From: Ingo Molnar <mingo@elte.hu>
To: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Christoph Lameter <cl@linux.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
Eric Dumazet <eric.dumazet@gmail.com>,
Thomas Gleixner <tglx@linutronix.de>,
akpm@linux-foundation.org, tj@kernel.org, npiggin@kernel.dk,
rientjes@google.com, linux-kernel@vger.kernel.org,
linux-mm@kvack.org
Subject: Re: [PATCH] slub: Disable the lockless allocator
Date: Mon, 28 Mar 2011 08:19:29 +0200 [thread overview]
Message-ID: <20110328061929.GA24328@elte.hu> (raw)
In-Reply-To: <4D9026C8.6060905@cs.helsinki.fi>
* Pekka Enberg <penberg@cs.helsinki.fi> wrote:
> On 3/27/11 4:57 AM, Christoph Lameter wrote:
> >But then the same fix must also be used in the asm code or the fallback
> >(turns out that the fallback is always used in kmem_cache_init since
> >the instruction patching comes later).
> >
> >Patch boots fine both in UP and SMP mode
> >
> >
> >
> >
> >Subject: percpu: Omit segment prefix in the UP case for cmpxchg_double
> >
> >Omit the segment prefix in the UP case. GS is not used then
> >and we will generate segfaults if cmpxchg16b is used otherwise.
> >
> >Signed-off-by: Linus Torvalds<torvalds@linux-foundation.org>
> >Signed-off-by: Christoph Lameter<cl@linux.com>
> >
> > arch/x86/include/asm/percpu.h | 10 ++++++----
> > 1 files changed, 6 insertions(+), 4 deletions(-)
> >
> >Index: linux-2.6/arch/x86/include/asm/percpu.h
> >===================================================================
> >--- linux-2.6.orig/arch/x86/include/asm/percpu.h 2011-03-26 20:43:03.994089001 -0500
> >+++ linux-2.6/arch/x86/include/asm/percpu.h 2011-03-26 20:43:22.414089004 -0500
> >@@ -45,7 +45,7 @@
> > #include<linux/stringify.h>
> >
> > #ifdef CONFIG_SMP
> >-#define __percpu_arg(x) "%%"__stringify(__percpu_seg)":%P" #x
> >+#define __percpu_prefix "%%"__stringify(__percpu_seg)":"
> > #define __my_cpu_offset percpu_read(this_cpu_off)
> >
> > /*
> >@@ -62,9 +62,11 @@
> > (typeof(*(ptr)) __kernel __force *)tcp_ptr__; \
> > })
> > #else
> >-#define __percpu_arg(x) "%P" #x
> >+#define __percpu_prefix ""
> > #endif
> >
> >+#define __percpu_arg(x) __percpu_prefix "%P" #x
> >+
> > /*
> > * Initialized pointers to per-cpu variables needed for the boot
> > * processor need to use these macros to get the proper address
> >@@ -516,11 +518,11 @@
> > typeof(o2) __n2 = n2; \
> > typeof(o2) __dummy; \
> > alternative_io("call this_cpu_cmpxchg16b_emu\n\t" P6_NOP4, \
> >- "cmpxchg16b %%gs:(%%rsi)\n\tsetz %0\n\t", \
> >+ "cmpxchg16b " __percpu_prefix "(%%rsi)\n\tsetz %0\n\t", \
> > X86_FEATURE_CX16, \
> > ASM_OUTPUT2("=a"(__ret), "=d"(__dummy)), \
> > "S" (&pcp1), "b"(__n1), "c"(__n2), \
> >- "a"(__o1), "d"(__o2)); \
> >+ "a"(__o1), "d"(__o2) : "memory"); \
> > __ret; \
> > })
> >
> >Index: linux-2.6/arch/x86/lib/cmpxchg16b_emu.S
> >===================================================================
> >--- linux-2.6.orig/arch/x86/lib/cmpxchg16b_emu.S 2011-03-26 20:43:57.384089004 -0500
> >+++ linux-2.6/arch/x86/lib/cmpxchg16b_emu.S 2011-03-26 20:48:42.684088999 -0500
> >@@ -10,6 +10,12 @@
> > #include<asm/frame.h>
> > #include<asm/dwarf2.h>
> >
> >+#ifdef CONFIG_SMP
> >+#define SEG_PREFIX %gs:
> >+#else
> >+#define SEG_PREFIX
> >+#endif
> >+
> > .text
> >
> > /*
> >@@ -37,13 +43,13 @@
> > pushf
> > cli
> >
> >- cmpq %gs:(%rsi), %rax
> >+ cmpq SEG_PREFIX(%rsi), %rax
> > jne not_same
> >- cmpq %gs:8(%rsi), %rdx
> >+ cmpq SEG_PREFIX 8(%rsi), %rdx
> > jne not_same
> >
> >- movq %rbx, %gs:(%rsi)
> >- movq %rcx, %gs:8(%rsi)
> >+ movq %rbx, SEG_PREFIX(%rsi)
> >+ movq %rcx, SEG_PREFIX 8(%rsi)
> >
> > popf
> > mov $1, %al
>
> Tejun, does this look good to you as well? I think it should go
> through the percpu tree. It's needed to fix a boot crash with
> lockless SLUB fastpaths enabled.
AFAICS Linus applied it already:
d7c3f8cee81f: percpu: Omit segment prefix in the UP case for cmpxchg_double
Thanks,
Ingo
WARNING: multiple messages have this Message-ID (diff)
From: Ingo Molnar <mingo@elte.hu>
To: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Christoph Lameter <cl@linux.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
Eric Dumazet <eric.dumazet@gmail.com>,
Thomas Gleixner <tglx@linutronix.de>,
akpm@linux-foundation.org, tj@kernel.org, npiggin@kernel.dk,
rientjes@google.com, linux-kernel@vger.kernel.org,
linux-mm@kvack.org
Subject: Re: [PATCH] slub: Disable the lockless allocator
Date: Mon, 28 Mar 2011 08:19:29 +0200 [thread overview]
Message-ID: <20110328061929.GA24328@elte.hu> (raw)
In-Reply-To: <4D9026C8.6060905@cs.helsinki.fi>
* Pekka Enberg <penberg@cs.helsinki.fi> wrote:
> On 3/27/11 4:57 AM, Christoph Lameter wrote:
> >But then the same fix must also be used in the asm code or the fallback
> >(turns out that the fallback is always used in kmem_cache_init since
> >the instruction patching comes later).
> >
> >Patch boots fine both in UP and SMP mode
> >
> >
> >
> >
> >Subject: percpu: Omit segment prefix in the UP case for cmpxchg_double
> >
> >Omit the segment prefix in the UP case. GS is not used then
> >and we will generate segfaults if cmpxchg16b is used otherwise.
> >
> >Signed-off-by: Linus Torvalds<torvalds@linux-foundation.org>
> >Signed-off-by: Christoph Lameter<cl@linux.com>
> >
> > arch/x86/include/asm/percpu.h | 10 ++++++----
> > 1 files changed, 6 insertions(+), 4 deletions(-)
> >
> >Index: linux-2.6/arch/x86/include/asm/percpu.h
> >===================================================================
> >--- linux-2.6.orig/arch/x86/include/asm/percpu.h 2011-03-26 20:43:03.994089001 -0500
> >+++ linux-2.6/arch/x86/include/asm/percpu.h 2011-03-26 20:43:22.414089004 -0500
> >@@ -45,7 +45,7 @@
> > #include<linux/stringify.h>
> >
> > #ifdef CONFIG_SMP
> >-#define __percpu_arg(x) "%%"__stringify(__percpu_seg)":%P" #x
> >+#define __percpu_prefix "%%"__stringify(__percpu_seg)":"
> > #define __my_cpu_offset percpu_read(this_cpu_off)
> >
> > /*
> >@@ -62,9 +62,11 @@
> > (typeof(*(ptr)) __kernel __force *)tcp_ptr__; \
> > })
> > #else
> >-#define __percpu_arg(x) "%P" #x
> >+#define __percpu_prefix ""
> > #endif
> >
> >+#define __percpu_arg(x) __percpu_prefix "%P" #x
> >+
> > /*
> > * Initialized pointers to per-cpu variables needed for the boot
> > * processor need to use these macros to get the proper address
> >@@ -516,11 +518,11 @@
> > typeof(o2) __n2 = n2; \
> > typeof(o2) __dummy; \
> > alternative_io("call this_cpu_cmpxchg16b_emu\n\t" P6_NOP4, \
> >- "cmpxchg16b %%gs:(%%rsi)\n\tsetz %0\n\t", \
> >+ "cmpxchg16b " __percpu_prefix "(%%rsi)\n\tsetz %0\n\t", \
> > X86_FEATURE_CX16, \
> > ASM_OUTPUT2("=a"(__ret), "=d"(__dummy)), \
> > "S" (&pcp1), "b"(__n1), "c"(__n2), \
> >- "a"(__o1), "d"(__o2)); \
> >+ "a"(__o1), "d"(__o2) : "memory"); \
> > __ret; \
> > })
> >
> >Index: linux-2.6/arch/x86/lib/cmpxchg16b_emu.S
> >===================================================================
> >--- linux-2.6.orig/arch/x86/lib/cmpxchg16b_emu.S 2011-03-26 20:43:57.384089004 -0500
> >+++ linux-2.6/arch/x86/lib/cmpxchg16b_emu.S 2011-03-26 20:48:42.684088999 -0500
> >@@ -10,6 +10,12 @@
> > #include<asm/frame.h>
> > #include<asm/dwarf2.h>
> >
> >+#ifdef CONFIG_SMP
> >+#define SEG_PREFIX %gs:
> >+#else
> >+#define SEG_PREFIX
> >+#endif
> >+
> > .text
> >
> > /*
> >@@ -37,13 +43,13 @@
> > pushf
> > cli
> >
> >- cmpq %gs:(%rsi), %rax
> >+ cmpq SEG_PREFIX(%rsi), %rax
> > jne not_same
> >- cmpq %gs:8(%rsi), %rdx
> >+ cmpq SEG_PREFIX 8(%rsi), %rdx
> > jne not_same
> >
> >- movq %rbx, %gs:(%rsi)
> >- movq %rcx, %gs:8(%rsi)
> >+ movq %rbx, SEG_PREFIX(%rsi)
> >+ movq %rcx, SEG_PREFIX 8(%rsi)
> >
> > popf
> > mov $1, %al
>
> Tejun, does this look good to you as well? I think it should go
> through the percpu tree. It's needed to fix a boot crash with
> lockless SLUB fastpaths enabled.
AFAICS Linus applied it already:
d7c3f8cee81f: percpu: Omit segment prefix in the UP case for cmpxchg_double
Thanks,
Ingo
--
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/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2011-03-28 6:20 UTC|newest]
Thread overview: 134+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-03-22 14:40 [GIT PULL] SLAB changes for v2.6.39-rc1 Pekka Enberg
2011-03-22 14:40 ` Pekka Enberg
2011-03-24 14:21 ` Ingo Molnar
2011-03-24 14:41 ` Christoph Lameter
2011-03-24 16:50 ` Pekka Enberg
2011-03-24 17:26 ` Ingo Molnar
2011-03-24 17:26 ` Ingo Molnar
2011-03-24 17:43 ` Christoph Lameter
2011-03-24 17:43 ` Christoph Lameter
2011-03-24 17:56 ` Pekka Enberg
2011-03-24 17:56 ` Pekka Enberg
2011-03-24 18:01 ` Christoph Lameter
2011-03-24 18:01 ` Christoph Lameter
2011-03-24 18:06 ` Pekka Enberg
2011-03-24 18:06 ` Pekka Enberg
2011-03-24 18:15 ` Christoph Lameter
2011-03-24 18:15 ` Christoph Lameter
2011-03-24 18:20 ` Eric Dumazet
2011-03-24 18:20 ` Eric Dumazet
2011-03-24 18:47 ` Christoph Lameter
2011-03-24 18:47 ` Christoph Lameter
2011-03-24 18:51 ` Eric Dumazet
2011-03-24 18:51 ` Eric Dumazet
2011-03-24 19:02 ` Christoph Lameter
2011-03-24 19:04 ` Pekka Enberg
2011-03-24 19:04 ` Pekka Enberg
2011-03-24 18:55 ` Pekka Enberg
2011-03-24 18:55 ` Pekka Enberg
2011-03-24 18:59 ` Ingo Molnar
2011-03-24 18:59 ` Ingo Molnar
2011-03-24 19:03 ` Pekka Enberg
2011-03-24 19:03 ` Pekka Enberg
2011-03-24 19:05 ` Christoph Lameter
2011-03-24 19:05 ` Christoph Lameter
2011-03-24 19:14 ` Pekka Enberg
2011-03-24 19:14 ` Pekka Enberg
2011-03-24 19:36 ` Ingo Molnar
2011-03-24 19:36 ` Ingo Molnar
2011-03-24 19:41 ` Pekka Enberg
2011-03-24 19:41 ` Pekka Enberg
2011-03-24 19:51 ` Christoph Lameter
2011-03-24 19:51 ` Christoph Lameter
2011-03-24 20:01 ` Ingo Molnar
2011-03-24 20:01 ` Ingo Molnar
2011-03-24 20:08 ` Eric Dumazet
2011-03-24 20:08 ` Eric Dumazet
2011-03-24 20:43 ` Christoph Lameter
2011-03-24 20:43 ` Christoph Lameter
2011-03-25 15:13 ` Tejun Heo
2011-03-25 15:13 ` Tejun Heo
2011-03-25 15:45 ` Christoph Lameter
2011-03-25 15:45 ` Christoph Lameter
2011-03-27 9:59 ` Avi Kivity
2011-03-27 9:59 ` Avi Kivity
2011-03-24 19:02 ` Christoph Lameter
2011-03-24 19:02 ` Christoph Lameter
2011-03-27 9:57 ` Avi Kivity
2011-03-27 9:57 ` Avi Kivity
2011-03-24 18:13 ` Pekka Enberg
2011-03-24 18:13 ` Pekka Enberg
2011-03-24 19:27 ` Ingo Molnar
2011-03-24 19:27 ` Ingo Molnar
2011-03-24 19:29 ` Pekka Enberg
2011-03-24 19:29 ` Pekka Enberg
2011-03-24 18:52 ` [boot crash #2] " Ingo Molnar
2011-03-24 18:52 ` Ingo Molnar
2011-03-24 19:10 ` Thomas Gleixner
2011-03-24 19:10 ` Thomas Gleixner
2011-03-24 19:22 ` Ingo Molnar
2011-03-24 19:22 ` Ingo Molnar
2011-03-24 19:27 ` Pekka Enberg
2011-03-24 19:27 ` Pekka Enberg
2011-03-25 20:00 ` Ingo Molnar
2011-03-25 20:00 ` Ingo Molnar
2011-03-26 11:27 ` Ingo Molnar
2011-03-26 11:47 ` [PATCH] slub: Disable the lockless allocator Ingo Molnar
2011-03-26 11:47 ` Ingo Molnar
2011-03-26 17:45 ` Eric Dumazet
2011-03-26 17:45 ` Eric Dumazet
2011-03-26 19:18 ` Christoph Lameter
2011-03-26 19:18 ` Christoph Lameter
2011-03-26 19:30 ` Christoph Lameter
2011-03-26 19:30 ` Christoph Lameter
2011-03-26 19:49 ` Christoph Lameter
2011-03-26 19:49 ` Christoph Lameter
2011-03-26 19:57 ` Ingo Molnar
2011-03-26 19:57 ` Ingo Molnar
2011-03-26 21:36 ` Christoph Lameter
2011-03-26 21:36 ` Christoph Lameter
2011-03-26 20:00 ` Ingo Molnar
2011-03-26 20:00 ` Ingo Molnar
2011-03-26 22:18 ` Linus Torvalds
2011-03-27 1:29 ` Christoph Lameter
2011-03-27 1:57 ` Christoph Lameter
2011-03-27 7:52 ` Eric Dumazet
2011-03-27 7:52 ` Eric Dumazet
2011-03-28 10:32 ` [PATCH] percpu: avoid extra NOP in percpu_cmpxchg16b_double Eric Dumazet
2011-03-28 10:32 ` Eric Dumazet
2011-03-28 13:46 ` Christoph Lameter
2011-03-28 13:46 ` Christoph Lameter
2011-03-28 16:06 ` Tejun Heo
2011-03-28 16:06 ` Tejun Heo
2011-03-28 6:12 ` [PATCH] slub: Disable the lockless allocator Pekka Enberg
2011-03-28 6:12 ` Pekka Enberg
2011-03-28 6:19 ` Ingo Molnar [this message]
2011-03-28 6:19 ` Ingo Molnar
2011-03-28 6:25 ` Pekka Enberg
2011-03-28 6:25 ` Pekka Enberg
2011-03-28 6:36 ` Ingo Molnar
2011-03-28 6:36 ` Ingo Molnar
2011-03-28 7:26 ` Eric Dumazet
2011-03-28 7:26 ` Eric Dumazet
2011-03-28 13:44 ` Christoph Lameter
2011-03-28 13:44 ` Christoph Lameter
2011-03-28 7:35 ` Tejun Heo
2011-03-28 7:35 ` Tejun Heo
2011-03-26 19:30 ` Linus Torvalds
2011-03-26 19:30 ` Linus Torvalds
2011-03-26 17:58 ` [boot crash #2] Re: [GIT PULL] SLAB changes for v2.6.39-rc1 Linus Torvalds
2011-03-26 17:58 ` Linus Torvalds
2011-03-24 19:16 ` Call Trace:Re: " Thomas Gleixner
2011-03-24 19:16 ` Thomas Gleixner
2011-03-24 19:23 ` Ingo Molnar
2011-03-24 19:23 ` Ingo Molnar
2011-03-24 20:04 ` Gustavo F. Padovan
2011-03-24 20:04 ` Gustavo F. Padovan
2011-03-24 16:03 ` Ingo Molnar
2011-03-24 16:03 ` Ingo Molnar
2011-03-24 16:14 ` Ingo Molnar
2011-03-24 16:14 ` Ingo Molnar
2011-03-24 17:07 ` Pekka Enberg
2011-03-24 17:07 ` Pekka Enberg
2011-03-24 17:21 ` Ingo Molnar
2011-03-24 17:21 ` Ingo Molnar
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=20110328061929.GA24328@elte.hu \
--to=mingo@elte.hu \
--cc=akpm@linux-foundation.org \
--cc=cl@linux.com \
--cc=eric.dumazet@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=npiggin@kernel.dk \
--cc=penberg@cs.helsinki.fi \
--cc=rientjes@google.com \
--cc=tglx@linutronix.de \
--cc=tj@kernel.org \
--cc=torvalds@linux-foundation.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.