public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Joel Brobecker <brobecker@adacore.com>
To: Jan Kratochvil <jan.kratochvil@redhat.com>
Cc: gdb-patches@sourceware.org, Phil Muldoon <pmuldoon@redhat.com>
Subject: Re: [patchv2] compile: Fix crash on cv-qualified self-reference
Date: Tue, 30 Jun 2015 19:57:00 -0000	[thread overview]
Message-ID: <20150630195707.GA10987@adacore.com> (raw)
In-Reply-To: <20150516132555.GB17266@host1.jankratochvil.net>

Would anyone be in a better position to review this patch?
It's one patch identified as a blocking issue before we create
the GDB 7.10 branch. It looks reasonable to me, but I can only
do a superficial review...

Thanks!

On Sat, May 16, 2015 at 03:25:55PM +0200, Jan Kratochvil wrote:
> On Sat, 18 Apr 2015 19:28:43 +0200, Jan Kratochvil wrote:
> Hi,
> 
> with this modified testcase GDB would:
> 
> compile code struct_object.selffield = &struct_object
> ./compile/compile-c-types.c:83: internal-error: insert_type: Assertion `add == NULL || add->gcc_type == gcc_type' failed.
> A problem internal to GDB has been detected,
> further debugging may prove unreliable.
> Quit this debugging session? (y or n) FAIL: gdb.compile/compile.exp: compile code struct_object.selffield = &struct_object (GDB internal error)
> 
> While the insert_type() assertion looks unclear trying to fix it one ends up
> with either GCC crash
> 	[gcc libcc1] build_qualified_type for self-referencing/incomplete types
> 	https://gcc.gnu.org/ml/gcc/2015-04/msg00108.html
> 	c_incomplete_type_error()
> or after fixing up the GCC type for proper error reporting one gets:
> 	gdb command line:1:1: error: invalid use of incomplete typedef ‘sv’
> which is the real culprit of this bug as explained in this patch.
> 
> This patch is related to the XFAIL introduced by
> 	[PATCH v3 5/9] compile: Use -Wall, not -w
> 	https://sourceware.org/ml/gdb-patches/2015-04/msg00429.html
> as for proper -Wall happiness the 'volatile' qualifier needs to be added there
> - but adding the qualifier has caused this crash.
> 
> No regressions on {x86_64,x86_64-m32,i686}-fedora23pre-linux-gnu.
> 
> 
> Thanks,
> Jan

> gdb/ChangeLog
> 2015-05-16  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	compile: Fix crash on cv-qualified self-reference.
> 	* compile/compile-c-types.c (convert_struct_or_union): Apply
> 	build_qualified_type.
> 	(convert_type_basic): Do not apply build_qualified_type for
> 	TYPE_CODE_STRUCT and TYPE_CODE_UNION.
> 
> gdb/testsuite/ChangeLog
> 2015-05-16  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	compile: Fix crash on cv-qualified self-reference.
> 	* gdb.compile/compile.c (struct struct_type): Add volatile for
> 	selffield.
> 	* gdb.compile/compile.exp
> 	(compile code struct_object.selffield = &struct_object): Remove XFAIL.
> 
> diff --git a/gdb/compile/compile-c-types.c b/gdb/compile/compile-c-types.c
> index 2b521bc..420f61d 100644
> --- a/gdb/compile/compile-c-types.c
> +++ b/gdb/compile/compile-c-types.c
> @@ -166,9 +166,13 @@ convert_struct_or_union (struct compile_c_instance *context, struct type *type)
>  {
>    int i;
>    gcc_type result;
> +  int quals;
>  
>    /* First we create the resulting type and enter it into our hash
> -     table.  This lets recursive types work.  */
> +     table.  This lets recursive types work.  We have to create gcc_type
> +     already with its qualifiers to prevent recursively calling
> +     build_qualified_type for unfinished TYPE as build_qualified_type
> +     creates a copy of the type, remaining unfinished forever.  */
>    if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
>      result = C_CTX (context)->c_ops->build_record_type (C_CTX (context));
>    else
> @@ -176,6 +180,15 @@ convert_struct_or_union (struct compile_c_instance *context, struct type *type)
>        gdb_assert (TYPE_CODE (type) == TYPE_CODE_UNION);
>        result = C_CTX (context)->c_ops->build_union_type (C_CTX (context));
>      }
> +  quals = 0;
> +  if (TYPE_CONST (type))
> +    quals |= GCC_QUALIFIER_CONST;
> +  if (TYPE_VOLATILE (type))
> +    quals |= GCC_QUALIFIER_VOLATILE;
> +  if (TYPE_RESTRICT (type))
> +    quals |= GCC_QUALIFIER_RESTRICT;
> +  result = C_CTX (context)->c_ops->build_qualified_type (C_CTX (context),
> +							 result, quals);
>    insert_type (context, type, result);
>  
>    for (i = 0; i < TYPE_NFIELDS (type); ++i)
> @@ -329,10 +342,13 @@ static gcc_type
>  convert_type_basic (struct compile_c_instance *context, struct type *type)
>  {
>    /* If we are converting a qualified type, first convert the
> -     unqualified type and then apply the qualifiers.  */
> +     unqualified type and then apply the qualifiers, except for the
> +     types handling qualifiers on their own.  */
>    if ((TYPE_INSTANCE_FLAGS (type) & (TYPE_INSTANCE_FLAG_CONST
>  				     | TYPE_INSTANCE_FLAG_VOLATILE
> -				     | TYPE_INSTANCE_FLAG_RESTRICT)) != 0)
> +				     | TYPE_INSTANCE_FLAG_RESTRICT)) != 0
> +      && (TYPE_CODE (type) != TYPE_CODE_STRUCT
> +	  && TYPE_CODE (type) != TYPE_CODE_UNION))
>      return convert_qualified (context, type);
>  
>    switch (TYPE_CODE (type))
> diff --git a/gdb/testsuite/gdb.compile/compile.c b/gdb/testsuite/gdb.compile/compile.c
> index 3d5f20a..41ff087 100644
> --- a/gdb/testsuite/gdb.compile/compile.c
> +++ b/gdb/testsuite/gdb.compile/compile.c
> @@ -42,7 +42,7 @@ struct struct_type {
>    float floatfield;
>    double doublefield;
>    const union union_type *ptrfield;
> -  struct struct_type *selffield;
> +  volatile struct struct_type *selffield;
>    int arrayfield[5];
>    _Complex double complexfield;
>    _Bool boolfield;
> diff --git a/gdb/testsuite/gdb.compile/compile.exp b/gdb/testsuite/gdb.compile/compile.exp
> index 07276bd..9668be8 100644
> --- a/gdb/testsuite/gdb.compile/compile.exp
> +++ b/gdb/testsuite/gdb.compile/compile.exp
> @@ -189,15 +189,7 @@ gdb_test "p localvar" " = 1"
>  # Test setting fields and also many different types.
>  #
>  
> -set test "compile code struct_object.selffield = &struct_object"
> -gdb_test_multiple $test $test {
> -    -re "^$test\r\n$gdb_prompt $" {
> -	pass "$test"
> -    }
> -    -re "gdb command line:1:25: warning: assignment discards 'volatile' qualifier from pointer target type \\\[-Wdiscarded-qualifiers\\\]\r\n$gdb_prompt $" {
> -	xfail "$test (PR compile/18202)"
> -    }
> -}
> +gdb_test_no_output "compile code struct_object.selffield = &struct_object"
>  gdb_test "print struct_object.selffield == &struct_object" " = 1"
>  
>  gdb_test_no_output "compile code struct_object.charfield = 1"


-- 
Joel

  reply	other threads:[~2015-06-30 19:57 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-04-18 17:28 [patch] " Jan Kratochvil
2015-05-16 13:26 ` [patchv2] " Jan Kratochvil
2015-06-30 19:57   ` Joel Brobecker [this message]
2015-07-01  9:44     ` Phil Muldoon
2015-07-01 10:32   ` Pedro Alves
2015-07-01 11:21   ` Yao Qi
2015-07-01 13:24     ` Jan Kratochvil
2015-07-01 13:54       ` Pedro Alves
2015-07-01 14:10         ` Jan Kratochvil
2015-07-01 14:59           ` Pedro Alves
2015-07-01 15:12             ` Jan Kratochvil
2015-07-01 15:24               ` Pedro Alves
2015-07-01 15:29                 ` Jan Kratochvil
2015-07-01 15:35                   ` Pedro Alves
2015-07-01 16:07                     ` Jan Kratochvil
2015-07-01 16:16                       ` Pedro Alves
2015-07-01 14:06       ` Yao Qi
2015-07-02 12:34       ` Jan Kratochvil
2015-07-04 17:11         ` [patchv3] " Jan Kratochvil
2015-07-08  9:29           ` Pedro Alves
2015-07-08 12:51             ` [commit] " Jan Kratochvil

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=20150630195707.GA10987@adacore.com \
    --to=brobecker@adacore.com \
    --cc=gdb-patches@sourceware.org \
    --cc=jan.kratochvil@redhat.com \
    --cc=pmuldoon@redhat.com \
    /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).