Re: [PATCH] vmap(): don't allow invalid pages

From: "Russell King (Oracle)" <linux@armlinux.org.uk>
To: Matthew Wilcox <willy@infradead.org>
Cc: Robin Murphy <robin.murphy@arm.com>,
	Yury Norov <yury.norov@gmail.com>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Nicholas Piggin <npiggin@gmail.com>,
	Ding Tianhong <dingtianhong@huawei.com>,
	Anshuman Khandual <anshuman.khandual@arm.com>,
	Alexey Klimov <aklimov@redhat.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [PATCH] vmap(): don't allow invalid pages
Date: Wed, 19 Jan 2022 22:38:42 +0000	[thread overview]
Message-ID: <YeiS8vDcQ6/B26fG@shell.armlinux.org.uk> (raw)
In-Reply-To: <YehoBQZvSmwXjqJ3@casper.infradead.org>

On Wed, Jan 19, 2022 at 07:35:33PM +0000, Matthew Wilcox wrote:
> On Wed, Jan 19, 2022 at 06:57:34PM +0000, Russell King (Oracle) wrote:
> > On Wed, Jan 19, 2022 at 06:01:24PM +0000, Matthew Wilcox wrote:
> > > On Wed, Jan 19, 2022 at 05:54:15PM +0000, Russell King (Oracle) wrote:
> > > > On Wed, Jan 19, 2022 at 04:27:32PM +0000, Matthew Wilcox wrote:
> > > > > On Wed, Jan 19, 2022 at 01:28:14PM +0000, Robin Murphy wrote:
> > > > > > > +		if (WARN_ON(!pfn_valid(page_to_pfn(page))))
> > > > > > 
> > > > > > Is it page_to_pfn() guaranteed to work without blowing up if page is invalid
> > > > > > in the first place? Looking at the CONFIG_SPARSEMEM case I'm not sure that's
> > > > > > true...
> > > > > 
> > > > > Even if it does blow up, at least it's blowing up here where someone
> > > > > can start to debug it, rather than blowing up on first access, where
> > > > > we no longer have the invlid struct page pointer.
> > > > > 
> > > > > I don't think we have a 'page_valid' function which will tell us whether
> > > > > a random pointer is actually a struct page or not.
> > > > 
> > > > Isn't it supposed to be:
> > > > 
> > > > 	if (!pfn_valid(pfn)) {
> > > > 		handle invalid pfn;
> > > > 	}
> > > > 
> > > > 	page = pfn_to_page(pfn);
> > > > 
> > > > Anything else - even trying to convert an invalid page back to a pfn,
> > > > could well be unreliable (sparsemem or discontigmem). 
> > > 
> > > This function is passed an array of pages.  We have no way of doing
> > > what you propose.
> > 
> > You can't go from a struct page to "this is valid", it's too late by the
> > time you call vmap() - that's my fundamental point.
> 
> Yes, and we have debugging code in __virt_to_phys() that would have
> caught this, had Yury enabled CONFIG_DEBUG_VIRTUAL.  My point is that
> in this instance, page_to_pfn() doesn't crash, which lets vmap() set
> up a mapping to a completely bogus physical address.  We're better
> off checking pfn_valid() here than not.

I don't disagree that pfn_valid() will catch _some_ but it should, no,
must not be a subsitute for ensuring that the proper checks are done
when creating e.g. an array of struct pages.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!