public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Simon Marchi <simark@simark.ca>
To: Jan Vrany <jan.vrany@fit.cvut.cz>, gdb-patches@sourceware.org
Subject: Re: [PATCH] Fix segfault when invoking -var-info-path-expression on a dynamic varobj
Date: Tue, 19 Jun 2018 00:42:00 -0000	[thread overview]
Message-ID: <fb201059-787c-7a78-8214-c2144e0c224a@simark.ca> (raw)
In-Reply-To: <20180604131503.31945-1-jan.vrany@fit.cvut.cz>

On 2018-06-04 09:15 AM, Jan Vrany wrote:
> Invoking -var-info-path-expression on a dynamic varobj lead either in wrong
> (nonsense) result or to a segmentation fault in cplus_describe_child().
> This was caused by the fact that varobj_get_path_expr() called
> cplus_path_expr_of_child() ignoring the fact the parent of the variable
> is dynamic. Then, cplus_describe_child() accessed the underlaying C type
> members by index, causing (i) either wrong (nonsense) expression being
> returned (since dynamic child may be completely arbibtrary value)
> or (ii) segmentation fault (in case the index higher than number of
> underlaying C type members.
> 
> This fixes the problem by checking whether a varobj is a child of a dynamic
> varobj and, if so, reporting an error to MI consumer as described in
> the documentation.

Hi Jan,

Thanks for the patch.  I think we should also handle the indirect children
of dynamic objects.  When trying this test program:

#include <vector>
struct foo { int a; };
int main() {
    std::vector<foo> vec;
    vec.push_back({ 1 });
    return 0;
}

With the following MI commands:

-enable-pretty-printing
b 6
r
-var-create - * vec
-var-list-children var1
-var-list-children var1.[0]
-var-list-children var1.[0].public
-var-info-path-expression var1.[0].public.a

I get this:

~"/home/simark/src/binutils-gdb/gdb/varobj.c:969: internal-error: const char* varobj_get_path_expr(const varobj*): Assertion `!varobj_is_dynamic_p(var->parent)' failed..."

I noted some small formatting nits below.

> gdb/ChangeLog:
> 
> 	* mi/mi-cmd-var.c: Report an error if varobj is a child of dynamic
> 	varobj as stated in documentation.
> 	* varobj.c: New assert.
> 
> gdb/testsuite/Changelog:
> 
> 	* gdb.pythonpy-mi-var-info-path-expression.c: New file.

You are missing a slash here (and in the ChangeLog file itself).

> 	* gdb.python/py-mi-var-info-path-expression.py: New file.
> 	* gdb.python/py-mi-var-info-path-expression.exp: New file.
> ---
>  gdb/ChangeLog                                 |  6 ++
>  gdb/mi/mi-cmd-var.c                           |  5 ++
>  gdb/testsuite/ChangeLog                       |  6 ++
>  .../py-mi-var-info-path-expression.c          | 40 ++++++++++
>  .../py-mi-var-info-path-expression.exp        | 77 +++++++++++++++++++
>  .../py-mi-var-info-path-expression.py         | 45 +++++++++++
>  gdb/varobj.c                                  |  6 +-
>  7 files changed, 184 insertions(+), 1 deletion(-)
>  create mode 100644 gdb/testsuite/gdb.python/py-mi-var-info-path-expression.c
>  create mode 100644 gdb/testsuite/gdb.python/py-mi-var-info-path-expression.exp
>  create mode 100644 gdb/testsuite/gdb.python/py-mi-var-info-path-expression.py
> 
> diff --git a/gdb/ChangeLog b/gdb/ChangeLog
> index a292f4c056..fc5edb9a9e 100644
> --- a/gdb/ChangeLog
> +++ b/gdb/ChangeLog
> @@ -1,3 +1,9 @@
> +2018-06-04  Jan Vrany  <jan.vrany@fit.cvut.cz>
> +
> +	* mi/mi-cmd-var.c: Report an error if varobj is a child of dynamic
> +	varobj as stated in documentation.
> +	* varobj.c: New assert.
> +
>  2018-06-04  Pedro Alves  <palves@redhat.com>
>  
>  	* darwin-nat.c (darwin_ops): Delete.
> diff --git a/gdb/mi/mi-cmd-var.c b/gdb/mi/mi-cmd-var.c
> index 38c59c7e66..f278281023 100644
> --- a/gdb/mi/mi-cmd-var.c
> +++ b/gdb/mi/mi-cmd-var.c
> @@ -438,6 +438,11 @@ mi_cmd_var_info_path_expression (const char *command, char **argv, int argc)
>  
>    /* Get varobj handle, if a valid var obj name was specified.  */
>    var = varobj_get_handle (argv[0]);
> +
> +  /* -var-info-path-expression is currently not valid for children of
> +     a dynamic varobj  */

Finish the sentence with a period.

> diff --git a/gdb/varobj.c b/gdb/varobj.c
> index a0df485ae9..87bfd44549 100644
> --- a/gdb/varobj.c
> +++ b/gdb/varobj.c
> @@ -962,9 +962,13 @@ varobj_get_path_expr (const struct varobj *var)
>        /* For root varobjs, we initialize path_expr
>  	 when creating varobj, so here it should be
>  	 child varobj.  */
> -      struct varobj *mutable_var = (struct varobj *) var;
>        gdb_assert (!is_root_p (var));
>  
> +      /* Computation of full rooted expression for children of dynamic
> +         varobjs is not supported. */

Two spaces after the final period (before the */).

> +      gdb_assert (!varobj_is_dynamic_p(var->parent));

Space after varobj_is_dynamic_p.

Thanks,

Simon

  parent reply	other threads:[~2018-06-19  0:42 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-06-04 13:15 Jan Vrany
2018-06-18  7:14 ` [PING] " Jan Vrany
2018-06-19  0:42 ` Simon Marchi [this message]
2018-06-19  8:59   ` Jan Vrany

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=fb201059-787c-7a78-8214-c2144e0c224a@simark.ca \
    --to=simark@simark.ca \
    --cc=gdb-patches@sourceware.org \
    --cc=jan.vrany@fit.cvut.cz \
    /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 read-only IMAP folder(s) and NNTP newsgroup(s).