[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] [RFC 16/22] x86/percpu: Adapt percpu for PIE support
- To: Thomas Garnier <thgarnie@xxxxxxxxxx>, Brian Gerst <brgerst@xxxxxxxxx>
- From: "H. Peter Anvin" <hpa@xxxxxxxxx>
- Date: Wed, 19 Jul 2017 20:03:16 -0700
- Cc: Michal Hocko <mhocko@xxxxxxxx>, kvm list <kvm@xxxxxxxxxxxxxxx>, Radim Krčmář <rkrcmar@xxxxxxxxxx>, Peter Zijlstra <peterz@xxxxxxxxxxxxx>, Catalin Marinas <catalin.marinas@xxxxxxx>, Christopher Li <sparse@xxxxxxxxxxx>, the arch/x86 maintainers <x86@xxxxxxxxxx>, Paul Gortmaker <paul.gortmaker@xxxxxxxxxxxxx>, Pavel Machek <pavel@xxxxxx>, Kernel Hardening <kernel-hardening@xxxxxxxxxxxxxxxxxx>, Christoph Lameter <cl@xxxxxxxxx>, Thomas Gleixner <tglx@xxxxxxxxxxxxx>, Chris Metcalf <cmetcalf@xxxxxxxxxxxx>, linux-arch <linux-arch@xxxxxxxxxxxxxxx>, Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>, Daniel Borkmann <daniel@xxxxxxxxxxxxx>, Matthew Wilcox <mawilcox@xxxxxxxxxxxxx>, Joerg Roedel <joro@xxxxxxxxxx>, Peter Foley <pefoley2@xxxxxxxxxxx>, Christian Borntraeger <borntraeger@xxxxxxxxxx>, linux-sparse@xxxxxxxxxxxxxxx, Matthias Kaehlcke <mka@xxxxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx, Borislav Petkov <bp@xxxxxxx>, Markus Trippelsdorf <markus@xxxxxxxxxxxxxxx>, Len Brown <len.brown@xxxxxxxxx>, Kees Cook <keescook@xxxxxxxxxxxx>, Arnd Bergmann <arnd@xxxxxxxx>, linux-pm <linux-pm@xxxxxxxxxxxxxxx>, "H . J . Lu" <hjl.tools@xxxxxxxxx>, Steven Rostedt <rostedt@xxxxxxxxxxx>, Borislav Petkov <bp@xxxxxxxxx>, Andy Lutomirski <luto@xxxxxxxxxx>, Josh Poimboeuf <jpoimboe@xxxxxxxxxx>, Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>, Ingo Molnar <mingo@xxxxxxxxxx>, Juergen Gross <jgross@xxxxxxxx>, Dou Liyang <douly.fnst@xxxxxxxxxxxxxx>, Paul Bolle <pebolle@xxxxxxxxxx>, "Paul E . McKenney" <paulmck@xxxxxxxxxxxxxxxxxx>, Baoquan He <bhe@xxxxxxxxxx>, Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>, Jiri Kosina <jkosina@xxxxxxx>, Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>, "Rafael J . Wysocki" <rjw@xxxxxxxxxxxxx>, Linux Kernel Mailing List <linux-kernel@xxxxxxxxxxxxxxx>, Daniel Micay <danielmicay@xxxxxxxxx>, "linux-crypto@xxxxxxxxxxxxxxx" <linux-crypto@xxxxxxxxxxxxxxx>, Rob Landley <rob@xxxxxxxxxxx>, Tejun Heo <tj@xxxxxxxxxx>, Paolo Bonzini <pbonzini@xxxxxxxxxx>, Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>, Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx>, "David S . Miller" <davem@xxxxxxxxxxxxx>, "Kirill A . Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx>
- Delivery-date: Thu, 20 Jul 2017 03:26:57 +0000
- List-id: Xen developer discussion <xen-devel.lists.xen.org>
On 07/19/17 19:21, H. Peter Anvin wrote:
> On 07/19/17 16:33, H. Peter Anvin wrote:
>>>
>>> I agree that it is odd but that's how the compiler generates code. I
>>> will re-explore PIC options with mcmodel=small or medium, as mentioned
>>> on other threads.
>>
>> Why should the way compiler generates code affect the way we do things
>> in assembly?
>>
>> That being said, the compiler now has support for generating this kind
>> of code explicitly via the __seg_gs pointer modifier. That should let
>> us drop the __percpu_prefix and just use variables directly. I suspect
>> we want to declare percpu variables as "volatile __seg_gs" to account
>> for the possibility of CPU switches.
>>
>> Older compilers won't be able to work with this, of course, but I think
>> that it is acceptable for those older compilers to not be able to
>> support PIE.
>>
>
> Grump. It turns out that the compiler doesn't do the right thing for
> symbols marked with the __seg_[fg]s markers. __thread does the right
> thing, but __thread a) has %fs: hard-coded, still, and b) I believe can
> still cache %seg:0 arbitrarily long.
I filed this bug report for gcc:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81490
It might still be possible to work around this by playing really ugly
games with __thread, but I haven't yet figured out how best to do that.
-hpa
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel
|
