linux-kernel - Re: [PATCH 1/2] debugobject: fix concurrency issues with is_static

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Sat, 04 Mar 2023 01:14:21 +0100
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Schspa Shi <schspa@...il.com>, Waiman Long <longman@...hat.com>
Cc:     swboyd@...omium.org, linux@...ck-us.net, wuchi.zero@...il.com,
        linux-kernel@...r.kernel.org,
        syzbot+5093ba19745994288b53@...kaller.appspotmail.com
Subject: Re: [PATCH 1/2] debugobject: fix concurrency issues with
 is_static_object

On Sat, Mar 04 2023 at 01:53, Schspa Shi wrote:
> Waiman Long <longman@...hat.com> writes:
>>> diff --git a/include/linux/debugobjects.h b/include/linux/debugobjects.h
>>> index 32444686b6ff4..544a6111b97f6 100644
>>> --- a/include/linux/debugobjects.h
>>> +++ b/include/linux/debugobjects.h
>>> @@ -30,6 +30,7 @@ struct debug_obj {
>>>   	enum debug_obj_state	state;
>>>   	unsigned int		astate;
>>>   	void			*object;
>>> +	bool			is_static;
>>>   	const struct debug_obj_descr *descr;
>>>   };
>>
>> The patch looks reasonable. My main concern is the increase in size of the
>> debug_obj structure. It is an additional 8 bytes on 64-bit arches. How much will
>> we save performance-wise by caching it in the debug_obj. Alternatively, you may
>> pack it within the current size by, maybe, reducing the size of state.
>>
>
> Yes, we can change this to:
>
> struct debug_obj {
> 	struct hlist_node	node;
> 	struct {
> 		enum debug_obj_state	state : 31;
> 		bool					is_static : 1;
> 	};

and thereby making debugobjects even more slower than it is right
now. Please check the resulting assembly code before proposing quick
"solutions".

Thanks,

        tglx