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] [day] [month] [year] [list]
Date:	Fri, 8 Aug 2014 19:34:59 +0000
From:	Josef Bacik <jbacik@...com>
To:	Steven Rostedt <rostedt@...dmis.org>
CC:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] trace-cmd: make sure we have a pending pid in trace-hist

Oops crap that's the second time I've done that this week.  When I get to my computer I'll send the actual fix by itself, sorry about that,

Josef

Steven Rostedt <rostedt@...dmis.org> wrote:


On Wed, 30 Jul 2014 16:54:45 -0400
Josef Bacik <jbacik@...com> wrote:

> I have a trace.dat file where we lost some events and so we didn't have the
> corresponding event to go with the stack trace.  This made trace-cmd hist
> segfault because it unconditionally pushes the pending_pid's function onto the
> stack trace.  In this case pending_pid < 0, which means we didn't have a pending
> stack trace, so we got garbage in the call chain and hilarity ensued.  Fix this
> by only pushing the pending call chain if there actually is one, and also don't
> reset the current call chain if there isn't actually one there either.  This
> fixed my segfault.  Thanks,
>
> Signed-off-by: Josef Bacik <jbacik@...com>
> ---
>  Makefile      |  9 +++++----
>  trace-cmd.c   |  3 +++
>  trace-hist.c  | 14 +++++++++-----
>  trace-local.h |  2 ++
>  trace-usage.c |  5 +++++
>  5 files changed, 24 insertions(+), 9 deletions(-)
>
> diff --git a/Makefile b/Makefile
> index cbe0eb9..0a46125 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -81,7 +81,7 @@ ifndef NO_PYTHON
>  PYTHON               := ctracecmd.so
>  PYTHON_GUI   := ctracecmd.so ctracecmdgui.so
>
> -PYTHON_VERS ?= python
> +PYTHON_VERS ?= python-2.7
>
>  # Can build python?
>  ifeq ($(shell sh -c "pkg-config --cflags $(PYTHON_VERS) > /dev/null 2>&1 && echo y"), y)
> @@ -164,6 +164,7 @@ LIBS = -L. -ltracecmd -ldl
>  LIB_FILE = libtracecmd.a
>
>  PACKAGES= gtk+-2.0 libxml-2.0 gthread-2.0
> +NOGUI_PACKAGES= glib-2.0
>
>  ifndef BUILDGUI
>   BUILDGUI = 0
> @@ -194,8 +195,8 @@ N                 = @/bin/true ||
>  CONFIG_FLAGS += $(HELP_DIR_SQ)
>  else
>
> -CONFIG_INCLUDES =
> -CONFIG_LIBS  =
> +CONFIG_INCLUDES = $(shell pkg-config --cflags $(NOGUI_PACKAGES))
> +CONFIG_LIBS  = $(shell pkg-config --libs $(NOGUI_PACKAGES))
>  CONFIG_FLAGS =
>
>  VERSION              = $(TC_VERSION)
> @@ -307,7 +308,7 @@ $(obj)/%.o: $(src)/%.c
>  TRACE_GUI_OBJS = trace-filter.o trace-compat.o trace-hash.o trace-dialog.o \
>               trace-xml.o
>  TRACE_CMD_OBJS = trace-cmd.o trace-record.o trace-read.o trace-split.o trace-listen.o \
> -      trace-stack.o trace-hist.o trace-mem.o trace-snapshot.o
> +      trace-stack.o trace-hist.o trace-mem.o trace-snapshot.o trace-latency.o
>  TRACE_VIEW_OBJS = trace-view.o trace-view-store.o
>  TRACE_GRAPH_OBJS = trace-graph.o trace-plot.o trace-plot-cpu.o trace-plot-task.o
>  TRACE_VIEW_MAIN_OBJS = trace-view-main.o $(TRACE_VIEW_OBJS) $(TRACE_GUI_OBJS)
> diff --git a/trace-cmd.c b/trace-cmd.c
> index ebf9c7a..a8ecad6 100644
> --- a/trace-cmd.c
> +++ b/trace-cmd.c
> @@ -432,6 +432,9 @@ int main (int argc, char **argv)
>       } else if (strcmp(argv[1], "stack") == 0) {
>               trace_stack(argc, argv);
>               exit(0);
> +     } else if (strcmp(argv[1], "latency") == 0) {
> +             trace_latency(argc, argv);
> +             exit(0);
>       } else if (strcmp(argv[1], "check-events") == 0) {
>               char *tracing;
>               int ret;
> diff --git a/trace-hist.c b/trace-hist.c
> index 73820de..574549e 100644
> --- a/trace-hist.c
> +++ b/trace-hist.c
> @@ -453,9 +453,11 @@ process_kernel_stack(struct pevent *pevent, struct pevent_record *record)
>               }
>       } else {
>               /* function stack trace? */
> -             copy_stack_to_pending(current_pid);
> -             free(ips);
> -             reset_stack();
> +             if (current_pid >= 0) {
> +                     copy_stack_to_pending(current_pid);
> +                     free(ips);
> +                     reset_stack();
> +             }
>       }
>
>       current_pid = pid;
> @@ -482,8 +484,10 @@ process_kernel_stack(struct pevent *pevent, struct pevent_record *record)
>                       push_stack_func(func);
>       }
>
> -     push_stack_func(pending_ips[pending_ips_idx - 1]);
> -     reset_pending_stack();
> +     if (pending_pid >= 0) {
> +             push_stack_func(pending_ips[pending_ips_idx - 1]);
> +             reset_pending_stack();
> +     }
>       save_call_chain(current_pid, ips, ips_idx, 1);
>       if (do_restore)
>               restore_stack(current_pid);

This blob looks to be what the change log suggest. I think you added a
bunch more than the fix.

Can you resend please with just the change from the change log. If you
want to add more, please send separate patches. This shouldn't be like
congress adding earmarks now, is it ;-)

Thanks!

-- Steve


> diff --git a/trace-local.h b/trace-local.h
> index 3c82c2f..1d6d69b 100644
> --- a/trace-local.h
> +++ b/trace-local.h
> @@ -57,6 +57,8 @@ void trace_snapshot(int argc, char **argv);
>
>  void trace_mem(int argc, char **argv);
>
> +void trace_latency(int argc, char **argv);
> +
>  /* --- instance manipulation --- */
>
>  struct func_list {
> diff --git a/trace-usage.c b/trace-usage.c
> index 0dec87e..ef73551 100644
> --- a/trace-usage.c
> +++ b/trace-usage.c
> @@ -231,6 +231,11 @@ static struct usage_help usage_help[] = {
>               "          -N do not load any plugins\n"
>       },
>       {
> +             "latency",
> +             "parse latency events",
> +             ""
> +     },
> +     {
>               NULL, NULL, NULL
>       }
>  };

--
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

Powered by Openwall GNU/*/Linux Powered by OpenVZ