From: Ingo Molnar <mingo@kernel.org>
To: Willy Tarreau <w@1wt.eu>
Cc: linux-kernel@vger.kernel.org, x86@kernel.org,
Andy Lutomirski <luto@kernel.org>, Borislav Petkov <bp@alien8.de>,
Brian Gerst <brgerst@gmail.com>,
Dave Hansen <dave.hansen@linux.intel.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
Peter Zijlstra <peterz@infradead.org>,
Thomas Gleixner <tglx@linutronix.de>,
Josh Poimboeuf <jpoimboe@redhat.com>,
"H. Peter Anvin" <hpa@zytor.com>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Kees Cook <keescook@chromium.org>
Subject: Re: [RFC PATCH v2 6/6] x86/entry/pti: don't switch PGD on when pti_disable is set
Date: Wed, 10 Jan 2018 08:15:10 +0100 [thread overview]
Message-ID: <20180110071510.nfice2pjtuylqirc@gmail.com> (raw)
In-Reply-To: <1515502580-12261-7-git-send-email-w@1wt.eu>
* Willy Tarreau <w@1wt.eu> wrote:
> When a syscall returns to userspace with pti_disable set, it means the
> current mm is configured to disable page table isolation (PTI). In this
> case, returns from kernel to user will not switch the CR3, leaving it
> to the kernel one which already maps both user and kernel pages. This
> avoids a TLB flush, and saves another one on next entry.
>
> Thanks to these changes, haproxy running under KVM went back from
> 12700 conn/s (without PCID) or 19700 (with PCID) to 23100 once loaded
> after calling prctl(), indicating that PTI has no measurable impact on
> this workload.
>
> Signed-off-by: Willy Tarreau <w@1wt.eu>
> Cc: Andy Lutomirski <luto@kernel.org>
> Cc: Borislav Petkov <bp@alien8.de>
> Cc: Brian Gerst <brgerst@gmail.com>
> Cc: Dave Hansen <dave.hansen@linux.intel.com>
> Cc: Ingo Molnar <mingo@kernel.org>
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Josh Poimboeuf <jpoimboe@redhat.com>
> Cc: "H. Peter Anvin" <hpa@zytor.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: Kees Cook <keescook@chromium.org>
>
> v2:
> - use pti_disable instead of task flag
> ---
> arch/x86/entry/calling.h | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/arch/x86/entry/calling.h b/arch/x86/entry/calling.h
> index 2c0d3b5..5361a10 100644
> --- a/arch/x86/entry/calling.h
> +++ b/arch/x86/entry/calling.h
> @@ -229,6 +229,11 @@
>
> .macro SWITCH_TO_USER_CR3_NOSTACK scratch_reg:req scratch_reg2:req
> ALTERNATIVE "jmp .Lend_\@", "", X86_FEATURE_PTI
> +
> + /* The "pti_disable" mm attribute is mirrored into this per-cpu var */
> + cmpb $0, PER_CPU_VAR(pti_disable)
> + jne .Lend_\@
Could you please do this small change for future iterations:
s/per-cpu
/per-CPU
... to make the spelling more consistent with the rest of the code base?
Thanks,
Ingo
next prev parent reply other threads:[~2018-01-10 7:15 UTC|newest]
Thread overview: 104+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-01-09 12:56 [RFC PATCH v2 0/6] Per process PTI activation Willy Tarreau
2018-01-09 12:56 ` [RFC PATCH v2 1/6] x86/mm: add a pti_disable entry in mm_context_t Willy Tarreau
2018-01-09 12:56 ` [RFC PATCH v2 2/6] x86/arch_prctl: add ARCH_GET_NOPTI and ARCH_SET_NOPTI to enable/disable PTI Willy Tarreau
2018-01-09 14:17 ` Borislav Petkov
2018-01-09 14:36 ` Willy Tarreau
2018-01-09 14:51 ` Borislav Petkov
2018-01-09 14:54 ` Willy Tarreau
2018-01-09 21:26 ` Andy Lutomirski
2018-01-09 21:29 ` Borislav Petkov
2018-01-09 21:32 ` Willy Tarreau
2018-01-09 21:46 ` Borislav Petkov
2018-01-09 22:06 ` Willy Tarreau
2018-01-09 22:20 ` Borislav Petkov
2018-01-09 22:29 ` Dave Hansen
2018-01-09 22:40 ` Willy Tarreau
2018-01-10 14:42 ` Borislav Petkov
2018-01-10 15:39 ` Willy Tarreau
2018-01-10 16:09 ` Borislav Petkov
2018-01-10 16:19 ` Willy Tarreau
2018-01-10 17:28 ` Borislav Petkov
2018-01-10 7:31 ` Ingo Molnar
2018-01-10 7:37 ` Willy Tarreau
2018-01-10 7:59 ` Ingo Molnar
2018-01-09 23:53 ` Andy Lutomirski
2018-01-10 4:25 ` Willy Tarreau
2018-01-10 7:25 ` Ingo Molnar
2018-01-10 14:45 ` Borislav Petkov
2018-01-10 15:43 ` Willy Tarreau
2018-01-10 15:45 ` Ingo Molnar
2018-01-09 21:34 ` Kees Cook
2018-01-09 21:41 ` Willy Tarreau
2018-01-09 21:50 ` Kees Cook
2018-01-09 22:03 ` Willy Tarreau
2018-01-10 7:13 ` Ingo Molnar
2018-01-12 15:03 ` David Laight
2018-01-12 15:06 ` Willy Tarreau
2018-01-09 12:56 ` [RFC PATCH v2 3/6] x86/pti: add a per-cpu variable pti_disable Willy Tarreau
2018-01-10 7:19 ` Ingo Molnar
2018-01-10 7:29 ` Willy Tarreau
2018-01-10 8:01 ` Ingo Molnar
2018-01-10 8:50 ` Willy Tarreau
2018-01-10 8:59 ` Ingo Molnar
2018-01-10 9:00 ` Willy Tarreau
2018-01-09 12:56 ` [RFC PATCH v2 4/6] x86/pti: don't mark the user PGD with _PAGE_NX Willy Tarreau
2018-01-09 12:56 ` [RFC PATCH v2 5/6] x86/entry/pti: avoid setting CR3 when it's already correct Willy Tarreau
2018-01-10 7:16 ` Ingo Molnar
2018-01-10 7:18 ` Willy Tarreau
2018-01-10 20:29 ` Dave Hansen
2018-01-11 6:46 ` Willy Tarreau
2018-01-09 12:56 ` [RFC PATCH v2 6/6] x86/entry/pti: don't switch PGD on when pti_disable is set Willy Tarreau
2018-01-10 7:15 ` Ingo Molnar [this message]
2018-01-10 7:23 ` Willy Tarreau
2018-01-10 8:22 ` Peter Zijlstra
2018-01-10 9:11 ` Willy Tarreau
2018-01-10 19:21 ` Andy Lutomirski
2018-01-10 19:39 ` Willy Tarreau
2018-01-10 19:44 ` Andy Lutomirski
2018-01-10 19:50 ` Linus Torvalds
2018-01-10 20:04 ` Andy Lutomirski
2018-01-11 6:42 ` Willy Tarreau
2018-01-11 15:29 ` Dave Hansen
2018-01-11 15:44 ` Willy Tarreau
2018-01-11 15:51 ` Dave Hansen
2018-01-11 17:02 ` Andy Lutomirski
2018-01-11 18:21 ` Alexei Starovoitov
2018-01-11 18:30 ` Dave Hansen
2018-01-11 18:32 ` Josh Poimboeuf
2018-01-11 18:36 ` Linus Torvalds
2018-01-11 18:38 ` Dave Hansen
2018-01-11 18:51 ` Linus Torvalds
2018-01-11 18:57 ` Dave Hansen
2018-01-11 19:05 ` Josh Poimboeuf
2018-01-11 19:07 ` Borislav Petkov
2018-01-11 19:17 ` Dave Hansen
2018-01-11 19:19 ` Olivier Galibert
2018-01-11 19:26 ` Josh Poimboeuf
2018-01-11 19:34 ` Alan Cox
2018-01-11 21:23 ` Willy Tarreau
2018-01-11 21:28 ` Linus Torvalds
2018-01-11 22:06 ` Willy Tarreau
2018-01-12 16:37 ` David Laight
2018-01-11 19:12 ` Linus Torvalds
2018-01-11 19:38 ` Alexei Starovoitov
2018-01-11 19:11 ` Willy Tarreau
2018-01-11 20:00 ` Dave Hansen
2018-01-11 17:09 ` Andy Lutomirski
2018-01-11 17:40 ` Willy Tarreau
2018-01-11 17:53 ` Andy Lutomirski
2018-01-11 18:05 ` Willy Tarreau
2018-01-11 18:15 ` Dave Hansen
2018-01-11 18:31 ` Linus Torvalds
2018-01-11 18:25 ` Linus Torvalds
2018-01-11 18:26 ` Linus Torvalds
2018-01-11 19:33 ` Andy Lutomirski
2018-01-12 20:22 ` Ingo Molnar
2018-01-12 21:18 ` Andy Lutomirski
2018-01-12 21:54 ` Willy Tarreau
2018-01-11 21:59 ` Willy Tarreau
2018-01-12 16:27 ` David Laight
2018-01-12 17:55 ` Linus Torvalds
2018-01-12 19:36 ` Willy Tarreau
2018-01-11 18:35 ` Dave Hansen
2018-01-11 21:49 ` Willy Tarreau
-- strict thread matches above, loose matches on Subject: below --
2018-01-11 23:11 Alexey Dobriyan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180110071510.nfice2pjtuylqirc@gmail.com \
--to=mingo@kernel.org \
--cc=bp@alien8.de \
--cc=brgerst@gmail.com \
--cc=dave.hansen@linux.intel.com \
--cc=gregkh@linuxfoundation.org \
--cc=hpa@zytor.com \
--cc=jpoimboe@redhat.com \
--cc=keescook@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=peterz@infradead.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=w@1wt.eu \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox