Re: [PATCH v4 06/11] perf dso: Add support for reading the e_machine type for a dso

From: Namhyung Kim <namhyung@kernel.org>
To: Ian Rogers <irogers@google.com>
Cc: "Peter Zijlstra" <peterz@infradead.org>,
	"Ingo Molnar" <mingo@redhat.com>,
	"Arnaldo Carvalho de Melo" <acme@kernel.org>,
	"Mark Rutland" <mark.rutland@arm.com>,
	"Alexander Shishkin" <alexander.shishkin@linux.intel.com>,
	"Jiri Olsa" <jolsa@kernel.org>,
	"Adrian Hunter" <adrian.hunter@intel.com>,
	"Kan Liang" <kan.liang@linux.intel.com>,
	"John Garry" <john.g.garry@oracle.com>,
	"Will Deacon" <will@kernel.org>,
	"James Clark" <james.clark@linaro.org>,
	"Mike Leach" <mike.leach@linaro.org>,
	"Leo Yan" <leo.yan@linux.dev>, guoren <guoren@kernel.org>,
	"Paul Walmsley" <paul.walmsley@sifive.com>,
	"Palmer Dabbelt" <palmer@dabbelt.com>,
	"Albert Ou" <aou@eecs.berkeley.edu>,
	"Charlie Jenkins" <charlie@rivosinc.com>,
	"Bibo Mao" <maobibo@loongson.cn>,
	"Huacai Chen" <chenhuacai@kernel.org>,
	"Catalin Marinas" <catalin.marinas@arm.com>,
	"Jiri Slaby" <jirislaby@kernel.org>,
	"Björn Töpel" <bjorn@rivosinc.com>,
	"Howard Chu" <howardchu95@gmail.com>,
	linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	"linux-csky@vger.kernel.org" <linux-csky@vger.kernel.org>,
	linux-riscv@lists.infradead.org, "Arnd Bergmann" <arnd@arndb.de>
Subject: Re: [PATCH v4 06/11] perf dso: Add support for reading the e_machine type for a dso
Date: Thu, 6 Mar 2025 17:22:20 -0800	[thread overview]
Message-ID: <Z8pKTE7tOqdqNUdA@google.com> (raw)
In-Reply-To: <20250304050305.901167-7-irogers@google.com>

On Mon, Mar 03, 2025 at 09:03:00PM -0800, Ian Rogers wrote:
> For ELF file dsos read the e_machine from the ELF header. For kernel
> types assume the e_machine matches the perf tool. In other cases
> return EM_NONE.
> 
> Signed-off-by: Ian Rogers <irogers@google.com>
> ---
>  tools/perf/util/dso.c | 54 +++++++++++++++++++++++++++++++++++++++++++
>  tools/perf/util/dso.h |  1 +
>  2 files changed, 55 insertions(+)
> 
> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
> index 5c6e85fdae0d..7f2f1af4f73b 100644
> --- a/tools/perf/util/dso.c
> +++ b/tools/perf/util/dso.c
> @@ -1170,6 +1170,60 @@ ssize_t dso__data_read_offset(struct dso *dso, struct machine *machine,
>  	return data_read_write_offset(dso, machine, offset, data, size, true);
>  }
>  
> +uint16_t dso__e_machine(struct dso *dso, struct machine *machine)
> +{
> +	uint16_t e_machine = EM_NONE;
> +	int fd;
> +
> +	switch (dso__binary_type(dso)) {
> +	case DSO_BINARY_TYPE__KALLSYMS:
> +	case DSO_BINARY_TYPE__GUEST_KALLSYMS:
> +	case DSO_BINARY_TYPE__VMLINUX:
> +	case DSO_BINARY_TYPE__GUEST_VMLINUX:
> +	case DSO_BINARY_TYPE__GUEST_KMODULE:
> +	case DSO_BINARY_TYPE__GUEST_KMODULE_COMP:
> +	case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE:
> +	case DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE_COMP:
> +	case DSO_BINARY_TYPE__KCORE:
> +	case DSO_BINARY_TYPE__GUEST_KCORE:
> +	case DSO_BINARY_TYPE__BPF_PROG_INFO:
> +	case DSO_BINARY_TYPE__BPF_IMAGE:
> +	case DSO_BINARY_TYPE__OOL:
> +	case DSO_BINARY_TYPE__JAVA_JIT:

I think some of them can be possible in recorded data.  But let's go
simple with EM_HOST as we haven't supported cross-arch trace.

> +		return EM_HOST;
> +	case DSO_BINARY_TYPE__DEBUGLINK:
> +	case DSO_BINARY_TYPE__BUILD_ID_CACHE:
> +	case DSO_BINARY_TYPE__BUILD_ID_CACHE_DEBUGINFO:
> +	case DSO_BINARY_TYPE__SYSTEM_PATH_DSO:
> +	case DSO_BINARY_TYPE__OPENEMBEDDED_DEBUGINFO:
> +	case DSO_BINARY_TYPE__FEDORA_DEBUGINFO:
> +	case DSO_BINARY_TYPE__UBUNTU_DEBUGINFO:
> +	case DSO_BINARY_TYPE__MIXEDUP_UBUNTU_DEBUGINFO:
> +	case DSO_BINARY_TYPE__BUILDID_DEBUGINFO:
> +		break;
> +	case DSO_BINARY_TYPE__NOT_FOUND:
> +	default:
> +		return EM_NONE;
> +	}
> +
> +	pthread_mutex_lock(&dso__data_open_lock);

Hmm.. I'm afraid it'd slow down perf trace a bit more.  It sees
occasional LOST events.  But it may be ok as it's cached in thread later.

> +
> +	/*
> +	 * dso__data(dso)->fd might be closed if other thread opened another
> +	 * file (dso) due to open file limit (RLIMIT_NOFILE).
> +	 */
> +	try_to_open_dso(dso, machine);
> +	fd = dso__data(dso)->fd;
> +	if (fd >= 0) {
> +		_Static_assert(offsetof(Elf32_Ehdr, e_machine) == 18, "Unexpected offset");
> +		_Static_assert(offsetof(Elf64_Ehdr, e_machine) == 18, "Unexpected offset");
> +		if (pread(fd, &e_machine, sizeof(e_machine), 18) != sizeof(e_machine))

I think it needs to check the endianess and swap the data.

Thanks,
Namhyung

> +			e_machine = EM_NONE;
> +	}
> +	pthread_mutex_unlock(&dso__data_open_lock);
> +	return e_machine;
> +}
> +
>  /**
>   * dso__data_read_addr - Read data from dso address
>   * @dso: dso object
> diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h
> index f3ca2a5e7670..ba9b83db061a 100644
> --- a/tools/perf/util/dso.h
> +++ b/tools/perf/util/dso.h
> @@ -818,6 +818,7 @@ int dso__data_file_size(struct dso *dso, struct machine *machine);
>  off_t dso__data_size(struct dso *dso, struct machine *machine);
>  ssize_t dso__data_read_offset(struct dso *dso, struct machine *machine,
>  			      u64 offset, u8 *data, ssize_t size);
> +uint16_t dso__e_machine(struct dso *dso, struct machine *machine);
>  ssize_t dso__data_read_addr(struct dso *dso, struct map *map,
>  			    struct machine *machine, u64 addr,
>  			    u8 *data, ssize_t size);
> -- 
> 2.48.1.711.g2feabab25a-goog
> 