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 --- Documentation/admin-guide/vga-softcursor.rst | 62 ++++++++++++++++++++++++++++ 1 file changed, 62 insertions(+) create mode 100644 Documentation/admin-guide/vga-softcursor.rst (limited to 'Documentation/admin-guide/vga-softcursor.rst') diff --git a/Documentation/admin-guide/vga-softcursor.rst b/Documentation/admin-guide/vga-softcursor.rst new file mode 100644 index 000000000..f52175457 --- /dev/null +++ b/Documentation/admin-guide/vga-softcursor.rst @@ -0,0 +1,62 @@ +Software cursor for VGA +======================= + +by Pavel Machek +and Martin Mares + +Linux now has some ability to manipulate cursor appearance. Normally, +you can set the size of hardware cursor. You can now play a few new +tricks: you can make your cursor look like a non-blinking red block, +make it inverse background of the character it's over or to highlight +that character and still choose whether the original hardware cursor +should remain visible or not. There may be other things I have never +thought of. + +The cursor appearance is controlled by a ``[?1;2;3c`` escape sequence +where 1, 2 and 3 are parameters described below. If you omit any of them, +they will default to zeroes. + +first Parameter + specifies cursor size:: + + 0=default + 1=invisible + 2=underline, + ... + 8=full block + + 16 if you want the software cursor to be applied + + 32 if you want to always change the background color + + 64 if you dislike having the background the same as the + foreground. + + Highlights are ignored for the last two flags. + +second parameter + selects character attribute bits you want to change + (by simply XORing them with the value of this parameter). On standard + VGA, the high four bits specify background and the low four the + foreground. In both groups, low three bits set color (as in normal + color codes used by the console) and the most significant one turns + on highlight (or sometimes blinking -- it depends on the configuration + of your VGA). + +third parameter + consists of character attribute bits you want to set. + + Bit setting takes place before bit toggling, so you can simply clear a + bit by including it in both the set mask and the toggle mask. + +Examples +-------- + +To get normal blinking underline, use:: + + echo -e '\033[?2c' + +To get blinking block, use:: + + echo -e '\033[?6c' + +To get red non-blinking block, use:: + + echo -e '\033[?17;0;64c' -- cgit v1.3.1