From: Alexandre Chartre <alexandre.chartre@oracle.com>
To: David Laight <David.Laight@ACULAB.COM>, Borislav Petkov <bp@alien8.de>
Cc: "tglx@linutronix.de" <tglx@linutronix.de>,
"mingo@redhat.com" <mingo@redhat.com>,
"hpa@zytor.com" <hpa@zytor.com>,
"x86@kernel.org" <x86@kernel.org>,
"dave.hansen@linux.intel.com" <dave.hansen@linux.intel.com>,
"luto@kernel.org" <luto@kernel.org>,
"peterz@infradead.org" <peterz@infradead.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"thomas.lendacky@amd.com" <thomas.lendacky@amd.com>,
"jroedel@suse.de" <jroedel@suse.de>,
"konrad.wilk@oracle.com" <konrad.wilk@oracle.com>,
"jan.setjeeilers@oracle.com" <jan.setjeeilers@oracle.com>,
"junaids@google.com" <junaids@google.com>,
"oweisse@google.com" <oweisse@google.com>,
"rppt@linux.vnet.ibm.com" <rppt@linux.vnet.ibm.com>,
"graf@amazon.de" <graf@amazon.de>,
"mgross@linux.intel.com" <mgross@linux.intel.com>,
"kuzuno@gmail.com" <kuzuno@gmail.com>
Subject: Re: [RFC][PATCH v2 00/21] x86/pti: Defer CR3 switch to C code
Date: Wed, 18 Nov 2020 11:29:52 +0100 [thread overview]
Message-ID: <0bedae59-5397-9cae-3c2a-66bc376f5616@oracle.com> (raw)
In-Reply-To: <ce8d862f498042d1bd7a6e8a071f06bf@AcuMS.aculab.com>
On 11/18/20 10:30 AM, David Laight wrote:
> From: Alexandre Chartre
>> Sent: 18 November 2020 07:42
>>
>>
>> On 11/17/20 10:26 PM, Borislav Petkov wrote:
>>> On Tue, Nov 17, 2020 at 07:12:07PM +0100, Alexandre Chartre wrote:
>>>> Some benchmarks are available, in particular from phoronix:
>>>
>>> What I was expecting was benchmarks *you* have run which show that
>>> perf penalty, not something one can find quickly on the internet and
>>> something one cannot always reproduce her-/himself.
>>>
>>> You do know that presenting convincing numbers with a patchset greatly
>>> improves its chances of getting it upstreamed, right?
>>>
>>
>> Well, it looks like I wrongfully assume that KPTI was a well known performance
>> overhead since it was introduced (because it adds extra page-table switches),
>> but you are right I should be presenting my own numbers.
>
> IIRC the penalty comes from the page table switch.
> Doing it at a different time is unlikely to make much difference.
>
Correct, this RFC is not changing the overhead. However, it is a step forward
for being able to execute some selected syscalls or interrupt handlers without
switching to the kernel page-table. The next step would be to identify and add
the necessary mapping to the user page-table so that specified syscalls can be
executed without switching the page-table.
> For some workloads the penalty is massive - getting on for 50%.
> We are still using old kernels on AWS.
>
Here are some micro benchmarks of the getppid and getpid syscalls which highlight
the PTI overhead. This uses the kernel tools/perf command, and the getpid command
from libMICRO (https://github.com/redhat-performance/libMicro):
system running 5.10-rc4 booted with nopti:
------------------------------------------
# perf bench syscall basic
# Running 'syscall/basic' benchmark:
# Executed 10000000 getppid() calls
Total time: 0.792 [sec]
0.079223 usecs/op
12622549 ops/sec
# getpid -B 100000
prc thr usecs/call samples errors cnt/samp
getpid 1 1 0.08029 102 0 100000
We can see that getpid and getppid syscall have the same execution
time around 0.08 usecs. These syscalls are very small and just return
a value, so the time is mostly spent entering/exiting the kernel.
same system booted with pti:
----------------------------
# perf bench syscall basic
# Running 'syscall/basic' benchmark:
# Executed 10000000 getppid() calls
Total time: 2.025 [sec]
0.202527 usecs/op
4937605 ops/sec
# getpid -B 100000
prc thr usecs/call samples errors cnt/samp
getpid 1 1 0.20241 102 0 100000
With PTI, the execution time jumps to 0.20 usecs (+0.12 usecs = +150%).
That's a very extreme case because these are very small syscalls, and
in that case the overhead to switch page-tables is significant compared
to the execution time of the syscall.
So with an overhead of +0.12 usecs per syscall, the PTI impact is significant
with workload which uses a lot of short syscalls. But if you use longer syscalls,
for example with an average execution time of 2.0 usecs per syscall then you
have a lower overhead of 6%.
alex.
next prev parent reply other threads:[~2020-11-18 10:28 UTC|newest]
Thread overview: 69+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-16 14:47 [RFC][PATCH v2 00/21] x86/pti: Defer CR3 switch to C code Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 01/21] x86/syscall: Add wrapper for invoking syscall function Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 02/21] x86/entry: Update asm_call_on_stack to support more function arguments Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 03/21] x86/entry: Consolidate IST entry from userspace Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 04/21] x86/sev-es: Define a setup stack function for the VC idtentry Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 05/21] x86/entry: Implement ret_from_fork body with C code Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 06/21] x86/pti: Provide C variants of PTI switch CR3 macros Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 07/21] x86/entry: Fill ESPFIX stack using C code Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 08/21] x86/pti: Introduce per-task PTI trampoline stack Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 09/21] x86/pti: Function to clone page-table entries from a specified mm Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 10/21] x86/pti: Function to map per-cpu page-table entry Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 11/21] x86/pti: Extend PTI user mappings Alexandre Chartre
2020-11-16 19:48 ` Andy Lutomirski
2020-11-16 20:21 ` Alexandre Chartre
2020-11-16 23:06 ` Andy Lutomirski
2020-11-17 8:42 ` Alexandre Chartre
2020-11-17 15:49 ` Andy Lutomirski
2020-11-19 19:15 ` Thomas Gleixner
2020-11-16 14:47 ` [RFC][PATCH v2 12/21] x86/pti: Use PTI stack instead of trampoline stack Alexandre Chartre
2020-11-16 16:57 ` Andy Lutomirski
2020-11-16 18:10 ` Alexandre Chartre
2020-11-16 18:34 ` Andy Lutomirski
2020-11-16 19:37 ` Alexandre Chartre
2020-11-17 15:09 ` Alexandre Chartre
2020-11-17 15:52 ` Andy Lutomirski
2020-11-17 17:01 ` Alexandre Chartre
2020-11-19 1:49 ` Andy Lutomirski
2020-11-19 8:05 ` Alexandre Chartre
2020-11-19 12:06 ` Alexandre Chartre
2020-11-19 16:06 ` Andy Lutomirski
2020-11-19 17:02 ` Alexandre Chartre
2020-11-16 21:24 ` David Laight
2020-11-17 8:27 ` Alexandre Chartre
2020-11-19 19:10 ` Thomas Gleixner
2020-11-19 19:55 ` Alexandre Chartre
2020-11-19 21:20 ` Thomas Gleixner
2020-11-24 7:20 ` [x86/pti] 5da9e742d1: PANIC:double_fault kernel test robot
2020-11-16 14:47 ` [RFC][PATCH v2 13/21] x86/pti: Execute syscall functions on the kernel stack Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 14/21] x86/pti: Execute IDT handlers " Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 15/21] x86/pti: Execute IDT handlers with error code " Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 16/21] x86/pti: Execute system vector handlers " Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 17/21] x86/pti: Execute page fault handler " Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 18/21] x86/pti: Execute NMI " Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 19/21] x86/pti: Defer CR3 switch to C code for IST entries Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 20/21] x86/pti: Defer CR3 switch to C code for non-IST and syscall entries Alexandre Chartre
2020-11-16 14:47 ` [RFC][PATCH v2 21/21] x86/pti: Use a different stack canary with the user and kernel page-table Alexandre Chartre
2020-11-16 16:56 ` Andy Lutomirski
2020-11-16 18:34 ` Alexandre Chartre
2020-11-16 20:17 ` [RFC][PATCH v2 00/21] x86/pti: Defer CR3 switch to C code Borislav Petkov
2020-11-17 7:56 ` Alexandre Chartre
2020-11-17 16:55 ` Borislav Petkov
2020-11-17 18:12 ` Alexandre Chartre
2020-11-17 18:28 ` Borislav Petkov
2020-11-17 19:02 ` Alexandre Chartre
2020-11-17 21:23 ` Borislav Petkov
2020-11-18 7:08 ` Alexandre Chartre
2020-11-17 21:26 ` Borislav Petkov
2020-11-18 7:41 ` Alexandre Chartre
2020-11-18 9:30 ` David Laight
2020-11-18 10:29 ` Alexandre Chartre [this message]
2020-11-18 13:22 ` David Laight
2020-11-18 17:15 ` Alexandre Chartre
2020-11-18 11:29 ` Borislav Petkov
2020-11-18 19:37 ` Alexandre Chartre
2020-11-16 20:24 ` Borislav Petkov
2020-11-17 8:19 ` Alexandre Chartre
2020-11-17 17:07 ` Borislav Petkov
2020-11-17 18:24 ` Alexandre Chartre
2020-11-19 19:32 ` Thomas Gleixner
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=0bedae59-5397-9cae-3c2a-66bc376f5616@oracle.com \
--to=alexandre.chartre@oracle.com \
--cc=David.Laight@ACULAB.COM \
--cc=bp@alien8.de \
--cc=dave.hansen@linux.intel.com \
--cc=graf@amazon.de \
--cc=hpa@zytor.com \
--cc=jan.setjeeilers@oracle.com \
--cc=jroedel@suse.de \
--cc=junaids@google.com \
--cc=konrad.wilk@oracle.com \
--cc=kuzuno@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=mgross@linux.intel.com \
--cc=mingo@redhat.com \
--cc=oweisse@google.com \
--cc=peterz@infradead.org \
--cc=rppt@linux.vnet.ibm.com \
--cc=tglx@linutronix.de \
--cc=thomas.lendacky@amd.com \
--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;
as well as URLs for NNTP newsgroup(s).