linux-kernel - Re: [RFC] perf callchain: Compare dsos (as well) for CCKEY

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20171004130834.GD23759@krava>
Date:   Wed, 4 Oct 2017 15:08:34 +0200
From:   Jiri Olsa <jolsa@...hat.com>
To:     Ravi Bangoria <ravi.bangoria@...ux.vnet.ibm.com>
Cc:     pozdneyev@...il.com, acme@...nel.org, linux-kernel@...r.kernel.org,
        linux-perf-users@...r.kernel.org, peterz@...radead.org,
        mingo@...hat.com, alexander.shishkin@...ux.intel.com,
        yao.jin@...ux.intel.com, ak@...ux.intel.com, jolsa@...nel.org,
        kjlx@...pleofstupid.com, milian.wolff@...b.com,
        zhangmengting@...wei.com
Subject: Re: [RFC] perf callchain: Compare dsos (as well) for CCKEY_FUNCTION

On Wed, Oct 04, 2017 at 11:43:08AM +0530, Ravi Bangoria wrote:
> Two functions from different binaries can have same start
> address. Thus, comparing only start address in match_chain()
> leads to inconsistent callchains. Fix this by adding a check
> for dsos as well.
> 
> Ex, https://www.spinics.net/lists/linux-perf-users/msg04067.html
> 
> Reported-by: Alexander Pozdneev <pozdneyev@...il.com>
> Signed-off-by: Ravi Bangoria <ravi.bangoria@...ux.vnet.ibm.com>
> ---
>  tools/perf/util/callchain.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c
> index 510b513..6d5a483 100644
> --- a/tools/perf/util/callchain.c
> +++ b/tools/perf/util/callchain.c
> @@ -678,6 +678,9 @@ static enum match_result match_chain(struct callchain_cursor_node *node,
>  {
>  	struct symbol *sym = node->sym;
>  	u64 left, right;
> +	struct dso *left_dso = NULL;
> +	struct dso *right_dso = NULL;
> +
>  
>  	if (callchain_param.key == CCKEY_SRCLINE) {
>  		enum match_result match = match_chain_srcline(node, cnode);
> @@ -689,12 +692,16 @@ static enum match_result match_chain(struct callchain_cursor_node *node,
>  	if (cnode->ms.sym && sym && callchain_param.key == CCKEY_FUNCTION) {
>  		left = cnode->ms.sym->start;
>  		right = sym->start;
> +		if (cnode->ms.map && node->map) {
> +			left_dso = cnode->ms.map->dso;
> +			right_dso = node->map->dso;

makes sense.. but why not to get those maps separately?

	if (cnode->ms.map)
		left_dso = cnode->ms.map->dso;
	if (node->map) {
		right_dso = node->map->dso;

I'd think that if one is missing, it's most likely different
map/dso and you want to fail the == check

jirka

> +		}
>  	} else {
>  		left = cnode->ip;
>  		right = node->ip;
>  	}
>  
> -	if (left == right) {
> +	if (left == right && left_dso == right_dso) {
>  		if (node->branch) {
>  			cnode->branch_count++;
>  
> -- 
> 1.8.3.1
>