Re: [PATCH v4 15/15] perf test: Remove /usr/bin/cc dependency from Intel PT shell test

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] The `jitdump` workload directly calls `gettid()` without a fallback wrapper, breaking compilation on systems with glibc < 2.30.
--

commit f141af2f5d5b51f3502b379b7b18fe4506f565d2
Author: Ian Rogers <irogers@google.com>

perf test: Remove /usr/bin/cc dependency from Intel PT shell test

This commit introduces a built-in 'jitdump' workload to replace the runtime
compilation step using /usr/bin/cc in the Intel PT shell test. The workload
features multi-architecture support to dynamically encode opcodes without 
requiring external compiler dependencies.

> diff --git a/tools/perf/tests/workloads/jitdump.c b/tools/perf/tests/workloads/jitdump.c
> new file mode 100644
> --- /dev/null
> +++ b/tools/perf/tests/workloads/jitdump.c

[ ... ]

> +static int write_jitdump(FILE *f, void *addr, const void *dat, size_t sz, uint64_t *idx)
> +{
> +	struct jr_code_load rec = {
> +		.p.id = JIT_CODE_LOAD,
> +		.p.total_size = sizeof(rec) + sz,
> +		.p.timestamp = get_timestamp(),
> +		.pid = getpid(),
> +		.tid = gettid(),

[Severity: Low]
Will this direct gettid() call break the build on systems with glibc < 2.30?

The gettid() function is not declared in <unistd.h> on older glibc versions 
(like those found in CentOS 7 or Ubuntu 18.04). Since the perf build system 
treats implicit function declarations as errors, this will unconditionally fail 
the build in those environments.

Other parts of the codebase, such as builtin-record.c or jvmti/jvmti_agent.c, 
use a fallback definition using syscall(__NR_gettid). Could a similar fallback 
be used here?

> +		.vma = (unsigned long)addr,
> +		.code_addr = (unsigned long)addr,
> +		.code_size = sz,
> +		.code_index = ++*idx,
> +	};

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260531082235.940833-1-irogers@google.com?part=15