mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Jialin Wang <wjl.linux@gmail.com>
To: Thomas Gleixner <tglx@kernel.org>, Ingo Molnar <mingo@redhat.com>,
	Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>
Cc: linux-kernel@vger.kernel.org, Jialin Wang <wjl.linux@gmail.com>
Subject: [PATCH] x86/reboot: Disable CET in machine_real_restart()
Date: Sun, 11 Oct 2026 23:18:03 +0800	[thread overview]
Message-ID: <20261011151807.263575-1-wjl.linux@gmail.com> (raw)

When a kernel panic occurs in an environment without ACPI reboot or
8042 keyboard reset support (such as QEMU microvm),
native_machine_emergency_restart() falls back through BOOT_ACPI,
BOOT_KBD, BOOT_EFI, and eventually invokes machine_real_restart(MRR_BIOS)
to switch back to real mode for a BIOS reboot.

When kernel IBT is enabled, the indirect far jump (ljmpl) in
machine_real_restart() jumps to the 32-bit compatibility mode entry point
(pa_machine_real_restart_asm). Because machine_real_restart_asm does not
begin with an ENDBR32 instruction, an indirect branch tracking #CP fault
is triggered. Since ibt_fatal is true, do_kernel_cp_fault() triggers:

  Missing ENDBR: 0x9a230
  kernel BUG at arch/x86/kernel/cet.c:133!

This issue can be reproduced with virtme-ng:

  vng -v -r -- 'echo c >/proc/sysrq-trigger'

which launches QEMU with the microvm architecture:

  qemu-system-x86_64 -M microvm,accel=kvm,pcie=on,rtc=on \
    -cpu host,topoext=on -smp 16 -no-reboot \
    -kernel <vmlinuz> -append '... panic=-1 ...'

Invoke cet_disable() in machine_real_restart() before loading the
trampoline page table and executing the far jump, ensuring IBT is disabled
and avoiding the #CP fault.

Assisted-by: LLM
Signed-off-by: Jialin Wang <wjl.linux@gmail.com>
---
 arch/x86/kernel/reboot.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 0fed6d0d7e32..eb23b8977de5 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -114,6 +114,13 @@ void __noreturn machine_real_restart(unsigned int type)
 	CMOS_WRITE(0x00, 0x8f);
 	spin_unlock(&rtc_lock);
 
+	/*
+	 * Disable CET before loading the trampoline page table and jumping
+	 * to the real-mode trampoline. This avoids missing-ENDBR #CP faults
+	 * during the far jump.
+	 */
+	cet_disable();
+
 	/*
 	 * Switch to the trampoline page table.
 	 */
-- 
2.56.0


             reply	other threads:[~2026-10-11 15:18 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-11 15:18 Jialin Wang [this message]
2026-10-11 15:56 ` H. Peter Anvin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20261011151807.263575-1-wjl.linux@gmail.com \
    --to=wjl.linux@gmail.com \
    --cc=bp@alien8.de \
    --cc=dave.hansen@linux.intel.com \
    --cc=hpa@zytor.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=tglx@kernel.org \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®