Re: [PATCH 2.6.14-rc2] fix incorrect mm->hiwater_vm and mm->hiwater_rss

[Frank van Maarseveen <frankvm@frankvm.com>]

On Wed, Sep 21, 2005 at 04:35:56PM +0100, Hugh Dickins wrote:
> On Wed, 21 Sep 2005, Frank van Maarseveen wrote:
> 
> > --- ./fs/proc/task_mmu.c.orig	2005-07-07 14:22:12.000000000 +0200
> > +++ ./fs/proc/task_mmu.c	2005-09-16 13:51:56.000000000 +0200
> > @@ -14,6 +14,7 @@
> >  	text = (PAGE_ALIGN(mm->end_code) - (mm->start_code & PAGE_MASK)) >> 10;
> >  	lib = (mm->exec_vm << (PAGE_SHIFT-10)) - text;
> >  	buffer += sprintf(buffer,
> > +		"VmPeak:\t%8lu kB\n"
> 
> Good naming.
> 
> >  		"VmSize:\t%8lu kB\n"
> >  		"VmLck:\t%8lu kB\n"
> >  		"VmRSS:\t%8lu kB\n"
> > @@ -22,6 +23,7 @@
> >  		"VmExe:\t%8lu kB\n"
> >  		"VmLib:\t%8lu kB\n"
> >  		"VmPTE:\t%8lu kB\n",
> > +		mm->hiwater_vm << (PAGE_SHIFT-10),
> >  		(mm->total_vm - mm->reserved_vm) << (PAGE_SHIFT-10),
> >  		mm->locked_vm << (PAGE_SHIFT-10),
> >  		get_mm_counter(mm, rss) << (PAGE_SHIFT-10),
> 
> I do keep wondering what's so interesting about this hiwater_vm
> (but would regret proposing other statistics to gather instead):
> perhaps you're showing it just because it's there?

No, hiwater_vm is exactly what we need. Before hiwater_vm I had my own patch
to get exactly the same result but the member was called "peak_vm" instead.

Most software we develop at my work uses this to test for regression
in memory usage. Just before program exit it reads /proc/pid/status:VmPeak
and reports it. Scripts do the rest.

-- 
Frank
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/