arm64: poe: Handle spurious Overlay faults

We do not currently issue an ISB after updating POR_EL0 when
context-switching it, for instance. The rationale is that if the old
value of POR_EL0 is more restrictive and causes a fault during
uaccess, the access will be retried [1]. In other words, we are
trading an ISB on every context-switching for the (unlikely)
possibility of a spurious fault. We may also miss faults if the new
value of POR_EL0 is more restrictive, but that's considered
acceptable.

However, as things stand, a spurious Overlay fault results in
uaccess failing right away since it causes fault_from_pkey() to
return true. If an Overlay fault is reported, we therefore need to
double check POR_EL0 against vma_pkey(vma) - this is what
arch_vma_access_permitted() already does.

As it turns out, we already perform that explicit check if no
Overlay fault is reported, and we need to keep that check (see
comment added in fault_from_pkey()). Net result: the Overlay ISS2
bit isn't of much help to decide whether a pkey fault occurred.

Remove the check for the Overlay bit from fault_from_pkey() and
add a comment to try and explain the situation. While at it, also
add a comment to permission_overlay_switch() in case anyone gets
surprised by the lack of ISB.

[1] https://lore.kernel.org/linux-arm-kernel/ZtYNGBrcE-j35fpw@arm.com/

Fixes: 160a8e1 ("arm64: context switch POR_EL0 register")
Signed-off-by: Kevin Brodsky <kevin.brodsky@arm.com>
Link: https://lore.kernel.org/r/20250619160042.2499290-2-kevin.brodsky@arm.com
Signed-off-by: Will Deacon <will@kernel.org>

Author: kevin-brodsky-arm

arch/arm64/kernel/process.c

@@ -673,6 +673,11 @@ static void permission_overlay_switch(struct task_struct *next)
 	current->thread.por_el0 = read_sysreg_s(SYS_POR_EL0);
 	if (current->thread.por_el0 != next->thread.por_el0) {
 		write_sysreg_s(next->thread.por_el0, SYS_POR_EL0);
+		/*
+		 * No ISB required as we can tolerate spurious Overlay faults -
+		 * the fault handler will check again based on the new value
+		 * of POR_EL0.
+		 */
 	}
 }
 

arch/arm64/mm/fault.c

@@ -487,17 +487,29 @@ static void do_bad_area(unsigned long far, unsigned long esr,
 	}
 }
 
-static bool fault_from_pkey(unsigned long esr, struct vm_area_struct *vma,
-			unsigned int mm_flags)
+static bool fault_from_pkey(struct vm_area_struct *vma, unsigned int mm_flags)
 {
-	unsigned long iss2 = ESR_ELx_ISS2(esr);
-
 	if (!system_supports_poe())
 		return false;
 
-	if (esr_fsc_is_permission_fault(esr) && (iss2 & ESR_ELx_Overlay))
-		return true;
-
+	/*
+	 * We do not check whether an Overlay fault has occurred because we
+	 * cannot make a decision based solely on its value:
+	 *
+	 * - If Overlay is set, a fault did occur due to POE, but it may be
+	 *   spurious in those cases where we update POR_EL0 without ISB (e.g.
+	 *   on context-switch). We would then need to manually check POR_EL0
+	 *   against vma_pkey(vma), which is exactly what
+	 *   arch_vma_access_permitted() does.
+	 *
+	 * - If Overlay is not set, we may still need to report a pkey fault.
+	 *   This is the case if an access was made within a mapping but with no
+	 *   page mapped, and POR_EL0 forbids the access (according to
+	 *   vma_pkey()). Such access will result in a SIGSEGV regardless
+	 *   because core code checks arch_vma_access_permitted(), but in order
+	 *   to report the correct error code - SEGV_PKUERR - we must handle
+	 *   that case here.
+	 */
 	return !arch_vma_access_permitted(vma,
 			mm_flags & FAULT_FLAG_WRITE,
 			mm_flags & FAULT_FLAG_INSTRUCTION,
@@ -635,7 +647,7 @@ static int __kprobes do_page_fault(unsigned long far, unsigned long esr,
 		goto bad_area;
 	}
 
-	if (fault_from_pkey(esr, vma, mm_flags)) {
+	if (fault_from_pkey(vma, mm_flags)) {
 		pkey = vma_pkey(vma);
 		vma_end_read(vma);
 		fault = 0;
@@ -679,7 +691,7 @@ static int __kprobes do_page_fault(unsigned long far, unsigned long esr,
 		goto bad_area;
 	}
 
-	if (fault_from_pkey(esr, vma, mm_flags)) {
+	if (fault_from_pkey(vma, mm_flags)) {
 		pkey = vma_pkey(vma);
 		mmap_read_unlock(mm);
 		fault = 0;