arch/x86/kernel/alternative.c | 72 +++++++++++++++++++++++-------------------- 1 file changed, 38 insertions(+), 34 deletions(-)
The switch of the FineIBT preamble from "subl $hash, %r10d" to the
shorter "subl $hash, %eax" moved the hash immediate from offset 7 to
offset 5 of the preamble. fineibt_preamble_hash was updated to match,
but the open-coded offset in cfi_get_func_hash() was missed and it
still reads the hash at offset 7.
cfi_get_func_hash() is used by the BPF JIT to give a struct_ops
trampoline the CFI hash of the stub function it stands in for. With
FineIBT the trampoline now gets the upper half of the real hash
followed by the first two bytes of the next instruction, so the first
indirect call from the kernel into a struct_ops program,
tcp_init_congestion_control() calling ->init() of a BPF congestion
control for example, fails the FineIBT check and the kernel dies with
a CFI failure.
Move the FineIBT preamble template and its offset defines above
cfi_get_func_hash() and use fineibt_preamble_hash there, so every
reader of the preamble shares one definition of its layout. The
CFI_FINEIBT arm is only built with CONFIG_FINEIBT, the only
configuration in which cfi_mode can take that value.
cfi_get_func_arity() does not need the same treatment: the __bhi_args
call whose displacement it reads still ends at the function address.
Fixes: 85a2d4a890dc ("x86,ibt: Use UDB instead of 0xEA")
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: stable@vger.kernel.org # 6.18+
Assisted-by: LLM
Signed-off-by: Soheil Hassas Yeganeh <soheil.kdev@gmail.com>
---
arch/x86/kernel/alternative.c | 72 +++++++++++++++++++++++--------------------
1 file changed, 38 insertions(+), 34 deletions(-)
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 91b1cdd16569..1b22062cabfe 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -1198,6 +1198,41 @@ static bool cfi_debug __ro_after_init;
bool cfi_bhi __ro_after_init = false;
#endif
+#ifdef CONFIG_FINEIBT
+/*
+ * <fineibt_preamble_start>:
+ * 0: f3 0f 1e fa endbr64
+ * 4: 2d 78 56 34 12 sub $0x12345678, %eax
+ * 9: 2e 0f 85 03 00 00 00 jne,pn 13 <fineibt_preamble_start+0x13>
+ * 10: 0f 1f 40 d6 nopl -0x2a(%rax)
+ *
+ * Note that the JNE target is the 0xD6 byte inside the NOPL, this decodes as
+ * UDB on x86_64 and raises #UD.
+ */
+asm( ".pushsection .rodata \n"
+ "fineibt_preamble_start: \n"
+ " endbr64 \n"
+ " subl $0x12345678, %eax \n"
+ "fineibt_preamble_bhi: \n"
+ " cs jne.d32 fineibt_preamble_start+0x13 \n"
+ "#fineibt_func: \n"
+ " nopl -42(%rax) \n"
+ "fineibt_preamble_end: \n"
+ ".popsection\n"
+);
+
+extern u8 fineibt_preamble_start[];
+extern u8 fineibt_preamble_bhi[];
+extern u8 fineibt_preamble_end[];
+
+#define fineibt_preamble_size (fineibt_preamble_end - fineibt_preamble_start)
+#define fineibt_preamble_bhi (fineibt_preamble_bhi - fineibt_preamble_start)
+#define fineibt_preamble_ud 0x13
+#define fineibt_preamble_hash 5
+
+#define fineibt_prefix_size (fineibt_preamble_size - ENDBR_INSN_SIZE)
+#endif /* CONFIG_FINEIBT */
+
#ifdef CONFIG_CFI
u32 cfi_get_func_hash(void *func)
{
@@ -1205,9 +1240,11 @@ u32 cfi_get_func_hash(void *func)
func -= cfi_get_offset();
switch (cfi_mode) {
+#ifdef CONFIG_FINEIBT
case CFI_FINEIBT:
- func += 7;
+ func += fineibt_preamble_hash;
break;
+#endif
case CFI_KCFI:
func += 1;
break;
@@ -1363,39 +1400,6 @@ early_param("cfi", cfi_parse_cmdline);
* anyway.
*/
-/*
- * <fineibt_preamble_start>:
- * 0: f3 0f 1e fa endbr64
- * 4: 2d 78 56 34 12 sub $0x12345678, %eax
- * 9: 2e 0f 85 03 00 00 00 jne,pn 13 <fineibt_preamble_start+0x13>
- * 10: 0f 1f 40 d6 nopl -0x2a(%rax)
- *
- * Note that the JNE target is the 0xD6 byte inside the NOPL, this decodes as
- * UDB on x86_64 and raises #UD.
- */
-asm( ".pushsection .rodata \n"
- "fineibt_preamble_start: \n"
- " endbr64 \n"
- " subl $0x12345678, %eax \n"
- "fineibt_preamble_bhi: \n"
- " cs jne.d32 fineibt_preamble_start+0x13 \n"
- "#fineibt_func: \n"
- " nopl -42(%rax) \n"
- "fineibt_preamble_end: \n"
- ".popsection\n"
-);
-
-extern u8 fineibt_preamble_start[];
-extern u8 fineibt_preamble_bhi[];
-extern u8 fineibt_preamble_end[];
-
-#define fineibt_preamble_size (fineibt_preamble_end - fineibt_preamble_start)
-#define fineibt_preamble_bhi (fineibt_preamble_bhi - fineibt_preamble_start)
-#define fineibt_preamble_ud 0x13
-#define fineibt_preamble_hash 5
-
-#define fineibt_prefix_size (fineibt_preamble_size - ENDBR_INSN_SIZE)
-
/*
* <fineibt_caller_start>:
* 0: b8 78 56 34 12 mov $0x12345678, %eax
---
base-commit: 08dbfad3f5040f5bdb6c529da20d6d4e81fefd72
change-id: 20260831-b4-x86-cfi-fineibt-func-hash-e160f6eee02c
Best regards,
--
Soheil Hassas Yeganeh <soheil.kdev@gmail.com>On Mon, Aug 31, 2026 at 10:49 AM Soheil Hassas Yeganeh
<soheil.kdev@gmail.com> wrote:
>
> The switch of the FineIBT preamble from "subl $hash, %r10d" to the
> shorter "subl $hash, %eax" moved the hash immediate from offset 7 to
> offset 5 of the preamble. fineibt_preamble_hash was updated to match,
> but the open-coded offset in cfi_get_func_hash() was missed and it
> still reads the hash at offset 7.
>
> cfi_get_func_hash() is used by the BPF JIT to give a struct_ops
> trampoline the CFI hash of the stub function it stands in for. With
> FineIBT the trampoline now gets the upper half of the real hash
> followed by the first two bytes of the next instruction, so the first
> indirect call from the kernel into a struct_ops program,
> tcp_init_congestion_control() calling ->init() of a BPF congestion
> control for example, fails the FineIBT check and the kernel dies with
> a CFI failure.
>
> Move the FineIBT preamble template and its offset defines above
> cfi_get_func_hash() and use fineibt_preamble_hash there, so every
> reader of the preamble shares one definition of its layout. The
> CFI_FINEIBT arm is only built with CONFIG_FINEIBT, the only
> configuration in which cfi_mode can take that value.
> cfi_get_func_arity() does not need the same treatment: the __bhi_args
> call whose displacement it reads still ends at the function address.
>
> Fixes: 85a2d4a890dc ("x86,ibt: Use UDB instead of 0xEA")
Gentle ping on this one. To recap: since 85a2d4a890dc the FineIBT preamble
stores the hash at +5, but cfi_get_func_hash() still reads it at +7, so BPF
struct_ops trampolines are emitted with a wrong hash and the first indirect call
through them (e.g. a bpf tcp_congestion_ops .cong_control) hits a FineIBT #CP.
We have large scale panics when we use BPF congestion control.
Happy to respin if a different shape is preferred (e.g. moving the
offset next to
the preamble definition), but this seemed like the best way to fix the issue.
Thanks,
Soheil
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Alexei Starovoitov <ast@kernel.org>
> Cc: Daniel Borkmann <daniel@iogearbox.net>
> Cc: stable@vger.kernel.org # 6.18+
> Assisted-by: LLM
> Signed-off-by: Soheil Hassas Yeganeh <soheil.kdev@gmail.com>
> ---
> arch/x86/kernel/alternative.c | 72 +++++++++++++++++++++++--------------------
> 1 file changed, 38 insertions(+), 34 deletions(-)
>
> diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
> index 91b1cdd16569..1b22062cabfe 100644
> --- a/arch/x86/kernel/alternative.c
> +++ b/arch/x86/kernel/alternative.c
> @@ -1198,6 +1198,41 @@ static bool cfi_debug __ro_after_init;
> bool cfi_bhi __ro_after_init = false;
> #endif
>
> +#ifdef CONFIG_FINEIBT
> +/*
> + * <fineibt_preamble_start>:
> + * 0: f3 0f 1e fa endbr64
> + * 4: 2d 78 56 34 12 sub $0x12345678, %eax
> + * 9: 2e 0f 85 03 00 00 00 jne,pn 13 <fineibt_preamble_start+0x13>
> + * 10: 0f 1f 40 d6 nopl -0x2a(%rax)
> + *
> + * Note that the JNE target is the 0xD6 byte inside the NOPL, this decodes as
> + * UDB on x86_64 and raises #UD.
> + */
> +asm( ".pushsection .rodata \n"
> + "fineibt_preamble_start: \n"
> + " endbr64 \n"
> + " subl $0x12345678, %eax \n"
> + "fineibt_preamble_bhi: \n"
> + " cs jne.d32 fineibt_preamble_start+0x13 \n"
> + "#fineibt_func: \n"
> + " nopl -42(%rax) \n"
> + "fineibt_preamble_end: \n"
> + ".popsection\n"
> +);
> +
> +extern u8 fineibt_preamble_start[];
> +extern u8 fineibt_preamble_bhi[];
> +extern u8 fineibt_preamble_end[];
> +
> +#define fineibt_preamble_size (fineibt_preamble_end - fineibt_preamble_start)
> +#define fineibt_preamble_bhi (fineibt_preamble_bhi - fineibt_preamble_start)
> +#define fineibt_preamble_ud 0x13
> +#define fineibt_preamble_hash 5
> +
> +#define fineibt_prefix_size (fineibt_preamble_size - ENDBR_INSN_SIZE)
> +#endif /* CONFIG_FINEIBT */
> +
> #ifdef CONFIG_CFI
> u32 cfi_get_func_hash(void *func)
> {
> @@ -1205,9 +1240,11 @@ u32 cfi_get_func_hash(void *func)
>
> func -= cfi_get_offset();
> switch (cfi_mode) {
> +#ifdef CONFIG_FINEIBT
> case CFI_FINEIBT:
> - func += 7;
> + func += fineibt_preamble_hash;
> break;
> +#endif
> case CFI_KCFI:
> func += 1;
> break;
> @@ -1363,39 +1400,6 @@ early_param("cfi", cfi_parse_cmdline);
> * anyway.
> */
>
> -/*
> - * <fineibt_preamble_start>:
> - * 0: f3 0f 1e fa endbr64
> - * 4: 2d 78 56 34 12 sub $0x12345678, %eax
> - * 9: 2e 0f 85 03 00 00 00 jne,pn 13 <fineibt_preamble_start+0x13>
> - * 10: 0f 1f 40 d6 nopl -0x2a(%rax)
> - *
> - * Note that the JNE target is the 0xD6 byte inside the NOPL, this decodes as
> - * UDB on x86_64 and raises #UD.
> - */
> -asm( ".pushsection .rodata \n"
> - "fineibt_preamble_start: \n"
> - " endbr64 \n"
> - " subl $0x12345678, %eax \n"
> - "fineibt_preamble_bhi: \n"
> - " cs jne.d32 fineibt_preamble_start+0x13 \n"
> - "#fineibt_func: \n"
> - " nopl -42(%rax) \n"
> - "fineibt_preamble_end: \n"
> - ".popsection\n"
> -);
> -
> -extern u8 fineibt_preamble_start[];
> -extern u8 fineibt_preamble_bhi[];
> -extern u8 fineibt_preamble_end[];
> -
> -#define fineibt_preamble_size (fineibt_preamble_end - fineibt_preamble_start)
> -#define fineibt_preamble_bhi (fineibt_preamble_bhi - fineibt_preamble_start)
> -#define fineibt_preamble_ud 0x13
> -#define fineibt_preamble_hash 5
> -
> -#define fineibt_prefix_size (fineibt_preamble_size - ENDBR_INSN_SIZE)
> -
> /*
> * <fineibt_caller_start>:
> * 0: b8 78 56 34 12 mov $0x12345678, %eax
>
> ---
> base-commit: 08dbfad3f5040f5bdb6c529da20d6d4e81fefd72
> change-id: 20260831-b4-x86-cfi-fineibt-func-hash-e160f6eee02c
>
> Best regards,
> --
> Soheil Hassas Yeganeh <soheil.kdev@gmail.com>
>
On Tue, Sep 8, 2026 at 6:22 AM Soheil Hassas Yeganeh
<soheil.kdev@gmail.com> wrote:
>
> On Mon, Aug 31, 2026 at 10:49 AM Soheil Hassas Yeganeh
> <soheil.kdev@gmail.com> wrote:
> >
> > The switch of the FineIBT preamble from "subl $hash, %r10d" to the
> > shorter "subl $hash, %eax" moved the hash immediate from offset 7 to
> > offset 5 of the preamble. fineibt_preamble_hash was updated to match,
> > but the open-coded offset in cfi_get_func_hash() was missed and it
> > still reads the hash at offset 7.
> >
> > cfi_get_func_hash() is used by the BPF JIT to give a struct_ops
> > trampoline the CFI hash of the stub function it stands in for. With
> > FineIBT the trampoline now gets the upper half of the real hash
> > followed by the first two bytes of the next instruction, so the first
> > indirect call from the kernel into a struct_ops program,
> > tcp_init_congestion_control() calling ->init() of a BPF congestion
> > control for example, fails the FineIBT check and the kernel dies with
> > a CFI failure.
> >
> > Move the FineIBT preamble template and its offset defines above
> > cfi_get_func_hash() and use fineibt_preamble_hash there, so every
> > reader of the preamble shares one definition of its layout. The
> > CFI_FINEIBT arm is only built with CONFIG_FINEIBT, the only
> > configuration in which cfi_mode can take that value.
> > cfi_get_func_arity() does not need the same treatment: the __bhi_args
> > call whose displacement it reads still ends at the function address.
> >
> > Fixes: 85a2d4a890dc ("x86,ibt: Use UDB instead of 0xEA")
>
> Gentle ping on this one. To recap: since 85a2d4a890dc the FineIBT preamble
> stores the hash at +5, but cfi_get_func_hash() still reads it at +7, so BPF
> struct_ops trampolines are emitted with a wrong hash and the first indirect call
> through them (e.g. a bpf tcp_congestion_ops .cong_control) hits a FineIBT #CP.
>
> We have large scale panics when we use BPF congestion control.
> Happy to respin if a different shape is preferred (e.g. moving the
> offset next to
> the preamble definition), but this seemed like the best way to fix the issue.
Looks fine to me. I guess it will go via tip tree?
On Tue, Sep 08, 2026 at 09:37:46AM -0700, Alexei Starovoitov wrote:
> On Tue, Sep 8, 2026 at 6:22 AM Soheil Hassas Yeganeh
> <soheil.kdev@gmail.com> wrote:
> >
> > On Mon, Aug 31, 2026 at 10:49 AM Soheil Hassas Yeganeh
> > <soheil.kdev@gmail.com> wrote:
> > >
> > > The switch of the FineIBT preamble from "subl $hash, %r10d" to the
> > > shorter "subl $hash, %eax" moved the hash immediate from offset 7 to
> > > offset 5 of the preamble. fineibt_preamble_hash was updated to match,
> > > but the open-coded offset in cfi_get_func_hash() was missed and it
> > > still reads the hash at offset 7.
> > >
> > > cfi_get_func_hash() is used by the BPF JIT to give a struct_ops
> > > trampoline the CFI hash of the stub function it stands in for. With
> > > FineIBT the trampoline now gets the upper half of the real hash
> > > followed by the first two bytes of the next instruction, so the first
> > > indirect call from the kernel into a struct_ops program,
> > > tcp_init_congestion_control() calling ->init() of a BPF congestion
> > > control for example, fails the FineIBT check and the kernel dies with
> > > a CFI failure.
> > >
> > > Move the FineIBT preamble template and its offset defines above
> > > cfi_get_func_hash() and use fineibt_preamble_hash there, so every
> > > reader of the preamble shares one definition of its layout. The
> > > CFI_FINEIBT arm is only built with CONFIG_FINEIBT, the only
> > > configuration in which cfi_mode can take that value.
> > > cfi_get_func_arity() does not need the same treatment: the __bhi_args
> > > call whose displacement it reads still ends at the function address.
> > >
> > > Fixes: 85a2d4a890dc ("x86,ibt: Use UDB instead of 0xEA")
> >
> > Gentle ping on this one. To recap: since 85a2d4a890dc the FineIBT preamble
> > stores the hash at +5, but cfi_get_func_hash() still reads it at +7, so BPF
> > struct_ops trampolines are emitted with a wrong hash and the first indirect call
> > through them (e.g. a bpf tcp_congestion_ops .cong_control) hits a FineIBT #CP.
> >
> > We have large scale panics when we use BPF congestion control.
> > Happy to respin if a different shape is preferred (e.g. moving the
> > offset next to
> > the preamble definition), but this seemed like the best way to fix the issue.
>
> Looks fine to me. I guess it will go via tip tree?
Yes, I'll pick it up. Thanks!
© 2016 - 2026 Red Hat, Inc.