Re: Is _PAGE_PROTNONE set only for user mappings?

From: Mel Gorman <mgorman@techsingularity.net>
To: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Cc: Dave Hansen <dave.hansen@intel.com>,
	Tom Lendacky <thomas.lendacky@amd.com>,
	Rick Edgecombe <rick.p.edgecombe@intel.com>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	Andy Lutomirski <luto@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
	x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
	"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
	Tianyu Lan <Tianyu.Lan@microsoft.com>,
	"Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org, vbabka@suse.cz,
	akpm@linux-foundation.org, willy@infradead.org
Subject: Re: Is _PAGE_PROTNONE set only for user mappings?
Date: Mon, 16 May 2022 14:03:20 +0100	[thread overview]
Message-ID: <20220516130319.GM3441@techsingularity.net> (raw)
In-Reply-To: <Yn3tssUR8w8mC1DJ@hyeyoo>

On Fri, May 13, 2022 at 02:33:38PM +0900, Hyeonggon Yoo wrote:
> On Thu, May 12, 2022 at 11:37:48AM +0100, Mel Gorman wrote:
> > On Wed, May 11, 2022 at 02:20:45PM +0900, Hyeonggon Yoo wrote:
> > > > pgprot_t vm_get_page_prot(unsigned long vm_flags)
> > > > {
> > > >        pgprot_t ret = __pgprot(pgprot_val(protection_map[vm_flags &
> > > >                                (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]) |
> > > >                        pgprot_val(arch_vm_get_page_prot(vm_flags)));
> > > >
> > > >        return arch_filter_pgprot(ret);
> > > > }
> > > > EXPORT_SYMBOL(vm_get_page_prot);
> > > 
> > > I guess it's only set for processes' VMA if no caller is abusing
> > > vm_get_page_prot() for kernel mappings.
> > > 
> > > But yeah, just quick guessing does not make us convinced.
> > > Let's Cc people working on mm.
> > > 
> > > If kernel never uses _PAGE_PROTNONE for kernel mappings, it's just okay
> > > not to clear _PAGE_GLOBAL at first in __change_page_attr() if it's not user address,
> > > because no user will confuse _PAGE_GLOBAL as _PAGE_PROTNONE if it's kernel
> > > address. right?
> > > 
> > 
> > I'm not aware of a case where _PAGE_BIT_PROTNONE is used for a kernel
> > address expecting PROT_NONE semantics instead of the global bit. NUMA
> > Balancing is not going to accidentally treat a kernel address as if it's
> > a NUMA hinting fault. By the time a fault is determining if a PTE access
> > is a numa hinting fault or accesssing a PROT_NONE region, it has been
> > established that it is a userspace address backed by a valid VMA.
> > 
> 
> Thanks Mel, and IIUC nor does do_kern_addr_fault() in arch/x86/mm/fault.c
> expect _PAGE_PROTNONE instead of _PAGE_GLOBAL. I want to make it clear
> in the code that _PAGE_PROTNONE is only used for user mappings.
> 
> How does below look?
> 

I've no strong objections. I worry that this somehow could be used to
set PAGE_USER on a kernel mapping page and maybe a comment would be more
appropriate. However, I'm failing to imagine how NUMA balancing could be
fooled into doing that.

-- 
Mel Gorman
SUSE Labs