public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Tom de Vries <tdevries@suse.de>
To: gdb-patches@sourceware.org
Subject: [RFC 1/5] [gdb] Catch gdb_exception_quit in munmap_list::~munmap_list
Date: Tue, 25 Oct 2022 13:19:41 +0200	[thread overview]
Message-ID: <20221025111945.23886-2-tdevries@suse.de> (raw)
In-Reply-To: <20221025111945.23886-1-tdevries@suse.de>

I tried out adding this exception slicing detection code in
munmap_list::~munmap_list () in gdb/compile/compile-object-load.c:
...
       try
         {
	   ...
         }
       catch (const gdb_exception_error &ex)
         {
           /* There's not much the user can do, so just ignore
              this.  */
         }
+      catch (const gdb_exception_quit &ex)
+        {
+          throw;
+        }
+      catch (const gdb_exception &ex)
+        {
+          gdb_assert (0);
+        }
...
and ran into:
...
gdb/compile/compile-object-load.c: In destructor ‘munmap_list::~munmap_list()’:
gdb/compile/compile-object-load.c:64:4: error: \
  throw will always call terminate() [-Werror=terminate]
    throw;
    ^~~~~
gdb/compile/compile-object-load.c:64:4: note: \
  in C++11 destructors default to noexcept
...

This raises the question of how a gdb_exception_quit should be handled in the
destructor.

Currently it's handled implicitly, and will result in std::terminate, which in
absense of a std::set_terminate should call abort, after printing the type
of the current exception, in other words gdb_exception_quit.

We have a couple of choices to handle this explicitly:
- std::terminate (no change in behaviour),
- gdb_assert (false) (also calls abort, but with more info), or
- ignore (as we do for gdb_exception_error)

It seems to me that a gdb_exception_quit should not cause an abort, so handle
explicitly by ignoring:
...
-      catch (const gdb_exception_error &ex)
+      catch (const gdb_exception &ex)
...

Tested on x86_64-linux (though I don't have a setup with libcc1).
---
 gdb/compile/compile-object-load.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gdb/compile/compile-object-load.c b/gdb/compile/compile-object-load.c
index 4c94ec53899..ca58775880b 100644
--- a/gdb/compile/compile-object-load.c
+++ b/gdb/compile/compile-object-load.c
@@ -54,7 +54,7 @@ munmap_list::~munmap_list ()
 	{
 	  gdbarch_infcall_munmap (target_gdbarch (), item.addr, item.size);
 	}
-      catch (const gdb_exception_error &ex)
+      catch (const gdb_exception &ex)
 	{
 	  /* There's not much the user can do, so just ignore
 	     this.  */
-- 
2.35.3


  reply	other threads:[~2022-10-25 11:19 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-25 11:19 [RFC 0/5] [gdb] Catch rethrow exception slicing Tom de Vries
2022-10-25 11:19 ` Tom de Vries [this message]
2022-10-25 11:19 ` [RFC 2/5] [gdbsupport] Add CATCH_ERROR and CATCH_ERROR_QUIT Tom de Vries
2022-10-25 11:19 ` [RFC 3/5] [gdb/contrib] Add refactor.py Tom de Vries
2022-10-25 11:19 ` [RFC 4/5] [gdb/contrib] Add refactor_gdb_exception.py Tom de Vries
2022-10-25 11:19 ` [RFC 5/5] [gdb] Do refactor_gdb_exception refactoring Tom de Vries

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=20221025111945.23886-2-tdevries@suse.de \
    --to=tdevries@suse.de \
    --cc=gdb-patches@sourceware.org \
    /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).