Re: [PATCH v2 7/7] tools/testing/selftests/bpf: replace open-coded 16 with TASK_COMM_LEN
From: Steven Rostedt <rostedt@goodmis.org>
Date: 2021-11-29 16:10:09
Also in:
bpf, linux-fsdevel, linux-mm, lkml, netdev
Possibly related (same subject, not in this thread)
- 2023-02-13 · Re: [PATCH v2 7/7] tools/testing/selftests/bpf: replace open-coded 16 with TASK_COMM_LEN · Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
- 2023-02-13 · Re: [PATCH v2 7/7] tools/testing/selftests/bpf: replace open-coded 16 with TASK_COMM_LEN · Namhyung Kim <hidden>
- 2023-02-12 · Re: [PATCH v2 7/7] tools/testing/selftests/bpf: replace open-coded 16 with TASK_COMM_LEN · Steven Rostedt <rostedt@goodmis.org>
- 2023-02-12 · Re: [PATCH v2 7/7] tools/testing/selftests/bpf: replace open-coded 16 with TASK_COMM_LEN · Yafang Shao <hidden>
- 2023-02-12 · Re: [PATCH v2 7/7] tools/testing/selftests/bpf: replace open-coded 16 with TASK_COMM_LEN · Yafang Shao <hidden>
On Mon, 29 Nov 2021 23:33:33 +0800 Yafang Shao [off-list ref] wrote:
quoted
TBH, i would vote for reverting the change. defining an array size as enum feels really odd.We changed it to enum because the BTF can't parse macro while it can parse the enum type.
I wonder if BTF could take advantage of the tracing: TRACE_DEFINE_ENUM() macros? This is how they are converted for user space tooling. Anyway, I'd have to go and look at why that trigger test failed. I don't see how the size of the array caused it to change the signage of value. -- Steve
Anyway I don't insist on keeping this change if you think reverting it is better.