linux-kernel - Re: [PATCH] Tight check of pfn_valid on sparsemem

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: <20100728155617.GA5401@barrios-desktop>
Date:	Thu, 29 Jul 2010 00:56:17 +0900
From:	Minchan Kim <minchan.kim@...il.com>
To:	Christoph Lameter <cl@...ux-foundation.org>
Cc:	KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>,
	Milton Miller <miltonm@....com>, linux-kernel@...r.kernel.org,
	linux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>,
	Russell King <linux@....linux.org.uk>,
	Mel Gorman <mel@....ul.ie>,
	Johannes Weiner <hannes@...xchg.org>,
	Kukjin Kim <kgene.kim@...sung.com>
Subject: Re: [PATCH] Tight check of pfn_valid on sparsemem - v4

On Wed, Jul 28, 2010 at 10:14:51AM -0500, Christoph Lameter wrote:
> On Wed, 28 Jul 2010, Minchan Kim wrote:
> 
> > static inline int memmap_valid(unsigned long pfn)
> > {
> >        struct page *page = pfn_to_page(pfn);
> >        struct page *__pg = virt_to_page(page);
> 
> Does that work both for vmemmap and real mmapping?

When Kame suggested this idea, he doesn't consider vmemmap model. 
(He prevent this featur's enabling by config !SPARSEMEM_VMEMMAP)

config SPARSEMEM_HAS_HOLE
       bool "allow holes in sparsemem's memmap"
       depends on ARM && SPARSEMEM && !SPARSEMEM_VMEMMAP
       default n

When I change it with ARCH_HAS_HOLES_MEMORYMODEL, it was my mistake.
I can change it with ARCH_HAS_HOLES_MEMORYMODEL && !SPARSE_VMEMMAP. 

I wonder whether we supports VMEMMAP. 
That's because hole problem of sparsemem is specific on ARM. 
ARM forks uses it for saving memory space but VMEMMAP does use more memory.
I think it's irony. 

> 
> >        return page_private(__pg) == MAGIC_MEMMAP && PageReserved(__pg);
> > }
> 
> Problem is that pages may be allocated for the mmap from a variety of
> places. The pages in mmap_init_zone() and allocated during boot may have
> PageReserved set whereas the page allocated via vmemmap_alloc_block() have
> PageReserved cleared since they came from the page allocator.
> 
> You need to have consistent use of PageReserved in page structs for the
> mmap in order to do this properly.

Yes if we supports both model. 

> 
> Simplest scheme would be to clear PageReserved() in all page struct
> associated with valid pages and clear those for page structs that do not
> refer to valid pages.

I can't understand your words.
Clear PG_resereved in valid pages and invalid pages both?

I guess your code look like that clear PG_revered on valid memmap
but set PG_reserved on invalid memmap.
Right?

invalid memmap pages will be freed by free_memmap and will be used 
on any place. How do we make sure it has PG_reserved?

Maybe I don't understand your point. 


> 
> Then
> 
> mmap_valid = !PageReserved(xxx(pfn_to_page(pfn))

-- 
Kind regards,
Minchan Kim
--
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/