2017-07-20 00:17:11

by H. Peter Anvin

[permalink] [raw]
Subject: Re: [RFC 06/22] kvm: Adapt assembly for PIE support

<[email protected]>,Chris Metcalf <[email protected]>,"Paul E . McKenney" <[email protected]>,Andrew Morton <[email protected]>,Christopher Li <[email protected]>,Dou Liyang <[email protected]>,Masahiro Yamada <[email protected]>,Daniel Borkmann <[email protected]>,Markus Trippelsdorf <[email protected]>,Peter Foley <[email protected]>,Steven Rostedt <[email protected]>,Tim Chen <[email protected]>,Catalin Marinas <[email protected]>,Matthew Wilcox <[email protected]>,Michal Hocko <[email protected]>,Rob Landley <[email protected]>,Jiri Kosina <[email protected]>,"H . J . Lu" <[email protected]>,Paul Bolle <[email protected]>,Baoquan He <[email protected]>,Daniel Micay <[email protected]>,the arch/x86 maintainers <[email protected]>,"[email protected]" <[email protected]>,Linux Kernel Mailing List <[email protected]>,[email protected],kvm list
<[email protected]>,linux-pm <[email protected]>,linux-arch <[email protected]>,Linux-Sparse <[email protected]>,Kernel Hardening <[email protected]>
From: [email protected]
Message-ID: <[email protected]>

On July 19, 2017 3:58:07 PM PDT, Ard Biesheuvel <[email protected]> wrote:
>On 19 July 2017 at 23:27, H. Peter Anvin <[email protected]> wrote:
>> On 07/19/17 08:40, Thomas Garnier wrote:
>>>>
>>>> This doesn't look right. It's accessing a per-cpu variable. The
>>>> per-cpu section is an absolute, zero-based section and not subject
>to
>>>> relocation.
>>>
>>> PIE does not respect the zero-based section, it tries to have
>>> everything relative. Patch 16/22 also adapt per-cpu to work with PIE
>>> (while keeping the zero absolute design by default).
>>>
>>
>> This is silly. The right thing is for PIE is to be explicitly
>absolute,
>> without (%rip). The use of (%rip) memory references for percpu is
>just
>> an optimization.
>>
>
>Sadly, there is an issue in binutils that may prevent us from doing
>this as cleanly as we would want.
>
>For historical reasons, bfd.ld emits special symbols like
>__GLOBAL_OFFSET_TABLE__ as absolute symbols with a section index of
>SHN_ABS, even though it is quite obvious that they are relative like
>any other symbol that points into the image. Unfortunately, this means
>that binutils needs to emit R_X86_64_RELATIVE relocations even for
>SHN_ABS symbols, which means we lose the ability to use both absolute
>and relocatable symbols in the same PIE image (unless the reloc tool
>can filter them out)
>
>More info here:
>https://sourceware.org/bugzilla/show_bug.cgi?id=19818

The reloc tool already has the ability to filter symbols.
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.