From: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
To: Linus Torvalds <torvalds@linux-foundation.org>,
x86@kernel.org, Thomas Gleixner <tglx@linutronix.de>,
Ingo Molnar <mingo@redhat.com>, "H. Peter Anvin" <hpa@zytor.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Andy Lutomirski <luto@amacapital.net>,
Dmitry Safonov <dsafonov@virtuozzo.com>,
Cyrill Gorcunov <gorcunov@openvz.org>,
Borislav Petkov <bp@suse.de>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Subject: [PATCHv5 19/19] x86/mm: Offset boot-time paging mode switching cost
Date: Mon, 21 Aug 2017 18:29:16 +0300 [thread overview]
Message-ID: <20170821152916.40124-20-kirill.shutemov@linux.intel.com> (raw)
In-Reply-To: <20170821152916.40124-1-kirill.shutemov@linux.intel.com>
By this point we have functioning boot-time switching between 4- and
5-level paging mode. But naive approach comes with cost.
Numbers below are for kernel build, allmodconfig, 5 times.
CONFIG_X86_5LEVEL=n:
Performance counter stats for 'sh -c make -j100 -B -k >/dev/null' (5 runs):
17308719.892691 task-clock:u (msec) # 26.772 CPUs utilized ( +- 0.11% )
0 context-switches:u # 0.000 K/sec
0 cpu-migrations:u # 0.000 K/sec
331,993,164 page-faults:u # 0.019 M/sec ( +- 0.01% )
43,614,978,867,455 cycles:u # 2.520 GHz ( +- 0.01% )
39,371,534,575,126 stalled-cycles-frontend:u # 90.27% frontend cycles idle ( +- 0.09% )
28,363,350,152,428 instructions:u # 0.65 insn per cycle
# 1.39 stalled cycles per insn ( +- 0.00% )
6,316,784,066,413 branches:u # 364.948 M/sec ( +- 0.00% )
250,808,144,781 branch-misses:u # 3.97% of all branches ( +- 0.01% )
646.531974142 seconds time elapsed ( +- 1.15% )
CONFIG_X86_5LEVEL=y:
Performance counter stats for 'sh -c make -j100 -B -k >/dev/null' (5 runs):
17411536.780625 task-clock:u (msec) # 26.426 CPUs utilized ( +- 0.10% )
0 context-switches:u # 0.000 K/sec
0 cpu-migrations:u # 0.000 K/sec
331,868,663 page-faults:u # 0.019 M/sec ( +- 0.01% )
43,865,909,056,301 cycles:u # 2.519 GHz ( +- 0.01% )
39,740,130,365,581 stalled-cycles-frontend:u # 90.59% frontend cycles idle ( +- 0.05% )
28,363,358,997,959 instructions:u # 0.65 insn per cycle
# 1.40 stalled cycles per insn ( +- 0.00% )
6,316,784,937,460 branches:u # 362.793 M/sec ( +- 0.00% )
251,531,919,485 branch-misses:u # 3.98% of all branches ( +- 0.00% )
658.886307752 seconds time elapsed ( +- 0.92% )
The patch tries to fix the performance regression by using
!cpu_feature_enabled(X86_FEATURE_LA57) instead of pgtable_l5_enabled in
all hot code paths. These will statically patch the target code for
additional performance.
CONFIG_X86_5LEVEL=y + the patch:
Performance counter stats for 'sh -c make -j100 -B -k >/dev/null' (5 runs):
17381990.268506 task-clock:u (msec) # 26.907 CPUs utilized ( +- 0.19% )
0 context-switches:u # 0.000 K/sec
0 cpu-migrations:u # 0.000 K/sec
331,862,625 page-faults:u # 0.019 M/sec ( +- 0.01% )
43,697,726,320,051 cycles:u # 2.514 GHz ( +- 0.03% )
39,480,408,690,401 stalled-cycles-frontend:u # 90.35% frontend cycles idle ( +- 0.05% )
28,363,394,221,388 instructions:u # 0.65 insn per cycle
# 1.39 stalled cycles per insn ( +- 0.00% )
6,316,794,985,573 branches:u # 363.410 M/sec ( +- 0.00% )
251,013,232,547 branch-misses:u # 3.97% of all branches ( +- 0.01% )
645.991174661 seconds time elapsed ( +- 1.19% )
Unfortunately, this approach doesn't help with text size:
vmlinux.before .text size: 9798404
vmlinux.after .text size: 9802566
The .text section is increased by about 4k. Not sure if we can do anything
about this.
Signed-eef-by: Kirill A. Shuemov <kirill.shutemov@linux.intel.com>
---
arch/x86/boot/compressed/misc.h | 5 +++++
arch/x86/entry/entry_64.S | 11 ++---------
arch/x86/include/asm/pgtable_64_types.h | 5 ++++-
arch/x86/kernel/head64.c | 9 +++++++--
arch/x86/kernel/head_64.S | 2 +-
arch/x86/mm/kasan_init_64.c | 6 ++++++
6 files changed, 25 insertions(+), 13 deletions(-)
diff --git a/arch/x86/boot/compressed/misc.h b/arch/x86/boot/compressed/misc.h
index 766a5211f827..5604f08aa405 100644
--- a/arch/x86/boot/compressed/misc.h
+++ b/arch/x86/boot/compressed/misc.h
@@ -11,6 +11,11 @@
#undef CONFIG_PARAVIRT_SPINLOCKS
#undef CONFIG_KASAN
+#ifdef CONFIG_X86_5LEVEL
+/* cpu_feature_enabled() cannot be used that early */
+#define pgtable_l5_enabled __pgtable_l5_enabled
+#endif
+
#include <linux/linkage.h>
#include <linux/screen_info.h>
#include <linux/elf.h>
diff --git a/arch/x86/entry/entry_64.S b/arch/x86/entry/entry_64.S
index 1dd5710315ef..702a1feb4991 100644
--- a/arch/x86/entry/entry_64.S
+++ b/arch/x86/entry/entry_64.S
@@ -274,15 +274,8 @@ return_from_SYSCALL_64:
* depending on paging mode) in the address.
*/
#ifdef CONFIG_X86_5LEVEL
- testl $1, pgtable_l5_enabled(%rip)
- jz 1f
- shl $(64 - 57), %rcx
- sar $(64 - 57), %rcx
- jmp 2f
-1:
- shl $(64 - 48), %rcx
- sar $(64 - 48), %rcx
-2:
+ ALTERNATIVE "shl $(64 - 48), %rcx; sar $(64 - 48), %rcx", \
+ "shl $(64 - 57), %rcx; sar $(64 - 57), %rcx", X86_FEATURE_LA57
#else
shl $(64 - (__VIRTUAL_MASK_SHIFT+1)), %rcx
sar $(64 - (__VIRTUAL_MASK_SHIFT+1)), %rcx
diff --git a/arch/x86/include/asm/pgtable_64_types.h b/arch/x86/include/asm/pgtable_64_types.h
index fa9f8b6592fa..0efb46fa1052 100644
--- a/arch/x86/include/asm/pgtable_64_types.h
+++ b/arch/x86/include/asm/pgtable_64_types.h
@@ -20,7 +20,10 @@ typedef unsigned long pgprotval_t;
typedef struct { pteval_t pte; } pte_t;
#ifdef CONFIG_X86_5LEVEL
-extern unsigned int pgtable_l5_enabled;
+extern unsigned int __pgtable_l5_enabled;
+#ifndef pgtable_l5_enabled
+#define pgtable_l5_enabled (cpu_feature_enabled(X86_FEATURE_LA57))
+#endif
#else
#define pgtable_l5_enabled 0
#endif
diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index 8a44b3d06261..c276785f257b 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -31,6 +31,11 @@
#include <asm/microcode.h>
#include <asm/kasan.h>
+#ifdef CONFIG_X86_5LEVEL
+#undef pgtable_l5_enabled
+#define pgtable_l5_enabled __pgtable_l5_enabled
+#endif
+
/*
* Manage page tables very early on.
*/
@@ -39,8 +44,8 @@ static unsigned int __initdata next_early_pgt;
pmdval_t early_pmd_flags = __PAGE_KERNEL_LARGE & ~(_PAGE_GLOBAL | _PAGE_NX);
#ifdef CONFIG_X86_5LEVEL
-unsigned int pgtable_l5_enabled __read_mostly = 0;
-EXPORT_SYMBOL(pgtable_l5_enabled);
+unsigned int __pgtable_l5_enabled __read_mostly = 0;
+EXPORT_SYMBOL(__pgtable_l5_enabled);
unsigned int pgdir_shift __read_mostly = 39;
EXPORT_SYMBOL(pgdir_shift);
unsigned int ptrs_per_p4d __read_mostly = 1;
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index e137f2665fc2..8a1fe9b63c03 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -121,7 +121,7 @@ ENTRY(secondary_startup_64)
/* Enable PAE mode, PGE and LA57 */
movl $(X86_CR4_PAE | X86_CR4_PGE), %ecx
#ifdef CONFIG_X86_5LEVEL
- testl $1, pgtable_l5_enabled(%rip)
+ testl $1, __pgtable_l5_enabled(%rip)
jz 1f
orl $X86_CR4_LA57, %ecx
1:
diff --git a/arch/x86/mm/kasan_init_64.c b/arch/x86/mm/kasan_init_64.c
index 9173ce1feba0..230e4ea1d3ae 100644
--- a/arch/x86/mm/kasan_init_64.c
+++ b/arch/x86/mm/kasan_init_64.c
@@ -1,5 +1,11 @@
#define DISABLE_BRANCH_PROFILING
#define pr_fmt(fmt) "kasan: " fmt
+
+#ifdef CONFIG_X86_5LEVEL
+/* Too early to use cpu_feature_enabled() */
+#define pgtable_l5_enabled __pgtable_l5_enabled
+#endif
+
#include <linux/bootmem.h>
#include <linux/kasan.h>
#include <linux/kdebug.h>
--
2.14.1
--
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: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
prev parent reply other threads:[~2017-08-21 15:29 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-08-21 15:28 [PATCHv5 00/19] Boot-time switching between 4- and 5-level paging Kirill A. Shutemov
2017-08-21 15:28 ` [PATCHv5 01/19] mm/sparsemem: Allocate mem_section at runtime for SPARSEMEM_EXTREME Kirill A. Shutemov
2017-08-22 16:28 ` Borislav Petkov
2017-08-22 17:50 ` Kirill A. Shutemov
2017-08-21 15:28 ` [PATCHv5 02/19] mm/zsmalloc: Prepare to variable MAX_PHYSMEM_BITS Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 03/19] x86/kasan: Use the same shadow offset for 4- and 5-level paging Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 04/19] x86/xen: Provide pre-built page tables only for XEN_PV and XEN_PVH Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 05/19] x86/xen: Drop 5-level paging support code from XEN_PV code Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 06/19] x86/boot/compressed/64: Detect and handle 5-level paging at boot-time Kirill A. Shutemov
2017-08-27 11:29 ` Cyrill Gorcunov
2017-09-04 13:02 ` Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 07/19] x86/mm: Make virtual memory layout movable for CONFIG_X86_5LEVEL Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 08/19] x86/mm: Make PGDIR_SHIFT and PTRS_PER_P4D variable Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 09/19] x86/mm: Make MAX_PHYSADDR_BITS and MAX_PHYSMEM_BITS dynamic Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 10/19] x86/mm: Make __PHYSICAL_MASK_SHIFT and __VIRTUAL_MASK_SHIFT dynamic Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 11/19] x86/mm: Make STACK_TOP_MAX dynamic Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 12/19] x86/mm: Adjust virtual address space layout in early boot Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 13/19] x86/mm: Make early boot code support boot-time switching of paging modes Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 14/19] x86/mm: Fold p4d page table layer at runtime Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 15/19] x86/mm: Replace compile-time checks for 5-level with runtime-time Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 16/19] x86/mm: Allow to boot without la57 if CONFIG_X86_5LEVEL=y Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 17/19] x86/xen: Allow XEN_PV and XEN_PVH to be enabled with X86_5LEVEL Kirill A. Shutemov
2017-08-21 15:29 ` [PATCHv5 18/19] x86/mm: Redefine some of page table helpers as macros Kirill A. Shutemov
2017-08-21 15:29 ` Kirill A. Shutemov [this message]
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=20170821152916.40124-20-kirill.shutemov@linux.intel.com \
--to=kirill.shutemov@linux.intel.com \
--cc=akpm@linux-foundation.org \
--cc=bp@suse.de \
--cc=dsafonov@virtuozzo.com \
--cc=gorcunov@openvz.org \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=luto@amacapital.net \
--cc=mingo@redhat.com \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--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).