[PATCH v3 0/2] Work around missing `bti c` in modules

[PATCH v3 0/2] Work around missing `bti c` in modules

[D Scott Phillips]

Changes since v2:
 - Split the kernel BTI exception behavior change into a separate patch

Changes since v1:
 - Add the gcc bug id into the traps.c comment
 - Cover the try_module_get with the preempt_disable
 - Add a CC_HAS_ config for the compiler bug that we'll eventually refine


D Scott Phillips (2):
  arm64: Handle kernel BTI exceptions
  arm64: Work around missing `bti c` in modules

 arch/arm64/Kconfig               |  3 +++
 arch/arm64/kernel/entry-common.c | 12 +++++++++
 arch/arm64/kernel/traps.c        | 43 ++++++++++++++++++++++++++++++--
 3 files changed, 56 insertions(+), 2 deletions(-)

-- 
2.37.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v3 1/2] arm64: Handle kernel BTI exceptions

[D Scott Phillips]

Send BTI exceptions to the do_bti handler, killing the current task with
SIGSEGV instead of panicking. This is to allow a later patch to apply a
compiler bug workaround.

Signed-off-by: D Scott Phillips <scott@os.amperecomputing.com>
---
 arch/arm64/kernel/entry-common.c | 12 ++++++++++++
 arch/arm64/kernel/traps.c        |  8 ++++++--
 2 files changed, 18 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c
index c75ca36b4a49..dad27e854d8c 100644
--- a/arch/arm64/kernel/entry-common.c
+++ b/arch/arm64/kernel/entry-common.c
@@ -388,6 +388,15 @@ static void noinstr el1_undef(struct pt_regs *regs)
 	exit_to_kernel_mode(regs);
 }
 
+static void noinstr el1_bti(struct pt_regs *regs)
+{
+	enter_from_kernel_mode(regs);
+	local_daif_inherit(regs);
+	do_bti(regs);
+	local_daif_mask();
+	exit_to_kernel_mode(regs);
+}
+
 static void noinstr el1_dbg(struct pt_regs *regs, unsigned long esr)
 {
 	unsigned long far = read_sysreg(far_el1);
@@ -427,6 +436,9 @@ asmlinkage void noinstr el1h_64_sync_handler(struct pt_regs *regs)
 	case ESR_ELx_EC_UNKNOWN:
 		el1_undef(regs);
 		break;
+	case ESR_ELx_EC_BTI:
+		el1_bti(regs);
+		break;
 	case ESR_ELx_EC_BREAKPT_CUR:
 	case ESR_ELx_EC_SOFTSTP_CUR:
 	case ESR_ELx_EC_WATCHPT_CUR:
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index b7fed33981f7..56e1782fcf54 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -501,8 +501,12 @@ NOKPROBE_SYMBOL(do_undefinstr);
 
 void do_bti(struct pt_regs *regs)
 {
-	BUG_ON(!user_mode(regs));
-	force_signal_inject(SIGILL, ILL_ILLOPC, regs->pc, 0);
+	if (user_mode(regs)) {
+		force_signal_inject(SIGILL, ILL_ILLOPC, regs->pc, 0);
+		return;
+	}
+
+	die("Oops - BTI", regs, 0);
 }
 NOKPROBE_SYMBOL(do_bti);
 
-- 
2.37.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v3 2/2] arm64: Work around missing `bti c` in modules

[D Scott Phillips]

GCC does not insert a `bti c` instruction at the beginning of a function
when all callers reach the function through a direct branch[1]. In the case
of cross-section calls (like __init to non __init), a thunk may be inserted
which uses an indirect branch. If that happens, the first instruction in
the callee function will result in a Branch Target Exception due to the
missing `bti c`.

Handle Branch Target Exceptions which happen in the kernel due to module
calls from __init to non-__init by clearing PSTATE.BTYPE and resuming.

[1]: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106671

Signed-off-by: D Scott Phillips <scott@os.amperecomputing.com>
---
 arch/arm64/Kconfig        |  3 +++
 arch/arm64/kernel/traps.c | 35 +++++++++++++++++++++++++++++++++++
 2 files changed, 38 insertions(+)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 9fb9fff08c94..8038842fa6b9 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -1896,6 +1896,9 @@ config ARM64_BTI_KERNEL
 	  is enabled and the system supports BTI all kernel code including
 	  modular code must have BTI enabled.
 
+config CC_HAS_CROSS_SECTION_BTI_MISSING
+	def_bool CC_IS_GCC # https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106671
+
 config CC_HAS_BRANCH_PROT_PAC_RET_BTI
 	# GCC 9 or later, clang 8 or later
 	def_bool $(cc-option,-mbranch-protection=pac-ret+leaf+bti)
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index 56e1782fcf54..315a305a4f1d 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -501,11 +501,46 @@ NOKPROBE_SYMBOL(do_undefinstr);
 
 void do_bti(struct pt_regs *regs)
 {
+	struct module *mod;
+
 	if (user_mode(regs)) {
 		force_signal_inject(SIGILL, ILL_ILLOPC, regs->pc, 0);
 		return;
 	}
 
+	/*
+	 * GCC does not insert a `bti c` instruction at the beginning
+	 * of a function when all callers reach the function through a
+	 * direct branch. In the case of cross-section calls (like
+	 * __init to non __init), a thunk may be inserted which uses
+	 * an indirect branch. If that happens, the first instruction
+	 * in the callee function will result in a Branch Target
+	 * Exception due to the missing `bti c`.
+	 *
+	 * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106671
+	 *
+	 * If that's the case here, clear PSTATE.BTYPE and resume.
+	 */
+	if (IS_ENABLED(CONFIG_CC_HAS_CROSS_SECTION_BTI_MISSING)) {
+		preempt_disable();
+		mod = __module_text_address(regs->pc);
+
+		if (mod && try_module_get(mod)) {
+			bool from_init;
+
+			from_init = within_module_init(regs->regs[30], mod);
+			module_put(mod);
+
+			if (from_init) {
+				preempt_enable();
+				regs->pstate &= ~PSR_BTYPE_MASK;
+				return;
+			}
+		}
+
+		preempt_enable();
+	}
+
 	die("Oops - BTI", regs, 0);
 }
 NOKPROBE_SYMBOL(do_bti);
-- 
2.37.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v3 1/2] arm64: Handle kernel BTI exceptions

[Mark Brown]

[-- Attachment #1.1: Type: text/plain, Size: 609 bytes --]

On Thu, Sep 01, 2022 at 05:15:50PM -0700, D Scott Phillips wrote:

> Send BTI exceptions to the do_bti handler, killing the current task with
> SIGSEGV instead of panicking. This is to allow a later patch to apply a
> compiler bug workaround.

There's arguments either way here - this is less destructive, but you
could say that since these exceptions are supposed to indicate a
security issue stopping the whole kernel has some safety advantages,
though it does also open up DoS opportunities.  I don't have a strong
opinion here, I mildly prefer this approach.

Reviewed-by: Mark Brown <broonie@kernel.org>

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[-- Attachment #2: Type: text/plain, Size: 176 bytes --]

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v3 2/2] arm64: Work around missing `bti c` in modules

[Mark Brown]

[-- Attachment #1.1: Type: text/plain, Size: 511 bytes --]

On Thu, Sep 01, 2022 at 05:15:51PM -0700, D Scott Phillips wrote:
> GCC does not insert a `bti c` instruction at the beginning of a function
> when all callers reach the function through a direct branch[1]. In the case
> of cross-section calls (like __init to non __init), a thunk may be inserted
> which uses an indirect branch. If that happens, the first instruction in
> the callee function will result in a Branch Target Exception due to the
> missing `bti c`.

Reviewed-by: Mark Brown <broonie@kernel.org>

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[-- Attachment #2: Type: text/plain, Size: 176 bytes --]

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel