public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Tom de Vries <tdevries@suse.de>
To: Pedro Alves <pedro@palves.net>, gdb-patches@sourceware.org
Subject: Re: [PATCH 2/2] [gdb] Fix rethrow exception slicing in insert_bp_location
Date: Tue, 25 Oct 2022 09:14:22 +0200	[thread overview]
Message-ID: <546f7fe4-7caa-4b78-bfd5-19480d60420e@suse.de> (raw)
In-Reply-To: <0203ad6e-426f-3e8f-fac8-1d2f0e8781a6@palves.net>

[-- Attachment #1: Type: text/plain, Size: 2449 bytes --]

On 10/24/22 18:51, Pedro Alves wrote:
> 
> 
> On 2022-10-24 5:43 p.m., Tom de Vries wrote:
>> On 10/24/22 18:36, Pedro Alves wrote:
>>> On 2022-10-24 9:49 a.m., Tom de Vries via Gdb-patches wrote:
>>>
>>>> +#define RETHROW_ON_TARGET_CLOSE_ERROR(E)                \
>>>> +  do                                    \
>>>> +    {                                    \
>>>> +      if ((E).reason != 0)                        \
>>>> +    {                                \
>>>> +      /* Can't set the breakpoint.  */                \
>>>> +                                    \
>>>> +      if ((E).error == TARGET_CLOSE_ERROR)                \
>>>> +        /* If the target has closed then it will have deleted any    \
>>>> +           breakpoints inserted within the target inferior, as a    \
>>>> +           result any further attempts to interact with the        \
>>>> +           breakpoint objects is not possible.  Just rethrow the    \
>>>> +           error.  Don't use E to rethrow, to prevent object    \
>>>> +           slicing of the exception.  */                \
>>>> +        throw;                            \
>>>> +    }                                \
>>>> +    } while (0)
>>>
>>> Is there a reason this is a macro instead of a function?
>>
>> yes, the throw without expression.
>>
>> Do you know of a way to do this using a function?
> 
> WDYM?  I believe it should Just Work.  E.g.:
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> #include <stdio.h>
> 
> struct excpt
> {
> };
> 
> struct excpt2 : excpt
> {
> };
> 
> void
> rethrow ()
> {
>    throw;
> }
> 
> int main ()
> {
>    try
>      {
>        try
> 	{
> 	  throw excpt2{};
> 	}
>        catch (const excpt &)
> 	{
> 	  rethrow ();
> 	}
>      }
>    catch (const excpt2 &)
>      {
>        printf ("caught excpt2\n");
>      }
>    return 0;
> }
> 
> $ g++ test_rethrow.cc -o test_rethrow
> $ ./test_rethrow
> caught excpt2
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Hmm, it does indeed, I didn't realize this was possible.

Fixed in attached patch.

Any further comments?

Thanks,
- Tom

[-- Attachment #2: 0001-gdb-Rewrite-RETHROW_ON_TARGET_CLOSE_ERROR-into-funct.patch --]
[-- Type: text/x-patch, Size: 3359 bytes --]

From 633dfb9645a281bd1d69b63d046d2ed8408cb0d3 Mon Sep 17 00:00:00 2001
From: Tom de Vries <tdevries@suse.de>
Date: Tue, 25 Oct 2022 08:17:46 +0200
Subject: [PATCH] [gdb] Rewrite RETHROW_ON_TARGET_CLOSE_ERROR into function

Recent commit b2829fcf9b5 ("[gdb] Fix rethrow exception slicing in
insert_bp_location") introduced macro RETHROW_ON_TARGET_CLOSE_ERROR.

I wrote this as a macro in order to have the rethrowing throw be part of the
same function as the catch, but as it turns out that not necessary.

Rewrite into a function.

Tested on x86_64-linux.
---
 gdb/breakpoint.c | 45 +++++++++++++++++++++++++--------------------
 1 file changed, 25 insertions(+), 20 deletions(-)

diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c
index a001e78cfb4..30826032360 100644
--- a/gdb/breakpoint.c
+++ b/gdb/breakpoint.c
@@ -2647,23 +2647,28 @@ breakpoint_kind (const struct bp_location *bl, CORE_ADDR *addr)
     return gdbarch_breakpoint_kind_from_pc (bl->gdbarch, addr);
 }
 
-#define RETHROW_ON_TARGET_CLOSE_ERROR(E)				\
-  do									\
-    {									\
-      if ((E).reason != 0)						\
-	{								\
-	  /* Can't set the breakpoint.  */				\
-									\
-	  if ((E).error == TARGET_CLOSE_ERROR)				\
-	    /* If the target has closed then it will have deleted any	\
-	       breakpoints inserted within the target inferior, as a	\
-	       result any further attempts to interact with the		\
-	       breakpoint objects is not possible.  Just rethrow the	\
-	       error.  Don't use E to rethrow, to prevent object	\
-	       slicing of the exception.  */				\
-	    throw;							\
-	}								\
-    } while (0)
+/* Rethrow the currently handled exception, if it's a TARGET_CLOSE_ERROR.
+   E is either the currently handled exception, or a copy, or a sliced copy,
+   so we can't rethrow that one, but we can use it to inspect the properties
+   of the currently handled exception.  */
+
+static void
+rethrow_on_target_close_error (const gdb_exception &e)
+{
+  if (e.reason == 0)
+    return;
+  /* Can't set the breakpoint.  */
+
+  if (e.error != TARGET_CLOSE_ERROR)
+    return;
+
+  /* If the target has closed then it will have deleted any breakpoints
+     inserted within the target inferior, as a result any further attempts
+     to interact with the breakpoint objects is not possible.  Just rethrow
+     the error.  Don't use e to rethrow, to prevent object slicing of the
+     exception.  */
+  throw;
+}
 
 /* Insert a low-level "breakpoint" of some type.  BL is the breakpoint
    location.  Any error messages are printed to TMP_ERROR_STREAM; and
@@ -2752,7 +2757,7 @@ insert_bp_location (struct bp_location *bl,
 	    }
 	  catch (gdb_exception &e)
 	    {
-	      RETHROW_ON_TARGET_CLOSE_ERROR (e);
+	      rethrow_on_target_close_error (e);
 	      bp_excpt = std::move (e);
 	    }
 	}
@@ -2792,7 +2797,7 @@ insert_bp_location (struct bp_location *bl,
 		    }
 		  catch (gdb_exception &e)
 		    {
-		      RETHROW_ON_TARGET_CLOSE_ERROR (e);
+		      rethrow_on_target_close_error (e);
 		      bp_excpt = std::move (e);
 		    }
 
@@ -2817,7 +2822,7 @@ insert_bp_location (struct bp_location *bl,
 		}
 	      catch (gdb_exception &e)
 		{
-		  RETHROW_ON_TARGET_CLOSE_ERROR (e);
+		  rethrow_on_target_close_error (e);
 		  bp_excpt = std::move (e);
 		}
 	    }

base-commit: c6d20401a20148b032cb7fb9dba079466a9383cc
-- 
2.35.3


  reply	other threads:[~2022-10-25  7:14 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-24  8:49 [PATCH 0/2] [gdb] Fix rethrow exception slicing Tom de Vries
2022-10-24  8:49 ` [PATCH 1/2] [gdb] Fix rethrow exception slicing in pretty_print_insn Tom de Vries
2022-10-24  8:49 ` [PATCH 2/2] [gdb] Fix rethrow exception slicing in insert_bp_location Tom de Vries
2022-10-24 16:36   ` Pedro Alves
2022-10-24 16:43     ` Tom de Vries
2022-10-24 16:48       ` Simon Marchi
2022-10-25  7:08         ` Tom de Vries
2022-10-24 16:51       ` Pedro Alves
2022-10-25  7:14         ` Tom de Vries [this message]
2022-10-25  9:26           ` Pedro Alves
2022-10-24 11:20 ` [PATCH 0/2] [gdb] Fix rethrow exception slicing Andrew Burgess

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=546f7fe4-7caa-4b78-bfd5-19480d60420e@suse.de \
    --to=tdevries@suse.de \
    --cc=gdb-patches@sourceware.org \
    --cc=pedro@palves.net \
    /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).