selftests/bpf: Reject graph kptr use after RCU unlockgrafted

Add a sleepable verifier test that loads a graph-node local kptr in an
explicit RCU read-side critical section, then passes its node to
bpf_rbtree_remove() after the section ends.

Before the verifier fix, the stale NON_OWN_REF flag makes the node look like
a live borrowed reference and the program is accepted. After the fix, the
pointer is demoted without NON_OWN_REF and the graph kfunc argument is
rejected.

Also exercise a graph kptr loaded while a spin lock provides implicit RCU
protection. The pointer must be invalidated when the lock is released, which
guards the required ordering between non-owning-reference invalidation and
RCU demotion.

Update the existing fault-protected load test state description. The
post-unlock pointer no longer carries NON_OWN_REF, but remains readable
because the load is rewritten to use BPF_PROBE_MEM.

Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260904084325.52250-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

1 files changed, 120 insertions, 0 deletions

diff --git a/arch/arm/lib/io-readsb.S b/arch/arm/lib/io-readsb.S
new file mode 100644
index 000000000..0def9388f
--- /dev/null
+++ b/arch/arm/lib/io-readsb.S
@@ -0,0 +1,120 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ *  linux/arch/arm/lib/io-readsb.S
+ *
+ *  Copyright (C) 1995-2000 Russell King
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+.Linsb_align:	rsb	ip, ip, #4
+		cmp	ip, r2
+		movgt	ip, r2
+		cmp	ip, #2
+		ldrb	r3, [r0]
+		strb	r3, [r1], #1
+		ldrbge	r3, [r0]
+		strbge	r3, [r1], #1
+		ldrbgt	r3, [r0]
+		strbgt	r3, [r1], #1
+		subs	r2, r2, ip
+		bne	.Linsb_aligned
+
+ENTRY(__raw_readsb)
+		teq	r2, #0		@ do we have to check for the zero len?
+		reteq	lr
+		ands	ip, r1, #3
+		bne	.Linsb_align
+
+.Linsb_aligned:	stmfd	sp!, {r4 - r6, lr}
+
+		subs	r2, r2, #16
+		bmi	.Linsb_no_16
+
+.Linsb_16_lp:	ldrb	r3, [r0]
+		ldrb	r4, [r0]
+		ldrb	r5, [r0]
+		mov	r3, r3,     put_byte_0
+		ldrb	r6, [r0]
+		orr	r3, r3, r4, put_byte_1
+		ldrb	r4, [r0]
+		orr	r3, r3, r5, put_byte_2
+		ldrb	r5, [r0]
+		orr	r3, r3, r6, put_byte_3
+		ldrb	r6, [r0]
+		mov	r4, r4,     put_byte_0
+		ldrb	ip, [r0]
+		orr	r4, r4, r5, put_byte_1
+		ldrb	r5, [r0]
+		orr	r4, r4, r6, put_byte_2
+		ldrb	r6, [r0]
+		orr	r4, r4, ip, put_byte_3
+		ldrb	ip, [r0]
+		mov	r5, r5,     put_byte_0
+		ldrb	lr, [r0]
+		orr	r5, r5, r6, put_byte_1
+		ldrb	r6, [r0]
+		orr	r5, r5, ip, put_byte_2
+		ldrb	ip, [r0]
+		orr	r5, r5, lr, put_byte_3
+		ldrb	lr, [r0]
+		mov	r6, r6,     put_byte_0
+		orr	r6, r6, ip, put_byte_1
+		ldrb	ip, [r0]
+		orr	r6, r6, lr, put_byte_2
+		orr	r6, r6, ip, put_byte_3
+		stmia	r1!, {r3 - r6}
+
+		subs	r2, r2, #16
+		bpl	.Linsb_16_lp
+
+		tst	r2, #15
+		ldmfdeq	sp!, {r4 - r6, pc}
+
+.Linsb_no_16:	tst	r2, #8
+		beq	.Linsb_no_8
+
+		ldrb	r3, [r0]
+		ldrb	r4, [r0]
+		ldrb	r5, [r0]
+		mov	r3, r3,     put_byte_0
+		ldrb	r6, [r0]
+		orr	r3, r3, r4, put_byte_1
+		ldrb	r4, [r0]
+		orr	r3, r3, r5, put_byte_2
+		ldrb	r5, [r0]
+		orr	r3, r3, r6, put_byte_3
+		ldrb	r6, [r0]
+		mov	r4, r4,     put_byte_0
+		ldrb	ip, [r0]
+		orr	r4, r4, r5, put_byte_1
+		orr	r4, r4, r6, put_byte_2
+		orr	r4, r4, ip, put_byte_3
+		stmia	r1!, {r3, r4}
+
+.Linsb_no_8:	tst	r2, #4
+		beq	.Linsb_no_4
+
+		ldrb	r3, [r0]
+		ldrb	r4, [r0]
+		ldrb	r5, [r0]
+		ldrb	r6, [r0]
+		mov	r3, r3,     put_byte_0
+		orr	r3, r3, r4, put_byte_1
+		orr	r3, r3, r5, put_byte_2
+		orr	r3, r3, r6, put_byte_3
+		str	r3, [r1], #4
+
+.Linsb_no_4:	ands	r2, r2, #3
+		ldmfdeq	sp!, {r4 - r6, pc}
+
+		cmp	r2, #2
+		ldrb	r3, [r0]
+		strb	r3, [r1], #1
+		ldrbge	r3, [r0]
+		strbge	r3, [r1], #1
+		ldrbgt	r3, [r0]
+		strbgt	r3, [r1]
+
+		ldmfd	sp!, {r4 - r6, pc}
+ENDPROC(__raw_readsb)