[PATCH] [gdb/build] Workaround gcc PR113599

[PATCH] [gdb/build] Workaround gcc PR113599

[Tom de Vries]

Since gcc commit d3f48f68227 ("c++: non-dependent .* operand folding
[PR112427]"), with gdb we run into PR gcc/113599 [1], a wrong-code bug, as
reported in PR build/31281.

Work around this by flipping inherit order:
...
-class thread_info : public refcounted_object,
-		    public intrusive_list_node<thread_info>
+class thread_info : public intrusive_list_node<thread_info>,
+		    public refcounted_object
...

An argument could be made that this isn't necessary, because this occurred in
an unreleased gcc version.

However, I think it could be useful when bisecting gcc for other problems in
building gdb.  Having this workaround means the bisect won't reintroduce the
problem.  Furthermore, the workaround is harmless.

Tested on Fedora rawhide x86_64.

Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31281

[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113599
---
 gdb/gdbthread.h | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/gdb/gdbthread.h b/gdb/gdbthread.h
index e7035d40ad4..1d9730fd1fa 100644
--- a/gdb/gdbthread.h
+++ b/gdb/gdbthread.h
@@ -244,10 +244,11 @@ using private_thread_info_up = std::unique_ptr<private_thread_info>;
    strong reference, and is thus not accounted for in the thread's
    refcount.
 
-   The intrusive_list_node base links threads in a per-inferior list.  */
+   The intrusive_list_node base links threads in a per-inferior list.
+   We place it first in the inherit orer to work around PR gcc/113599.  */
 
-class thread_info : public refcounted_object,
-		    public intrusive_list_node<thread_info>
+class thread_info : public intrusive_list_node<thread_info>,
+		    public refcounted_object
 {
 public:
   explicit thread_info (inferior *inf, ptid_t ptid);

base-commit: 726f209eb1b05842d816eac8b0b8f9c7f6cd9fbc
-- 
2.35.3

Re: [PATCH] [gdb/build] Workaround gcc PR113599

[Tom Tromey]

>>>>> "Tom" == Tom de Vries <tdevries@suse.de> writes:

Tom> However, I think it could be useful when bisecting gcc for other problems in
Tom> building gdb.  Having this workaround means the bisect won't reintroduce the
Tom> problem.  Furthermore, the workaround is harmless.

I think the change is harmless and well-documented, so I think it is ok.
Approved-By: Tom Tromey <tom@tromey.com>

thanks,
Tom

Re: [PATCH] [gdb/build] Workaround gcc PR113599

[Sam James]

Tom de Vries <tdevries@suse.de> writes:

> Since gcc commit d3f48f68227 ("c++: non-dependent .* operand folding
> [PR112427]"), with gdb we run into PR gcc/113599 [1], a wrong-code bug, as
> reported in PR build/31281.
>
> Work around this by flipping inherit order:
> ...
> -class thread_info : public refcounted_object,
> -		    public intrusive_list_node<thread_info>
> +class thread_info : public intrusive_list_node<thread_info>,
> +		    public refcounted_object
> ...
>
> An argument could be made that this isn't necessary, because this occurred in
> an unreleased gcc version.
>
> However, I think it could be useful when bisecting gcc for other problems in
> building gdb.  Having this workaround means the bisect won't reintroduce the
> problem.  Furthermore, the workaround is harmless.
>
> Tested on Fedora rawhide x86_64.
>
> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31281
>
> [1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113599
> ---
>  gdb/gdbthread.h | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/gdb/gdbthread.h b/gdb/gdbthread.h
> index e7035d40ad4..1d9730fd1fa 100644
> --- a/gdb/gdbthread.h
> +++ b/gdb/gdbthread.h
> @@ -244,10 +244,11 @@ using private_thread_info_up = std::unique_ptr<private_thread_info>;
>     strong reference, and is thus not accounted for in the thread's
>     refcount.
>  
> -   The intrusive_list_node base links threads in a per-inferior list.  */
> +   The intrusive_list_node base links threads in a per-inferior list.
> +   We place it first in the inherit orer to work around PR gcc/113599.  */

order

>  
> -class thread_info : public refcounted_object,
> -		    public intrusive_list_node<thread_info>
> +class thread_info : public intrusive_list_node<thread_info>,
> +		    public refcounted_object
>  {
>  public:
>    explicit thread_info (inferior *inf, ptid_t ptid);
>
> base-commit: 726f209eb1b05842d816eac8b0b8f9c7f6cd9fbc

Re: [PATCH] [gdb/build] Workaround gcc PR113599

[Tom de Vries]

On 1/25/24 16:27, Sam James wrote:
> 
> Tom de Vries <tdevries@suse.de> writes:
> 
>> Since gcc commit d3f48f68227 ("c++: non-dependent .* operand folding
>> [PR112427]"), with gdb we run into PR gcc/113599 [1], a wrong-code bug, as
>> reported in PR build/31281.
>>
>> Work around this by flipping inherit order:
>> ...
>> -class thread_info : public refcounted_object,
>> -		    public intrusive_list_node<thread_info>
>> +class thread_info : public intrusive_list_node<thread_info>,
>> +		    public refcounted_object
>> ...
>>
>> An argument could be made that this isn't necessary, because this occurred in
>> an unreleased gcc version.
>>
>> However, I think it could be useful when bisecting gcc for other problems in
>> building gdb.  Having this workaround means the bisect won't reintroduce the
>> problem.  Furthermore, the workaround is harmless.
>>
>> Tested on Fedora rawhide x86_64.
>>
>> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31281
>>
>> [1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113599
>> ---
>>   gdb/gdbthread.h | 7 ++++---
>>   1 file changed, 4 insertions(+), 3 deletions(-)
>>
>> diff --git a/gdb/gdbthread.h b/gdb/gdbthread.h
>> index e7035d40ad4..1d9730fd1fa 100644
>> --- a/gdb/gdbthread.h
>> +++ b/gdb/gdbthread.h
>> @@ -244,10 +244,11 @@ using private_thread_info_up = std::unique_ptr<private_thread_info>;
>>      strong reference, and is thus not accounted for in the thread's
>>      refcount.
>>   
>> -   The intrusive_list_node base links threads in a per-inferior list.  */
>> +   The intrusive_list_node base links threads in a per-inferior list.
>> +   We place it first in the inherit orer to work around PR gcc/113599.  */
> 
> order
> 

Thanks for catching that, fixed.

- Tom

>>   
>> -class thread_info : public refcounted_object,
>> -		    public intrusive_list_node<thread_info>
>> +class thread_info : public intrusive_list_node<thread_info>,
>> +		    public refcounted_object
>>   {
>>   public:
>>     explicit thread_info (inferior *inf, ptid_t ptid);
>>
>> base-commit: 726f209eb1b05842d816eac8b0b8f9c7f6cd9fbc
>