Skip to contentSkip to conversation

selftests/bpf: fix timer/test_bad_ret subtest on test_progs-cpuv4 flavor

PATCH Applied Andrii Nakryiko 2 participants 2 messages Dec 8, 2023 – Dec 9, 2023

Subsystems (3)
lore
Share & tools
lore

Andrii Nakryiko

author maintainer
#0 lore raw

PATCH 1/1 selftests/bpf: fix timer/test_bad_ret subtest on test_progs-cpuv4 flavor

Because test_bad_ret main program is not written in assembly, we don't control instruction indices in timer_cb_ret_bad() subprog. This bites us in timer/test_bad_ret subtest, where we see difference between cpuv4 and other flavors.

For now, make __msg() expectations not rely on instruction indices by anchoring them around bpf_get_prandom_u32 call. Once we have regex/glob support for __msg(), this can be expressed a bit more nicely, but for now just mitigating the problem with available means.

Fixes e02dea158dda ("selftests/bpf: validate async callback return value check correctness")
---
 tools/testing/selftests/bpf/progs/timer_failure.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
tools/testing/selftests/bpf/progs/timer_failure.c +3 -2
index 9fbc69c77bbb..0996c2486f05 100644
@@ -47,9 +47,10 @@ __log_level(2)
 __flag(BPF_F_TEST_STATE_FREQ)
 __failure
 /* check that fallthrough code path marks r0 as precise */
-__msg("mark_precise: frame0: regs=r0 stack= before 22: (b7) r0 = 0")
+__msg("mark_precise: frame0: regs=r0 stack= before")
+__msg(": (85) call bpf_get_prandom_u32#7") /* anchor message */
 /* check that branch code path marks r0 as precise */
-__msg("mark_precise: frame0: regs=r0 stack= before 24: (85) call bpf_get_prandom_u32#7")
+__msg("mark_precise: frame0: regs=r0 stack= before ") __msg(": (85) call bpf_get_prandom_u32#7")
 __msg("should have been in [0, 0]")
 long BPF_PROG2(test_bad_ret, int, a)
 {
Signature · 1 line
2.34.1
#1 lore raw
This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Quoted 11 lines · On Fri, 8 Dec 2023 15:30:28 -0800 you wrote:

Because test_bad_ret main program is not written in assembly, we don't control instruction indices in timer_cb_ret_bad() subprog. This bites us in timer/test_bad_ret subtest, where we see difference between cpuv4 and other flavors.

For now, make __msg() expectations not rely on instruction indices by anchoring them around bpf_get_prandom_u32 call. Once we have regex/glob support for __msg(), this can be expressed a bit more nicely, but for now just mitigating the problem with available means.

[...]

Here is the summary with links:
- [bpf-next] selftests/bpf: fix timer/test_bad_ret subtest on test_progs-cpuv4 flavor

    https://git.kernel.org/bpf/bpf-next/c/1720c42b90c8

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html