause bits according to the + * capabilities of provided in @tx and @rx. + * + * We convert as follows: + * tx rx Pause AsymDir + * 0 0 0 0 + * 0 1 1 1 + * 1 0 0 1 + * 1 1 1 0 + * + * Note: this translation from ethtool tx/rx notation to the advertisement + * is actually very problematical. Here are some examples: + * + * For tx=0 rx=1, meaning transmit is unsupported, receive is supported: + * + * Local device Link partner + * Pause AsymDir Pause AsymDir Result + * 1 1 1 0 TX + RX - but we have no TX support. + * 1 1 0 1 Only this gives RX only + * + * For tx=1 rx=1, meaning we have the capability to transmit and receive + * pause frames: + * + * Local device Link partner + * Pause AsymDir Pause AsymDir Result + * 1 0 0 1 Disabled - but since we do support tx and rx, + * this should resolve to RX only. + * + * Hence, asking for: + * rx=1 tx=0 gives Pause+AsymDir advertisement, but we may end up + * resolving to tx+rx pause or only rx pause depending on + * the partners advertisement. + * rx=0 tx=1 gives AsymDir only, which will only give tx pause if + * the partners advertisement allows it. + * rx=1 tx=1 gives Pause only, which will only allow tx+rx pause + * if the other end also advertises Pause. + */ +void linkmode_set_pause(unsigned long *advertisement, bool tx, bool rx) +{ + linkmode_mod_bit(ETHTOOL_LINK_MODE_Pause_BIT, advertisement, rx); + linkmode_mod_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, advertisement, + rx ^ tx); +} +EXPORT_SYMBOL_GPL(linkmode_set_pause); -- cgit v1.3.1