Re: [PATCH v1] arm64: poe: fix stale POR_EL0 values for ptrace

On Wed, Jan 21, 2026 at 03:44:21PM +0000, Mark Rutland wrote:

Hi Joey,

On Wed, Jan 21, 2026 at 01:56:39PM +0000, Joey Gouly wrote:

quoted

If a process wrote to POR_EL0 and then crashed before a context switch
happened, the coredump would contain an incorrect value for POR_EL0.

The value read in poe_get() would be a stale value left in thread.por_el0.  Fix
this by reading the value from the system register, if the target thread is the
current thread.

This matches what gcs/fpsimd do.

Fixes: 175198199262 ("arm64/ptrace: add support for FEAT_POE")
Reported-by: David Spickett <redacted>
Cc: stable@vger.kernel.org
Signed-off-by: Joey Gouly <joey.gouly@arm.com>
Cc: Kevin Brodsky <redacted>
Cc: Mark Rutland <mark.rutland@arm.com>

I have a couple of comments below, but as-is this looks functionally
correct to me. With or without the changes suggested below:

Acked-by: Mark Rutland <mark.rutland@arm.com>

quoted

---
 arch/arm64/include/asm/por.h | 2 ++
 arch/arm64/kernel/process.c  | 7 ++++++-
 arch/arm64/kernel/ptrace.c   | 5 +++++
 3 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/include/asm/por.h b/arch/arm64/include/asm/por.h
index d913d5b529e4..46f1356837e2 100644
--- a/arch/arm64/include/asm/por.h
+++ b/arch/arm64/include/asm/por.h

@@ -31,4 +31,6 @@ static inline bool por_elx_allows_exec(u64 por, u8 pkey)
 	return perm & POE_X;
 }
 
+void poe_preserve_current_state(void);

Is it possible to have a static inline here, i.e.

	static inline void poe_preserve_current_state(void)
	{
		current->thread.por_el0 = read_sysreg_s(SYS_POR_EL0);
	}

... or will that cause some header dependency problem?

It is possible to do as a static inline..

If we can have this as a static inline, we can use it everywhere
consistently, and avoid needing a function call for a trivial number of
instructions.

Otherwise, see below for another option.

quoted

+
 #endif /* _ASM_ARM64_POR_H */

diff --git a/arch/arm64/kernel/process.c b/arch/arm64/kernel/process.c
index 489554931231..400182099784 100644
--- a/arch/arm64/kernel/process.c
+++ b/arch/arm64/kernel/process.c

@@ -665,12 +665,17 @@ static int do_set_tsc_mode(unsigned int val)
 	return 0;
 }
 
+void poe_preserve_current_state(void)
+{
+	current->thread.por_el0 = read_sysreg_s(SYS_POR_EL0);
+}
+
 static void permission_overlay_switch(struct task_struct *next)
 {
 	if (!system_supports_poe())
 		return;
 
-	current->thread.por_el0 = read_sysreg_s(SYS_POR_EL0);
+	poe_preserve_current_state();
 	if (current->thread.por_el0 != next->thread.por_el0) {
 		write_sysreg_s(next->thread.por_el0, SYS_POR_EL0);
 		/*

diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index b9bdd83fbbca..276d8ee630cd 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c

@@ -37,6 +37,7 @@
 #include <asm/gcs.h>
 #include <asm/mte.h>
 #include <asm/pointer_auth.h>
+#include <asm/por.h>
 #include <asm/stacktrace.h>
 #include <asm/syscall.h>
 #include <asm/traps.h>

@@ -1486,6 +1487,10 @@ static int poe_get(struct task_struct *target,
 	if (!system_supports_poe())
 		return -EINVAL;
 
+	if (target == current) {
+		poe_preserve_current_state();
+	}

If we can't do the static inline, it might be best to just open code the
read here, i.e. make this:

	if (target == current)
		current->thread.por_el0 = read_sysreg_s(SYS_POR_EL0);

.. however I think this approach is better, no need for abstraction over this.

I will send a v2 in a day or so, in case there are other comments /
disagreements with that approach.

Thanks,
Joey

... since permission_overlay_switch() writes the register directly,
we're not really gaining abstraction by factoring this out. Open coding
would make the diff a bit smaller, and avoid the function call.

That all said, this is functionally correct either way, so if Catalin or
Will disagree, go with whatever they prefer!

Mark.