ARM: net: bpf: clarify tail_call index
authorRussell King <rmk+kernel@armlinux.org.uk>
Sat, 13 Jan 2018 12:11:26 +0000 (12:11 +0000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 31 Jan 2018 13:03:44 +0000 (14:03 +0100)
commit 091f02483df7b56615b524491f404e574c5e0668 upstream.

As per 90caccdd8cc0 ("bpf: fix bpf_tail_call() x64 JIT"), the index used
for array lookup is defined to be 32-bit wide. Update a misleading
comment that suggests it is 64-bit wide.

Fixes: 39c13c204bb1 ("arm: eBPF JIT compiler")
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/arm/net/bpf_jit_32.c

index 715e725..323a4df 100644 (file)
@@ -1016,7 +1016,7 @@ static int emit_bpf_tail_call(struct jit_ctx *ctx)
        emit_a32_mov_i(tmp[1], off, false, ctx);
        emit(ARM_LDR_I(tmp2[1], ARM_SP, STACK_VAR(r2[1])), ctx);
        emit(ARM_LDR_R(tmp[1], tmp2[1], tmp[1]), ctx);
-       /* index (64 bit) */
+       /* index is 32-bit for arrays */
        emit(ARM_LDR_I(tmp2[1], ARM_SP, STACK_VAR(r3[1])), ctx);
        /* index >= array->map.max_entries */
        emit(ARM_CMP_R(tmp2[1], tmp[1]), ctx);