linux-kernel - Re: [patch 2/2] mm: add dirty_background_bytes and dirty

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: <alpine.DEB.2.00.0811241651370.1550@chino.kir.corp.google.com>
Date:	Mon, 24 Nov 2008 17:00:08 -0800 (PST)
From:	David Rientjes <rientjes@...gle.com>
To:	Andrew Morton <akpm@...ux-foundation.org>
cc:	peterz@...radead.org, david@...morbit.com, cl@...ux-foundation.org,
	linux-kernel@...r.kernel.org
Subject: Re: [patch 2/2] mm: add dirty_background_bytes and dirty_bytes
 sysctls

On Mon, 24 Nov 2008, Andrew Morton wrote:

> > @@ -365,23 +429,29 @@ void
> >  get_dirty_limits(unsigned long *pbackground, unsigned long *pdirty,
> >  		 unsigned long *pbdi_dirty, struct backing_dev_info *bdi)
> >  {
> > -	int background_ratio;		/* Percentages */
> > -	int dirty_ratio;
> >  	unsigned long background;
> >  	unsigned long dirty;
> >  	unsigned long available_memory = determine_dirtyable_memory();
> >  	struct task_struct *tsk;
> >  
> > -	dirty_ratio = vm_dirty_ratio;
> > -	if (dirty_ratio < 5)
> > -		dirty_ratio = 5;
> > +	if (vm_dirty_bytes)
> > +		dirty = (vm_dirty_bytes + PAGE_SIZE) / PAGE_SIZE;
> 
> It would be conventional to use DIV_ROUND_UP() here.
> 

Ah, indeed.  I forgot about those.

> > +	else {
> > +		int dirty_ratio;
> 
> hm, I wonder why vm_dirty_ratio has a signed type.  
> 

vm_dirty_ratio only has an acceptable range between 0 and 100 since it 
uses proc_dointvec_minmax(), so it's not really important.  It could 
definitely be changed, however.

There's a limitation in the sysctl interface where we don't handle 
unsigned ints very well (ignore the reference to unsigned int in the 
proc_dointvec comment).  We lack an unsigned int handler, so all 
proc_dointvec() users are implicitly signed.

For example, if a sysctl's data object were declared with a type specifier 
of unsigned int and it uses proc_dointvec() as the handler, it would have 
a x86_64 max of 2147483648, that of a signed integer.

> > -	background_ratio = dirty_background_ratio;
> > -	if (background_ratio >= dirty_ratio)
> > -		background_ratio = dirty_ratio / 2;
> > +		dirty_ratio = vm_dirty_ratio;
> > +		if (dirty_ratio < 5)
> > +			dirty_ratio = 5;
> > +		dirty = (dirty_ratio * available_memory) / 100;
> > +	}
> > +
> > +	if (dirty_background_bytes)
> > +		background = (dirty_background_bytes + PAGE_SIZE) / PAGE_SIZE;
> 
> DIV_ROUND_UP()?
> 
> > +	else
> > +		background = (dirty_background_ratio * available_memory) / 100;
> >  
> > -	background = (background_ratio * available_memory) / 100;
> > -	dirty = (dirty_ratio * available_memory) / 100;
> > +	if (background >= dirty)
> > +		background = dirty / 2;
> >  	tsk = current;
> >  	if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk)) {
> >  		background += background / 4;
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/