From: Simon Marchi <simark@simark.ca>
To: Tom Tromey <tom@tromey.com>, gdb-patches@sourceware.org
Subject: Re: [PP?] [PATCH 4/6] Transfer module ownership to do_module_cleanup
Date: Sun, 9 Aug 2020 20:48:46 -0400 [thread overview]
Message-ID: <2be49232-0543-b73a-aa35-477a9f2a7895@simark.ca> (raw)
In-Reply-To: <20200809135258.8207-5-tom@tromey.com>
On 2020-08-09 9:52 a.m., Tom Tromey wrote:
> This changes the do_module_cleanup structure to simply hold on to the
> module itself. This lets us remove most members from
> do_module_cleanup.
>
> gdb/ChangeLog
> 2020-08-08 Tom Tromey <tom@tromey.com>
>
> * compile/compile-object-run.c (struct do_module_cleanup): Add
> parameters to constructor. Update destructor.
> <source_file, scope, scope_data, out_value_type, out_value_addr,
> munmap_list_head, objfile_name_string>: Remove.
> <module>: New member.
> (do_module_cleanup): Update.
> (compile_object_run): Update.
> ---
> gdb/ChangeLog | 10 +++++
> gdb/compile/compile-object-run.c | 67 ++++++++++++--------------------
> 2 files changed, 35 insertions(+), 42 deletions(-)
>
> diff --git a/gdb/compile/compile-object-run.c b/gdb/compile/compile-object-run.c
> index 5a680a6723f..ac0a995fee9 100644
> --- a/gdb/compile/compile-object-run.c
> +++ b/gdb/compile/compile-object-run.c
> @@ -32,13 +32,16 @@
>
> struct do_module_cleanup
> {
> - do_module_cleanup () = default;
> + do_module_cleanup (int *ptr, compile_module_up &&mod)
> + : executedp (ptr),
> + module (std::move (mod))
> + {
> + }
>
> ~do_module_cleanup ()
> {
> - delete munmap_list_head;
> - xfree (source_file);
> - xfree (objfile_name_string);
> + delete module->munmap_list_head;
> + xfree (module->source_file);
> }
>
> DISABLE_COPY_AND_ASSIGN (do_module_cleanup);
> @@ -47,22 +50,8 @@ struct do_module_cleanup
> The pointer may be NULL. */
> int *executedp;
>
> - /* .c file OBJFILE was built from. It needs to be xfree-d. */
> - char *source_file;
> -
> - /* Copy from struct compile_module. */
> - enum compile_i_scope_types scope;
> - void *scope_data;
> -
> - /* Copy from struct compile_module. */
> - struct type *out_value_type;
> - CORE_ADDR out_value_addr;
> -
> - /* Copy from struct compile_module. */
> - struct munmap_list *munmap_list_head;
> -
> - /* objfile_name of our objfile. */
> - char *objfile_name_string;
> + /* The compile module. */
> + compile_module_up module;
> };
>
> /* Cleanup everything after the inferior function dummy frame gets
> @@ -80,22 +69,26 @@ do_module_cleanup (void *arg, int registers_valid)
>
> /* This code cannot be in compile_object_run as OUT_VALUE_TYPE
> no longer exists there. */
> - if (data->scope == COMPILE_I_PRINT_ADDRESS_SCOPE
> - || data->scope == COMPILE_I_PRINT_VALUE_SCOPE)
> + if (data->module->scope == COMPILE_I_PRINT_ADDRESS_SCOPE
> + || data->module->scope == COMPILE_I_PRINT_VALUE_SCOPE)
> {
> struct value *addr_value;
> - struct type *ptr_type = lookup_pointer_type (data->out_value_type);
> + struct type *ptr_type
> + = lookup_pointer_type (data->module->out_value_type);
>
> - addr_value = value_from_pointer (ptr_type, data->out_value_addr);
> + addr_value = value_from_pointer (ptr_type,
> + data->module->out_value_addr);
>
> /* SCOPE_DATA would be stale unless EXECUTEDP != NULL. */
> - compile_print_value (value_ind (addr_value), data->scope_data);
> + compile_print_value (value_ind (addr_value),
> + data->module->scope_data);
> }
> }
>
> + const char *objfile_name_s = objfile_name (data->module->objfile);
> for (objfile *objfile : current_program_space->objfiles ())
> if ((objfile->flags & OBJF_USERLOADED) == 0
> - && (strcmp (objfile_name (objfile), data->objfile_name_string) == 0))
> + && (strcmp (objfile_name (objfile), objfile_name_s) == 0))
> {
> objfile->unlink ();
>
> @@ -106,10 +99,10 @@ do_module_cleanup (void *arg, int registers_valid)
> }
>
> /* Delete the .c file. */
> - unlink (data->source_file);
> + unlink (data->module->source_file);
>
> /* Delete the .o file. */
> - unlink (data->objfile_name_string);
> + unlink (objfile_name_s);
The objfile gets destroyed when unlinked. Is it guaranteed that objfile_name_s is still alive at that point?
Simon
next prev parent reply other threads:[~2020-08-10 0:48 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-09 13:52 [PATCH 0/6] Avoid manual memory management in gdb/compile/ Tom Tromey
2020-08-09 13:52 ` [PATCH 1/6] Remove some manual memory management from compile interface Tom Tromey
2020-08-09 22:34 ` Simon Marchi
2020-09-19 23:45 ` Tom Tromey
2020-09-19 23:52 ` Simon Marchi
2020-09-23 12:56 ` Tom Tromey
2020-08-09 13:52 ` [PATCH 2/6] Use new/delete for do_module_cleanup Tom Tromey
2020-08-09 22:45 ` [PP?] " Simon Marchi
2020-09-23 13:36 ` Tom Tromey
2020-08-09 13:52 ` [PATCH 3/6] Introduce and use compile_module_up Tom Tromey
2020-08-09 13:52 ` [PATCH 4/6] Transfer module ownership to do_module_cleanup Tom Tromey
2020-08-10 0:48 ` Simon Marchi [this message]
2020-09-23 14:55 ` [PP?] " Tom Tromey
2020-08-09 13:52 ` [PATCH 5/6] Simplify compile_module cleanup Tom Tromey
2020-08-09 13:52 ` [PATCH 6/6] Avoid manual memory management of argv arrays in gdb/compile Tom Tromey
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=2be49232-0543-b73a-aa35-477a9f2a7895@simark.ca \
--to=simark@simark.ca \
--cc=gdb-patches@sourceware.org \
--cc=tom@tromey.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