public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Anton Blanchard <anton@samba.org>
To: Eric B Munson <ebmunson@us.ibm.com>
Cc: mingo@elte.hu, a.p.zijlstra@chello.nl, paulus@samba.org,
	acme@redhat.com, linux-kernel@vger.kernel.org, amodra@gmail.com
Subject: Re: [PATCH] perf lookup of ppc64 symbols
Date: Fri, 18 Jun 2010 00:46:08 +1000	[thread overview]
Message-ID: <20100617144608.GS28295@kryten> (raw)
In-Reply-To: <1276523793-15422-1-git-send-email-ebmunson@us.ibm.com>


Hi,

> Currently symbol resolution does not work for 64-bit programs
> using perf.  For this to work perf needs to lookup symbol names in
> the OPD.  This patch adds this functionality.  The OPD data is
> used to find symbol names when available.

To clarify, this is an issue with 64bit architectures that use function
descriptors (eg ppc64). The problem is that a symbol doesn't point to a text
address, it points to a data area that contains (amongst other things) a
pointer to the text address.

We look for a section called ".opd" which is the function descriptor
area. To create the full symbol table, when we see a symbol in the function
descriptor section we load the first pointer and use that as the text
address. The only tricky part here is working out which section the
text address is in. Right now we have a loop which is sub optimal:

> +	while ((sec = elf_nextscn(elf, sec)) != NULL) {
> +		gelf_getshdr(sec, &shdr);
> +
> +		if ((addr >= shdr.sh_addr) &&
> +		    (addr < (shdr.sh_addr + shdr.sh_size)))
> +			return cnt;
> +
> +		++cnt;
> +	}

Anton

> Signed-off-by: Anton Blanchard <anton@samba.org>
> Signed-off-by: Eric B Munson <ebmunson@us.ibm.com>
> ---
>  tools/perf/util/symbol.c |   37 ++++++++++++++++++++++++++++++++++---
>  1 files changed, 34 insertions(+), 3 deletions(-)
> 
> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> index b63e571..971d0a0 100644
> --- a/tools/perf/util/symbol.c
> +++ b/tools/perf/util/symbol.c
> @@ -933,6 +933,25 @@ static bool elf_sec__is_a(GElf_Shdr *self, Elf_Data *secstrs, enum map_type type
>  	}
>  }
>  
> +static size_t elf_addr_to_index(Elf *elf, GElf_Addr addr)
> +{
> +	Elf_Scn *sec = NULL;
> +	GElf_Shdr shdr;
> +	size_t cnt = 1;
> +
> +	while ((sec = elf_nextscn(elf, sec)) != NULL) {
> +		gelf_getshdr(sec, &shdr);
> +
> +		if ((addr >= shdr.sh_addr) &&
> +		    (addr < (shdr.sh_addr + shdr.sh_size)))
> +			return cnt;
> +
> +		++cnt;
> +	}
> +
> +	return -1;
> +}
> +
>  static int dso__load_sym(struct dso *self, struct map *map, const char *name,
>  			 int fd, symbol_filter_t filter, int kmodule)
>  {
> @@ -944,12 +963,13 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name,
>  	int err = -1;
>  	uint32_t idx;
>  	GElf_Ehdr ehdr;
> -	GElf_Shdr shdr;
> -	Elf_Data *syms;
> +	GElf_Shdr shdr, opdshdr;
> +	Elf_Data *syms, *opddata = NULL;
>  	GElf_Sym sym;
> -	Elf_Scn *sec, *sec_strndx;
> +	Elf_Scn *sec, *sec_strndx, *opdsec;
>  	Elf *elf;
>  	int nr = 0;
> +	size_t opdidx = 0;
>  
>  	elf = elf_begin(fd, PERF_ELF_C_READ_MMAP, NULL);
>  	if (elf == NULL) {
> @@ -969,6 +989,10 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name,
>  			goto out_elf_end;
>  	}
>  
> +	opdsec = elf_section_by_name(elf, &ehdr, &opdshdr, ".opd", &opdidx);
> +	if (opdsec)
> +		opddata = elf_rawdata(opdsec, NULL);
> +
>  	syms = elf_getdata(sec, NULL);
>  	if (syms == NULL)
>  		goto out_elf_end;
> @@ -1013,6 +1037,13 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name,
>  		if (!is_label && !elf_sym__is_a(&sym, map->type))
>  			continue;
>  
> +		if (opdsec && sym.st_shndx == opdidx) {
> +			u32 offset = sym.st_value - opdshdr.sh_addr;
> +			u64 *opd = opddata->d_buf + offset;
> +			sym.st_value = *opd;
> +			sym.st_shndx = elf_addr_to_index(elf, sym.st_value);
> +		}
> +
>  		sec = elf_getscn(elf, sym.st_shndx);
>  		if (!sec)
>  			goto out_elf_end;
> -- 
> 1.7.0.4

  reply	other threads:[~2010-06-17 14:49 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-06-14 13:56 [PATCH] perf lookup of ppc64 symbols Eric B Munson
2010-06-17 14:46 ` Anton Blanchard [this message]
2010-06-17 15:17   ` Arnaldo Carvalho de Melo
2010-06-17 15:19     ` Arnaldo Carvalho de Melo
2010-06-18 10:16 ` [tip:perf/core] perf symbols: Function descriptor symbol lookup tip-bot for Eric B Munson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100617144608.GS28295@kryten \
    --to=anton@samba.org \
    --cc=a.p.zijlstra@chello.nl \
    --cc=acme@redhat.com \
    --cc=amodra@gmail.com \
    --cc=ebmunson@us.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=paulus@samba.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox