linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: Mark Rutland <mark.rutland@arm.com>
To: linux-arm-kernel@lists.infradead.org
Cc: catalin.marinas@arm.com, james.morse@arm.com, joey.gouly@arm.com,
	mark.rutland@arm.com, peterz@infradead.org, will@kernel.org
Subject: [PATCH 2/9] arm64: split EL0/EL1 UNDEF handlers
Date: Wed, 28 Sep 2022 11:26:52 +0100	[thread overview]
Message-ID: <20220928102659.247510-3-mark.rutland@arm.com> (raw)
In-Reply-To: <20220928102659.247510-1-mark.rutland@arm.com>

In general, exceptions taken from EL1 need to be handled separately from
exceptions taken from EL0, as the logic to handle the two cases can be
significantly divergent, and exceptions taken from EL1 typically have
more stringent requirements on locking and instrumentation.

Subsequent patches will rework the way EL1 UNDEFs are handled in order
to address longstanding soundness issues with instrumentation and RCU.
In preparation for that rework, this patch splits the existing
do_undefinstr() handler into separate do_el0_undef() and do_el1_undef()
handlers.

Prior to this patch, do_undefinstr() was marked with NOKPROBE_SYMBOL(),
preventing instrumentation via kprobes. However, do_undefinstr() invokes
other code which can be instrumented, and:

* For UNDEFINED exceptions taken from EL0, there is no risk of recursion
  within kprobes. Therefore it is safe for do_el0_undef to be
  instrumented with kprobes, and it does not need to be marked with
  NOKPROBE_SYMBOL().

* For UNDEFINED exceptions taken from EL1, either:

  (a) The exception is has been taken when manipulating SSBS; these cases
      are limited and do not occur within code that can be invoked
      recursively via kprobes. Hence, in these cases instrumentation
      with kprobes is benign.

  (b) The exception has been taken for an unknown reason, as other than
      manipulating SSBS we do not expect to take UNDEFINED exceptions
      from EL1. Any handling of these exception is best-effort.

  ... and in either case, marking do_el1_undef() with NOKPROBE_SYMBOL()
  isn't sufficient to prvent recursion via kprobes as functions it calls
  (including die()) are instrumentable via kprobes.

  Hence, it's not worthwhile to mark do_el1_undef() with
  NOKPROBE_SYMBOL(). The same applies to do_el1_bti() and do_el1_fpac(),
  so their NOKPROBE_SYMBOL() annotations are also removed.

Aside from the new instrumentability, there should be no functional
change as a result of this patch.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: James Morse <james.morse@arm.com>
Cc: Joey Gouly <joey.gouly@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
---
 arch/arm64/include/asm/exception.h |  3 ++-
 arch/arm64/kernel/entry-common.c   |  4 ++--
 arch/arm64/kernel/traps.c          | 22 ++++++++++++----------
 3 files changed, 16 insertions(+), 13 deletions(-)

diff --git a/arch/arm64/include/asm/exception.h b/arch/arm64/include/asm/exception.h
index bc0dd002106d1..92963f98afece 100644
--- a/arch/arm64/include/asm/exception.h
+++ b/arch/arm64/include/asm/exception.h
@@ -58,7 +58,8 @@ asmlinkage void call_on_irq_stack(struct pt_regs *regs,
 asmlinkage void asm_exit_to_user_mode(struct pt_regs *regs);
 
 void do_mem_abort(unsigned long far, unsigned long esr, struct pt_regs *regs);
-void do_undefinstr(struct pt_regs *regs, unsigned long esr);
+void do_el0_undef(struct pt_regs *regs, unsigned long esr);
+void do_el1_undef(struct pt_regs *regs, unsigned long esr);
 void do_el0_bti(struct pt_regs *regs);
 void do_el1_bti(struct pt_regs *regs, unsigned long esr);
 void do_debug_exception(unsigned long addr_if_watchpoint, unsigned long esr,
diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c
index 135d4a5868fe5..a38c3c4f94c05 100644
--- a/arch/arm64/kernel/entry-common.c
+++ b/arch/arm64/kernel/entry-common.c
@@ -383,7 +383,7 @@ static void noinstr el1_undef(struct pt_regs *regs, unsigned long esr)
 {
 	enter_from_kernel_mode(regs);
 	local_daif_inherit(regs);
-	do_undefinstr(regs, esr);
+	do_el1_undef(regs, esr);
 	local_daif_mask();
 	exit_to_kernel_mode(regs);
 }
@@ -598,7 +598,7 @@ static void noinstr el0_undef(struct pt_regs *regs, unsigned long esr)
 {
 	enter_from_user_mode(regs);
 	local_daif_restore(DAIF_PROCCTX);
-	do_undefinstr(regs, esr);
+	do_el0_undef(regs, esr);
 	exit_to_user_mode(regs);
 }
 
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index 63d3b79205b3d..37dea5062bc1f 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -485,7 +485,7 @@ void arm64_notify_segfault(unsigned long addr)
 	force_signal_inject(SIGSEGV, code, addr, 0);
 }
 
-void do_undefinstr(struct pt_regs *regs, unsigned long esr)
+void do_el0_undef(struct pt_regs *regs, unsigned long esr)
 {
 	/* check for AArch32 breakpoint instructions */
 	if (!aarch32_break_handler(regs))
@@ -494,12 +494,16 @@ void do_undefinstr(struct pt_regs *regs, unsigned long esr)
 	if (call_undef_hook(regs) == 0)
 		return;
 
-	if (!user_mode(regs))
-		die("Oops - Undefined instruction", regs, esr);
-
 	force_signal_inject(SIGILL, ILL_ILLOPC, regs->pc, 0);
 }
-NOKPROBE_SYMBOL(do_undefinstr);
+
+void do_el1_undef(struct pt_regs *regs, unsigned long esr)
+{
+	if (call_undef_hook(regs) == 0)
+		return;
+
+	die("Oops - Undefined instruction", regs, esr);
+}
 
 void do_el0_bti(struct pt_regs *regs)
 {
@@ -510,7 +514,6 @@ void do_el1_bti(struct pt_regs *regs, unsigned long esr)
 {
 	die("Oops - BTI", regs, esr);
 }
-NOKPROBE_SYMBOL(do_el1_bti);
 
 void do_el0_fpac(struct pt_regs *regs, unsigned long esr)
 {
@@ -525,7 +528,6 @@ void do_el1_fpac(struct pt_regs *regs, unsigned long esr)
 	 */
 	die("Oops - FPAC", regs, esr);
 }
-NOKPROBE_SYMBOL(do_el1_fpac)
 
 #define __user_cache_maint(insn, address, res)			\
 	if (address >= TASK_SIZE_MAX) {				\
@@ -768,7 +770,7 @@ void do_el0_cp15(unsigned long esr, struct pt_regs *regs)
 		hook_base = cp15_64_hooks;
 		break;
 	default:
-		do_undefinstr(regs, esr);
+		do_el0_undef(regs, esr);
 		return;
 	}
 
@@ -783,7 +785,7 @@ void do_el0_cp15(unsigned long esr, struct pt_regs *regs)
 	 * EL0. Fall back to our usual undefined instruction handler
 	 * so that we handle these consistently.
 	 */
-	do_undefinstr(regs, esr);
+	do_el0_undef(regs, esr);
 }
 #endif
 
@@ -802,7 +804,7 @@ void do_el0_sys(unsigned long esr, struct pt_regs *regs)
 	 * back to our usual undefined instruction handler so that we handle
 	 * these consistently.
 	 */
-	do_undefinstr(regs, esr);
+	do_el0_undef(regs, esr);
 }
 
 static const char *esr_class_str[] = {
-- 
2.30.2


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

  parent reply	other threads:[~2022-09-28 10:28 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-28 10:26 [PATCH 0/9] arm64: rework UNDEFINED instruction traps Mark Rutland
2022-09-28 10:26 ` [PATCH 1/9] arm64: allow kprobes on EL0 handlers Mark Rutland
2022-09-28 10:26 ` Mark Rutland [this message]
2022-09-28 10:26 ` [PATCH 3/9] arm64: factor out EL1 SSBS emulation hook Mark Rutland
2022-09-28 10:26 ` [PATCH 4/9] arm64: factor insn read out of call_undef_hook() Mark Rutland
2022-09-28 10:26 ` [PATCH 5/9] arm64: rework EL0 MRS emulation Mark Rutland
2022-09-28 10:26 ` [PATCH 6/9] arm64: armv8_deprecated: fold ops into insn_emulation Mark Rutland
2022-09-28 10:26 ` [PATCH 7/9] arm64: armv8_deprecated move emulation functions Mark Rutland
2022-09-28 10:26 ` [PATCH 8/9] arm64: armv8_deprecated: move aarch32 helper earlier Mark Rutland
2022-09-28 10:26 ` [PATCH 9/9] arm64: armv8_deprecated: rework deprected instruction handling Mark Rutland

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=20220928102659.247510-3-mark.rutland@arm.com \
    --to=mark.rutland@arm.com \
    --cc=catalin.marinas@arm.com \
    --cc=james.morse@arm.com \
    --cc=joey.gouly@arm.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=peterz@infradead.org \
    --cc=will@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;
as well as URLs for NNTP newsgroup(s).