From: David Ahern <dsahern@gmail.com>
To: Stephane Eranian <eranian@google.com>
Cc: linux-kernel@vger.kernel.org, acme@infradead.org,
peterz@infradead.org, mingo@elte.hu, fweisbec@gmail.com
Subject: Re: [PATCH] perf: fix handling of .debug cache for kernel symbols
Date: Thu, 01 Dec 2011 23:05:21 -0700 [thread overview]
Message-ID: <4ED86AA1.7070506@gmail.com> (raw)
In-Reply-To: <20111202044825.GA21751@quad>
On 12/01/2011 09:48 PM, Stephane Eranian wrote:
> There was a bug in the way perf was looking for
> kernel symbols when the .debug cache was not in $HOME.
By chance have you encountered problems with annotate in this scenario?
i.e., remote analysis with a copy of /proc/kallsyms from the target and
using the --vmlinux argument.
For example, 'perf report --kallsyms kallsyms --vmlinux vmlinux ...' and
then attempting to use the annotate option? I have a hit a few issues
with this path - and kernel modules with a symfs path -- that I have not
had time to debug yet.
David
>
> With the /etc/perfconfig or $HOME/.perfconfig files, it
> is possible to use an alternate location for the .debug
> buildid + binary cache.
>
> The problem is that the alternate location was ignored
> for kernel symbols, leading to partial profiles in case
> the host kernel was not the one used to collect the profile,
> e.g., with remote profiling.
>
> This patch fixes the problem by adding the buildi_dir prefix
> to the lookup path used in dso__load_kernel_sym().
>
> The change to util.h is necessary to avoid a redefinition
> of _GNU_SOURCE.
>
> Signed-off-by: Stephane Eranian <eranian@google.com>
> ---
>
> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> index 632b50c..8ace954 100644
> --- a/tools/perf/util/symbol.c
> +++ b/tools/perf/util/symbol.c
> @@ -16,6 +16,7 @@
> #include "debug.h"
> #include "symbol.h"
> #include "strlist.h"
> +#include "util.h"
>
> #include <libelf.h>
> #include <gelf.h>
> @@ -2082,8 +2083,8 @@ static int dso__load_kernel_sym(struct dso *dso, struct map *map,
> sbuild_id);
>
> if (asprintf(&kallsyms_allocated_filename,
> - "%s/.debug/[kernel.kallsyms]/%s",
> - getenv("HOME"), sbuild_id) == -1) {
> + "%s/[kernel.kallsyms]/%s",
> + buildid_dir, sbuild_id) == -1) {
> pr_err("Not enough memory for kallsyms file lookup\n");
> return -1;
> }
> diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
> index 0128906..4b90e93 100644
> --- a/tools/perf/util/util.h
> +++ b/tools/perf/util/util.h
> @@ -40,7 +40,9 @@
> #define decimal_length(x) ((int)(sizeof(x) * 2.56 + 0.5) + 1)
>
> #define _ALL_SOURCE 1
> +#ifndef _GNU_SOURCE
> #define _GNU_SOURCE 1
> +#endif
> #define _BSD_SOURCE 1
> #define HAS_BOOL
>
prev parent reply other threads:[~2011-12-02 6:05 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-12-02 4:48 [PATCH] perf: fix handling of .debug cache for kernel symbols Stephane Eranian
2011-12-02 6:05 ` David Ahern [this message]
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=4ED86AA1.7070506@gmail.com \
--to=dsahern@gmail.com \
--cc=acme@infradead.org \
--cc=eranian@google.com \
--cc=fweisbec@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=peterz@infradead.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;
as well as URLs for NNTP newsgroup(s).