[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <52AE7086.2030708@gmail.com>
Date: Sun, 15 Dec 2013 20:16:22 -0700
From: David Ahern <dsahern@...il.com>
To: Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
Arnaldo Carvalho de Melo <acme@...stprotocols.net>
CC: Ingo Molnar <mingo@...hat.com>, linux-kernel@...r.kernel.org,
Frederic Weisbecker <fweisbec@...il.com>,
Jiri Olsa <jolsa@...hat.com>, Mike Galbraith <efault@....de>,
Namhyung Kim <namhyung@...il.com>,
Paul Mackerras <paulus@...ba.org>,
Stephane Eranian <eranian@...gle.com>,
Andi Kleen <ak@...ux.intel.com>,
Adrian Hunter <adrian.hunter@...el.com>
Subject: Re: [PATCH v0 07/71] perf tools: Record whether a dso is 64-bit
On 12/11/13, 5:36 AM, Alexander Shishkin wrote:
> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
> index a0c7c59..80817ec 100644
> --- a/tools/perf/util/dso.c
> +++ b/tools/perf/util/dso.c
> @@ -446,6 +446,7 @@ struct dso *dso__new(const char *name)
> dso->cache = RB_ROOT;
> dso->symtab_type = DSO_BINARY_TYPE__NOT_FOUND;
> dso->data_type = DSO_BINARY_TYPE__NOT_FOUND;
> + dso->is_64_bit = (sizeof(void *) == 8);
> dso->loaded = 0;
> dso->rel = 0;
> dso->sorted_by_name = 0;
> diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h
> index 384f2d9..62680e1 100644
> --- a/tools/perf/util/dso.h
> +++ b/tools/perf/util/dso.h
> @@ -91,6 +91,7 @@ struct dso {
> u8 annotate_warned:1;
> u8 sname_alloc:1;
> u8 lname_alloc:1;
> + u8 is_64_bit:1;
> u8 sorted_by_name;
> u8 loaded;
> u8 rel;
> diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
> index eed0b96..a0fc81b 100644
> --- a/tools/perf/util/symbol-elf.c
> +++ b/tools/perf/util/symbol-elf.c
> @@ -595,6 +595,8 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
> goto out_elf_end;
> }
>
> + ss->is_64_bit = (gelf_getclass(elf) == ELFCLASS64);
> +
> ss->symtab = elf_section_by_name(elf, &ehdr, &ss->symshdr, ".symtab",
> NULL);
> if (ss->symshdr.sh_type != SHT_SYMTAB)
> @@ -694,6 +696,7 @@ int dso__load_sym(struct dso *dso, struct map *map,
> bool remap_kernel = false, adjust_kernel_syms = false;
>
> dso->symtab_type = syms_ss->type;
> + dso->is_64_bit = syms_ss->is_64_bit;
> dso->rel = syms_ss->ehdr.e_type == ET_REL;
>
> /*
> diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c
> index ac7070a..b9d1119 100644
> --- a/tools/perf/util/symbol-minimal.c
> +++ b/tools/perf/util/symbol-minimal.c
> @@ -1,3 +1,4 @@
> +#include "util.h"
> #include "symbol.h"
>
> #include <stdio.h>
> @@ -287,6 +288,23 @@ int dso__synthesize_plt_symbols(struct dso *dso __maybe_unused,
> return 0;
> }
>
> +static int fd__is_64_bit(int fd)
> +{
> + u8 e_ident[EI_NIDENT];
> +
> + if (lseek(fd, 0, SEEK_SET))
> + return -1;
> +
> + if (readn(fd, e_ident, sizeof(e_ident)) != sizeof(e_ident))
> + return -1;
> +
> + if (memcmp(e_ident, ELFMAG, SELFMAG) ||
> + e_ident[EI_VERSION] != EV_CURRENT)
> + return -1;
> +
> + return e_ident[EI_CLASS] == ELFCLASS64;
> +}
> +
> int dso__load_sym(struct dso *dso, struct map *map __maybe_unused,
> struct symsrc *ss,
> struct symsrc *runtime_ss __maybe_unused,
> @@ -294,6 +312,11 @@ int dso__load_sym(struct dso *dso, struct map *map __maybe_unused,
> int kmodule __maybe_unused)
> {
> unsigned char *build_id[BUILD_ID_SIZE];
> + int ret;
> +
> + ret = fd__is_64_bit(ss->fd);
> + if (ret >= 0)
> + dso->is_64_bit = ret;
>
> if (filename__read_build_id(ss->name, build_id, BUILD_ID_SIZE) > 0) {
> dso__set_build_id(dso, build_id);
Here's what is wrong with this API: you are determining DSO bitness at
symbol load time, not when the DSO is created and added to the maps.
As I pointed out in a prior comment you are initializing dso->is_64_bit
to perf's bitness when the dso object is created (dso__new) but the
value is not correctly set until dso__load time. Some tools (perf-trace)
never load symbols, so that value is always wrong in the sense that its
value has no correlation to the dso object.
David
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists