[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <160097312263.12106.17437529842919754724.stgit@john-Precision-5820-Tower>
Date: Thu, 24 Sep 2020 11:45:22 -0700
From: John Fastabend <john.fastabend@...il.com>
To: ast@...nel.org, daniel@...earbox.net
Cc: netdev@...r.kernel.org, john.fastabend@...il.com
Subject: [bpf-next PATCH 2/2] bpf: Add AND verifier test case where 32bit
and 64bit bounds differ
If we AND two values together that are known in the 32bit subregs, but not
known in the 64bit registers we rely on the tnum value to report the 32bit
subreg is known. And do not use mark_reg_known() directly from
scalar32_min_max_and()
Add an AND test to cover the case with known 32bit subreg, but unknown
64bit reg.
Signed-off-by: John Fastabend <john.fastabend@...il.com>
---
tools/testing/selftests/bpf/verifier/and.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/tools/testing/selftests/bpf/verifier/and.c b/tools/testing/selftests/bpf/verifier/and.c
index d781bc86e100..ca8fdb1b3f01 100644
--- a/tools/testing/selftests/bpf/verifier/and.c
+++ b/tools/testing/selftests/bpf/verifier/and.c
@@ -48,3 +48,19 @@
.result = REJECT,
.flags = F_NEEDS_EFFICIENT_UNALIGNED_ACCESS,
},
+{
+ "check known subreg with unknown reg",
+ .insns = {
+ BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0, BPF_FUNC_get_prandom_u32),
+ BPF_ALU64_IMM(BPF_LSH, BPF_REG_0, 32),
+ BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1),
+ BPF_ALU64_IMM(BPF_AND, BPF_REG_0, 0xFFFF1234),
+ /* Upper bits are unknown but AND above masks out 1 zero'ing lower bits */
+ BPF_JMP32_IMM(BPF_JLT, BPF_REG_0, 1, 1),
+ BPF_LDX_MEM(BPF_W, BPF_REG_1, BPF_REG_1, 512),
+ BPF_MOV64_IMM(BPF_REG_0, 0),
+ BPF_EXIT_INSN(),
+ },
+ .result = ACCEPT,
+ .retval = 0
+},
Powered by blists - more mailing lists