linux-kernel - Re: [PATCH v5 2/3] fbcon: Call WARN_CONSOLE

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]

Message-ID: <386d987c-899a-2e98-66ba-6e672dd532ef@redhat.com>
Date:   Wed, 11 Jul 2018 17:01:43 +0200
From:   Hans de Goede <hdegoede@...hat.com>
To:     Steven Rostedt <rostedt@...dmis.org>,
        Thomas Zimmermann <tzimmermann@...e.de>
Cc:     Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
        Petr Mladek <pmladek@...e.com>,
        Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        linux-fbdev@...r.kernel.org, dri-devel@...ts.freedesktop.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5 2/3] fbcon: Call WARN_CONSOLE_UNLOCKED() where
 applicable

Hi,

On 11-07-18 16:52, Steven Rostedt wrote:
> On Wed, 11 Jul 2018 16:46:11 +0200
> Thomas Zimmermann <tzimmermann@...e.de> wrote:
> 
>> Hi
>>
>> Am 28.06.2018 um 11:03 schrieb Hans de Goede:
>>> Replace comments about places where the console lock should be held with
>>> calls to WARN_CONSOLE_UNLOCKED() to assert that it is actually held.
>>
>> Debugging fbcon sometimes requires to not take the console lock. This
>> patch breaks the debugging workaround provided by
>> 'fb.lockless_register_fb'. The dmesg is now filled with warnings about
>> the missing lock.
>>
> 
> What if you make lockless_register_fb visible to fbcon, and then we can
> have a macro:
> 
> #define WARN_FB_CONSOLE_UNLOCKED() 			\
> 	do {						\
> 		if (unlikely(!lockless_register_fb))	\
> 			WARN_CONSOLE_UNLOCKED();	\
> 	} while (0)
> 
> And use that instead?

Yes I'm already preparing a patch like that :)

Without the unlikely though, this is not used in hot paths.

Regards,

Hans



> 
> -- Steve
> 
> 
>> Best regards
>> Thomas
>>
>>> Acked-by: Steven Rostedt (VMware) <rostedt@...dmis.org>
>>> Reviewed-by: Daniel Vetter <daniel.vetter@...ll.ch>
>>> Reviewed-by: Sergey Senozhatsky <sergey.senozhatsky@...il.com>
>>> Signed-off-by: Hans de Goede <hdegoede@...hat.com>
>>> ---
>>> Changes in v3:
>>> -New patch in v3 of this patchset
>>>
>>> Changes in v4:
>>> -Keep the comments about which fbcon functions need locks in place
>>> ---
>>>   drivers/video/fbdev/core/fbcon.c | 11 +++++++++++
>>>   1 file changed, 11 insertions(+)
>>>
>>> diff --git a/drivers/video/fbdev/core/fbcon.c b/drivers/video/fbdev/core/fbcon.c
>>> index c910e74d46ff..cd8d52a967aa 100644
>>> --- a/drivers/video/fbdev/core/fbcon.c
>>> +++ b/drivers/video/fbdev/core/fbcon.c
>>> @@ -828,6 +828,8 @@ static int set_con2fb_map(int unit, int newidx, int user)
>>>   	struct fb_info *oldinfo = NULL;
>>>    	int found, err = 0;
>>>   
>>> +	WARN_CONSOLE_UNLOCKED();
>>> +
>>>   	if (oldidx == newidx)
>>>   		return 0;
>>>   
>>> @@ -3044,6 +3046,8 @@ static int fbcon_fb_unbind(int idx)
>>>   {
>>>   	int i, new_idx = -1, ret = 0;
>>>   
>>> +	WARN_CONSOLE_UNLOCKED();
>>> +
>>>   	if (!fbcon_has_console_bind)
>>>   		return 0;
>>>   
>>> @@ -3094,6 +3098,8 @@ static int fbcon_fb_unregistered(struct fb_info *info)
>>>   {
>>>   	int i, idx;
>>>   
>>> +	WARN_CONSOLE_UNLOCKED();
>>> +
>>>   	idx = info->node;
>>>   	for (i = first_fb_vc; i <= last_fb_vc; i++) {
>>>   		if (con2fb_map[i] == idx)
>>> @@ -3131,6 +3137,9 @@ static int fbcon_fb_unregistered(struct fb_info *info)
>>>   static void fbcon_remap_all(int idx)
>>>   {
>>>   	int i;
>>> +
>>> +	WARN_CONSOLE_UNLOCKED();
>>> +
>>>   	for (i = first_fb_vc; i <= last_fb_vc; i++)
>>>   		set_con2fb_map(i, idx, 0);
>>>   
>>> @@ -3177,6 +3186,8 @@ static int fbcon_fb_registered(struct fb_info *info)
>>>   {
>>>   	int ret = 0, i, idx;
>>>   
>>> +	WARN_CONSOLE_UNLOCKED();
>>> +
>>>   	idx = info->node;
>>>   	fbcon_select_primary(info);
>>>     
>>
>