From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-it0-f71.google.com (mail-it0-f71.google.com [209.85.214.71]) by kanga.kvack.org (Postfix) with ESMTP id 797DF6B0038 for ; Sun, 19 Mar 2017 04:27:00 -0400 (EDT) Received: by mail-it0-f71.google.com with SMTP id g138so87429591itb.4 for ; Sun, 19 Mar 2017 01:27:00 -0700 (PDT) Received: from mail-it0-x233.google.com (mail-it0-x233.google.com. [2607:f8b0:4001:c0b::233]) by mx.google.com with ESMTPS id c195si7235749itb.108.2017.03.19.01.26.59 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Sun, 19 Mar 2017 01:26:59 -0700 (PDT) Received: by mail-it0-x233.google.com with SMTP id y18so4914652itc.0 for ; Sun, 19 Mar 2017 01:26:59 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <877f3lfzdo.fsf@skywalker.in.ibm.com> References: <20170313055020.69655-1-kirill.shutemov@linux.intel.com> <20170313055020.69655-27-kirill.shutemov@linux.intel.com> <87a88jg571.fsf@skywalker.in.ibm.com> <20170317175714.3bvpdylaaudf4ig2@node.shutemov.name> <877f3lfzdo.fsf@skywalker.in.ibm.com> From: "Kirill A. Shutemov" Date: Sun, 19 Mar 2017 11:26:58 +0300 Message-ID: Subject: Re: [PATCH 26/26] x86/mm: allow to have userspace mappings above 47-bits Content-Type: multipart/alternative; boundary=001a11449ffca93443054b112ced Sender: owner-linux-mm@kvack.org List-ID: To: "Aneesh Kumar K.V" Cc: linux-arch , Thomas Gleixner , Ingo Molnar , Michal Hocko , linux-kernel@vger.kernel.org, Andrew Morton , "Kirill A. Shutemov" , Arnd Bergmann , Andy Lutomirski , Dave Hansen , linux-mm@kvack.org, "H. Peter Anvin" , x86@kernel.org, Andi Kleen , Linus Torvalds --001a11449ffca93443054b112ced Content-Type: text/plain; charset=UTF-8 On Mar 19, 2017 09:25, "Aneesh Kumar K.V" wrote: "Kirill A. Shutemov" writes: > On Fri, Mar 17, 2017 at 11:23:54PM +0530, Aneesh Kumar K.V wrote: >> "Kirill A. Shutemov" writes: >> >> > On x86, 5-level paging enables 56-bit userspace virtual address space. >> > Not all user space is ready to handle wide addresses. It's known that >> > at least some JIT compilers use higher bits in pointers to encode their >> > information. It collides with valid pointers with 5-level paging and >> > leads to crashes. >> > >> > To mitigate this, we are not going to allocate virtual address space >> > above 47-bit by default. >> > >> > But userspace can ask for allocation from full address space by >> > specifying hint address (with or without MAP_FIXED) above 47-bits. >> > >> > If hint address set above 47-bit, but MAP_FIXED is not specified, we try >> > to look for unmapped area by specified address. If it's already >> > occupied, we look for unmapped area in *full* address space, rather than >> > from 47-bit window. >> > >> > This approach helps to easily make application's memory allocator aware >> > about large address space without manually tracking allocated virtual >> > address space. >> > >> >> So if I have done a successful mmap which returned > 128TB what should a >> following mmap(0,...) return ? Should that now search the *full* address >> space or below 128TB ? > > No, I don't think so. And this implementation doesn't do this. > > It's safer this way: if an library can't handle high addresses, it's > better not to switch it automagically to full address space if other part > of the process requested high address. > What is the epectation when the hint addr is below 128TB but addr + len > 128TB ? Should such mmap request fail ? Yes, I believe so. --001a11449ffca93443054b112ced Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable


On Mar 19, 2017 09:25, "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com= > wrote:
"Kirill A. Shutemov" <kirill@shutemov.name> writes:

> On Fri, Mar 17, 2017 at 11:23:54PM +0530, Aneesh Kumar K.V wrote:
>> "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> writes: >>
>> > On x86, 5-level paging enables 56-bit userspace virtual addre= ss space.
>> > Not all user space is ready to handle wide addresses. It'= s known that
>> > at least some JIT compilers use higher bits in pointers to en= code their
>> > information. It collides with valid pointers with 5-level pag= ing and
>> > leads to crashes.
>> >
>> > To mitigate this, we are not going to allocate virtual addres= s space
>> > above 47-bit by default.
>> >
>> > But userspace can ask for allocation from full address space = by
>> > specifying hint address (with or without MAP_FIXED) above 47-= bits.
>> >
>> > If hint address set above 47-bit, but MAP_FIXED is not specif= ied, we try
>> > to look for unmapped area by specified address. If it's a= lready
>> > occupied, we look for unmapped area in *full* address space, = rather than
>> > from 47-bit window.
>> >
>> > This approach helps to easily make application's memory a= llocator aware
>> > about large address space without manually tracking allocated= virtual
>> > address space.
>> >
>>
>> So if I have done a successful mmap which returned > 128TB what= should a
>> following mmap(0,...) return ? Should that now search the *full* a= ddress
>> space or below 128TB ?
>
> No, I don't think so. And this implementation doesn't do this.=
>
> It's safer this way: if an library can't handle high addresses= , it's
> better not to switch it automagically to full address space if other p= art
> of the process requested high address.
>

What is the epectation when the hint addr is below 128TB but addr + l= en >
128TB ? Should such mmap request fail ?

Yes, I believe so.


--001a11449ffca93443054b112ced-- -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org