This patchseries fixes all failing bpf verifier selftests:
root@intel-x86-64:~# ./test_verifier ... #1054/p XDP pkt read, pkt_meta' <= pkt_data, bad access 1 OK #1055/p XDP pkt read, pkt_meta' <= pkt_data, bad access 2 OK #1056/p XDP pkt read, pkt_data <= pkt_meta', good access OK #1057/p XDP pkt read, pkt_data <= pkt_meta', bad access 1 OK #1058/p XDP pkt read, pkt_data <= pkt_meta', bad access 2 OK Summary: 1566 PASSED, 0 SKIPPED, 0 FAILED
Daniel Borkmann (1): bpf, selftests: Adjust few selftest result_unpriv outcomes
tools/testing/selftests/bpf/verifier/stack_ptr.c | 2 -- tools/testing/selftests/bpf/verifier/value_ptr_arith.c | 8 -------- 2 files changed, 10 deletions(-)
From: Daniel Borkmann daniel@iogearbox.net
Given we don't need to simulate the speculative domain for registers with immediates anymore since the verifier uses direct imm-based rewrites instead of having to mask, we can also lift a few cases that were previously rejected.
Signed-off-by: Daniel Borkmann daniel@iogearbox.net Acked-by: Alexei Starovoitov ast@kernel.org [OP: backport to 5.4, small context adjustment in stack_ptr.c] Signed-off-by: Ovidiu Panait ovidiu.panait@windriver.com --- tools/testing/selftests/bpf/verifier/stack_ptr.c | 2 -- tools/testing/selftests/bpf/verifier/value_ptr_arith.c | 8 -------- 2 files changed, 10 deletions(-)
diff --git a/tools/testing/selftests/bpf/verifier/stack_ptr.c b/tools/testing/selftests/bpf/verifier/stack_ptr.c index 7276620ef242..53d2a5a5ec58 100644 --- a/tools/testing/selftests/bpf/verifier/stack_ptr.c +++ b/tools/testing/selftests/bpf/verifier/stack_ptr.c @@ -291,8 +291,6 @@ BPF_LDX_MEM(BPF_B, BPF_REG_0, BPF_REG_1, 0), BPF_EXIT_INSN(), }, - .result_unpriv = REJECT, - .errstr_unpriv = "invalid stack off=0 size=1", .result = ACCEPT, .retval = 42, }, diff --git a/tools/testing/selftests/bpf/verifier/value_ptr_arith.c b/tools/testing/selftests/bpf/verifier/value_ptr_arith.c index 28d44e6aa0b7..f9c91b95080e 100644 --- a/tools/testing/selftests/bpf/verifier/value_ptr_arith.c +++ b/tools/testing/selftests/bpf/verifier/value_ptr_arith.c @@ -300,8 +300,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT, - .result_unpriv = REJECT, - .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, { @@ -371,8 +369,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT, - .result_unpriv = REJECT, - .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, { @@ -472,8 +468,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT, - .result_unpriv = REJECT, - .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, { @@ -766,8 +760,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT, - .result_unpriv = REJECT, - .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, {
On 04.08.2021 20:20, Ovidiu Panait wrote:
From: Daniel Borkmann daniel@iogearbox.net
Given we don't need to simulate the speculative domain for registers with immediates anymore since the verifier uses direct imm-based rewrites instead of having to mask, we can also lift a few cases that were previously rejected.
Signed-off-by: Daniel Borkmann daniel@iogearbox.net Acked-by: Alexei Starovoitov ast@kernel.org [OP: backport to 5.4, small context adjustment in stack_ptr.c] Signed-off-by: Ovidiu Panait ovidiu.panait@windriver.com
Hi Greg,
It seems that this patch was missed for the previous 5.4 release, could it be included in the upcoming release?
Thanks!
Ovidiu
tools/testing/selftests/bpf/verifier/stack_ptr.c | 2 -- tools/testing/selftests/bpf/verifier/value_ptr_arith.c | 8 -------- 2 files changed, 10 deletions(-)
diff --git a/tools/testing/selftests/bpf/verifier/stack_ptr.c b/tools/testing/selftests/bpf/verifier/stack_ptr.c index 7276620ef242..53d2a5a5ec58 100644 --- a/tools/testing/selftests/bpf/verifier/stack_ptr.c +++ b/tools/testing/selftests/bpf/verifier/stack_ptr.c @@ -291,8 +291,6 @@ BPF_LDX_MEM(BPF_B, BPF_REG_0, BPF_REG_1, 0), BPF_EXIT_INSN(), },
- .result_unpriv = REJECT,
- .errstr_unpriv = "invalid stack off=0 size=1", .result = ACCEPT, .retval = 42, },
diff --git a/tools/testing/selftests/bpf/verifier/value_ptr_arith.c b/tools/testing/selftests/bpf/verifier/value_ptr_arith.c index 28d44e6aa0b7..f9c91b95080e 100644 --- a/tools/testing/selftests/bpf/verifier/value_ptr_arith.c +++ b/tools/testing/selftests/bpf/verifier/value_ptr_arith.c @@ -300,8 +300,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT,
- .result_unpriv = REJECT,
- .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, {
@@ -371,8 +369,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT,
- .result_unpriv = REJECT,
- .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, {
@@ -472,8 +468,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT,
- .result_unpriv = REJECT,
- .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, {
@@ -766,8 +760,6 @@ }, .fixup_map_array_48b = { 3 }, .result = ACCEPT,
- .result_unpriv = REJECT,
- .errstr_unpriv = "R0 pointer arithmetic of map value goes out of range", .retval = 1, }, {
On Tue, Aug 10, 2021 at 11:53:06AM +0300, Ovidiu Panait wrote:
On 04.08.2021 20:20, Ovidiu Panait wrote:
From: Daniel Borkmann daniel@iogearbox.net
Given we don't need to simulate the speculative domain for registers with immediates anymore since the verifier uses direct imm-based rewrites instead of having to mask, we can also lift a few cases that were previously rejected.
Signed-off-by: Daniel Borkmann daniel@iogearbox.net Acked-by: Alexei Starovoitov ast@kernel.org [OP: backport to 5.4, small context adjustment in stack_ptr.c] Signed-off-by: Ovidiu Panait ovidiu.panait@windriver.com
Hi Greg,
It seems that this patch was missed for the previous 5.4 release, could it be included in the upcoming release?
Ick, sorry about that, I missed it, my fault.
I'll go queue it up now, thanks.
greg k-h
linux-stable-mirror@lists.linaro.org
-
Greg KH -
Ovidiu Panait