All of lore.kernel.org
 help / color / mirror / Atom feed
From: Zhang Yanfei <zhangyanfei.yes@gmail.com>
To: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	"Rafael J . Wysocki" <rjw@sisk.pl>,
	lenb@kernel.org, Thomas Gleixner <tglx@linutronix.de>,
	mingo@elte.hu, Tejun Heo <tj@kernel.org>,
	Toshi Kani <toshi.kani@hp.com>,
	Wanpeng Li <liwanp@linux.vnet.ibm.com>,
	Thomas Renninger <trenn@suse.de>, Yinghai Lu <yinghai@kernel.org>,
	Jiang Liu <jiang.liu@huawei.com>,
	Wen Congyang <wency@cn.fujitsu.com>,
	Lai Jiangshan <laijs@cn.fujitsu.com>,
	isimatu.yasuaki@jp.fujitsu.com, izumi.taku@jp.fujitsu.com,
	Mel Gorman <mgorman@suse.de>, Minchan Kim <minchan@kernel.org>,
	mina86@mina86.com, gong.chen@linux.intel.com,
	vasilis.liaskovitis@profitbricks.com, lwoodman@redhat.com,
	Rik van Riel <riel@redhat.com>,
	jweiner@redhat.com, prarit@redhat.com,
	"x86@kernel.org" <x86@kernel.org>,
	linux-doc@vger.kernel.org,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Linux MM <linux-mm@kv>
Subject: Re: [PATCH part1 v6 4/6] x86/mem-hotplug: Support initialize page tables in bottom-up
Date: Mon, 07 Oct 2013 22:17:32 +0800	[thread overview]
Message-ID: <5252C27C.4030506@gmail.com> (raw)
In-Reply-To: <5251F9AB.6000203@zytor.com>

Hello peter,

On 10/07/2013 08:00 AM, H. Peter Anvin wrote:
> On 10/03/2013 07:00 PM, Zhang Yanfei wrote:
>> From: Tang Chen <tangchen@cn.fujitsu.com>
>>
>> The Linux kernel cannot migrate pages used by the kernel. As a
>> result, kernel pages cannot be hot-removed. So we cannot allocate
>> hotpluggable memory for the kernel.
>>
>> In a memory hotplug system, any numa node the kernel resides in
>> should be unhotpluggable. And for a modern server, each node could
>> have at least 16GB memory. So memory around the kernel image is
>> highly likely unhotpluggable.
>>
>> ACPI SRAT (System Resource Affinity Table) contains the memory
>> hotplug info. But before SRAT is parsed, memblock has already
>> started to allocate memory for the kernel. So we need to prevent
>> memblock from doing this.
>>
>> So direct memory mapping page tables setup is the case. init_mem_mapping()
>> is called before SRAT is parsed. To prevent page tables being allocated
>> within hotpluggable memory, we will use bottom-up direction to allocate
>> page tables from the end of kernel image to the higher memory.
>>
>> Acked-by: Tejun Heo <tj@kernel.org>
>> Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
>> Signed-off-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
> 
> I'm still seriously concerned about this.  This unconditionally
> introduces new behavior which may very well break some classes of

Well, this new behaviour is not unconditional, if user doesn't specify
the movable_node option, the kernel will act as before, allocating
memory top-down.

> systems -- the whole point of creating the page tables top down is
> because the kernel tends to be allocated in lower memory, which is also
> the memory that some devices need for DMA.

How much memory does these devices needed for DMA? And you mean memory
under 16MB or 4GB?

> 
> +#ifdef CONFIG_X86
> +		kernel_end = __pa_symbol(_end);
> +#else
> +		kernel_end = __pa(RELOC_HIDE((unsigned long)(_end), 0));
> +#endif
> 
> We really should make __pa_symbol() available everywhere by putting
> something like the above in a global define (under #ifndef __pa_symbol).

Hmmmm...in include/asm-generic/page.h?

> 
> Is RELOC_HIDE() even correct here?

Sorry, could you explain a bit?

-- 
Thanks.
Zhang Yanfei

--
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>

WARNING: multiple messages have this Message-ID (diff)
From: Zhang Yanfei <zhangyanfei.yes@gmail.com>
To: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	"Rafael J . Wysocki" <rjw@sisk.pl>,
	lenb@kernel.org, Thomas Gleixner <tglx@linutronix.de>,
	mingo@elte.hu, Tejun Heo <tj@kernel.org>,
	Toshi Kani <toshi.kani@hp.com>,
	Wanpeng Li <liwanp@linux.vnet.ibm.com>,
	Thomas Renninger <trenn@suse.de>, Yinghai Lu <yinghai@kernel.org>,
	Jiang Liu <jiang.liu@huawei.com>,
	Wen Congyang <wency@cn.fujitsu.com>,
	Lai Jiangshan <laijs@cn.fujitsu.com>,
	isimatu.yasuaki@jp.fujitsu.com, izumi.taku@jp.fujitsu.com,
	Mel Gorman <mgorman@suse.de>, Minchan Kim <minchan@kernel.org>,
	mina86@mina86.com, gong.chen@linux.intel.com,
	vasilis.liaskovitis@profitbricks.com, lwoodman@redhat.com,
	Rik van Riel <riel@redhat.com>,
	jweiner@redhat.com, prarit@redhat.com,
	"x86@kernel.org" <x86@kernel.org>,
	linux-doc@vger.kernel.org,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Linux MM <linux-mm@kvack.org>,
	linux-acpi@vger.kernel.org, imtangchen@gmail.com,
	Zhang Yanfei <zhangyanfei@cn.fujitsu.com>,
	Tang Chen <tangchen@cn.fujitsu.com>
Subject: Re: [PATCH part1 v6 4/6] x86/mem-hotplug: Support initialize page tables in bottom-up
Date: Mon, 07 Oct 2013 22:17:32 +0800	[thread overview]
Message-ID: <5252C27C.4030506@gmail.com> (raw)
In-Reply-To: <5251F9AB.6000203@zytor.com>

Hello peter,

On 10/07/2013 08:00 AM, H. Peter Anvin wrote:
> On 10/03/2013 07:00 PM, Zhang Yanfei wrote:
>> From: Tang Chen <tangchen@cn.fujitsu.com>
>>
>> The Linux kernel cannot migrate pages used by the kernel. As a
>> result, kernel pages cannot be hot-removed. So we cannot allocate
>> hotpluggable memory for the kernel.
>>
>> In a memory hotplug system, any numa node the kernel resides in
>> should be unhotpluggable. And for a modern server, each node could
>> have at least 16GB memory. So memory around the kernel image is
>> highly likely unhotpluggable.
>>
>> ACPI SRAT (System Resource Affinity Table) contains the memory
>> hotplug info. But before SRAT is parsed, memblock has already
>> started to allocate memory for the kernel. So we need to prevent
>> memblock from doing this.
>>
>> So direct memory mapping page tables setup is the case. init_mem_mapping()
>> is called before SRAT is parsed. To prevent page tables being allocated
>> within hotpluggable memory, we will use bottom-up direction to allocate
>> page tables from the end of kernel image to the higher memory.
>>
>> Acked-by: Tejun Heo <tj@kernel.org>
>> Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
>> Signed-off-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
> 
> I'm still seriously concerned about this.  This unconditionally
> introduces new behavior which may very well break some classes of

Well, this new behaviour is not unconditional, if user doesn't specify
the movable_node option, the kernel will act as before, allocating
memory top-down.

> systems -- the whole point of creating the page tables top down is
> because the kernel tends to be allocated in lower memory, which is also
> the memory that some devices need for DMA.

How much memory does these devices needed for DMA? And you mean memory
under 16MB or 4GB?

> 
> +#ifdef CONFIG_X86
> +		kernel_end = __pa_symbol(_end);
> +#else
> +		kernel_end = __pa(RELOC_HIDE((unsigned long)(_end), 0));
> +#endif
> 
> We really should make __pa_symbol() available everywhere by putting
> something like the above in a global define (under #ifndef __pa_symbol).

Hmmmm...in include/asm-generic/page.h?

> 
> Is RELOC_HIDE() even correct here?

Sorry, could you explain a bit?

-- 
Thanks.
Zhang Yanfei

--
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>

  reply	other threads:[~2013-10-07 14:17 UTC|newest]

Thread overview: 109+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-10-04  1:56 [PATCH part1 v6 0/6] x86, memblock: Allocate memory near kernel image before SRAT parsed Zhang Yanfei
2013-10-04  1:56 ` Zhang Yanfei
2013-10-04  1:57 ` [PATCH part1 v6 1/6] memblock: Factor out of top-down allocation Zhang Yanfei
2013-10-04  1:57   ` Zhang Yanfei
2013-10-04  1:58 ` [PATCH part1 v6 2/6] memblock: Introduce bottom-up allocation mode Zhang Yanfei
2013-10-04  1:58   ` Zhang Yanfei
2013-10-05 21:30   ` Toshi Kani
2013-10-05 21:30     ` Toshi Kani
2013-10-04  1:59 ` [PATCH part1 v6 3/6] x86/mm: Factor out of top-down direct mapping setup Zhang Yanfei
2013-10-04  1:59   ` Zhang Yanfei
2013-10-04  2:00 ` [PATCH part1 v6 4/6] x86/mem-hotplug: Support initialize page tables in bottom-up Zhang Yanfei
2013-10-04  2:00   ` Zhang Yanfei
2013-10-05 22:09   ` Toshi Kani
2013-10-05 22:09     ` Toshi Kani
2013-10-07  0:00   ` H. Peter Anvin
2013-10-07  0:00     ` H. Peter Anvin
2013-10-07 14:17     ` Zhang Yanfei [this message]
2013-10-07 14:17       ` Zhang Yanfei
2013-10-08 17:36     ` Zhang Yanfei
2013-10-08 17:36       ` Zhang Yanfei
2013-10-08 17:36       ` Zhang Yanfei
2013-10-09 16:44       ` Tejun Heo
2013-10-09 16:44         ` Tejun Heo
2013-10-09 17:14         ` Zhang Yanfei
2013-10-09 17:14           ` Zhang Yanfei
2013-10-09 19:20           ` Tejun Heo
2013-10-09 19:20             ` Tejun Heo
2013-10-09 19:30             ` Dave Hansen
2013-10-09 19:30               ` Dave Hansen
2013-10-09 19:47               ` Tejun Heo
2013-10-09 19:47                 ` Tejun Heo
2013-10-09 20:58             ` Toshi Kani
2013-10-09 20:58               ` Toshi Kani
2013-10-09 21:11               ` Tejun Heo
2013-10-09 21:11                 ` Tejun Heo
2013-10-09 21:14                 ` H. Peter Anvin
2013-10-09 21:14                   ` H. Peter Anvin
2013-10-09 21:45                   ` Zhang Yanfei
2013-10-09 21:45                     ` Zhang Yanfei
2013-10-09 23:10                     ` H. Peter Anvin
2013-10-09 23:10                       ` H. Peter Anvin
2013-10-09 23:26                       ` Zhang Yanfei
2013-10-09 23:26                         ` Zhang Yanfei
2013-10-10  1:20                         ` Zhang Yanfei
2013-10-10  1:20                           ` Zhang Yanfei
2013-10-10  1:20                           ` Zhang Yanfei
2013-10-10  0:25                   ` Toshi Kani
2013-10-10  0:25                     ` Toshi Kani
2013-10-09 23:58                 ` Toshi Kani
2013-10-09 23:58                   ` Toshi Kani
2013-10-10  1:00                   ` Tejun Heo
2013-10-10  1:00                     ` Tejun Heo
2013-10-10 14:36                     ` Toshi Kani
2013-10-10 14:36                       ` Toshi Kani
2013-10-10 15:35                       ` Tejun Heo
2013-10-10 15:35                         ` Tejun Heo
2013-10-10 16:24                         ` Toshi Kani
2013-10-10 16:24                           ` Toshi Kani
2013-10-10 16:46                           ` Tejun Heo
2013-10-10 16:46                             ` Tejun Heo
2013-10-10 16:50                             ` Toshi Kani
2013-10-10 16:50                               ` Toshi Kani
2013-10-10 16:55                               ` Tejun Heo
2013-10-10 16:55                                 ` Tejun Heo
2013-10-10 16:59                                 ` Toshi Kani
2013-10-10 16:59                                   ` Toshi Kani
2013-10-10 17:12                                   ` H. Peter Anvin
2013-10-10 17:12                                     ` H. Peter Anvin
2013-10-10 19:17                                     ` Toshi Kani
2013-10-10 19:17                                       ` Toshi Kani
2013-10-10 22:19                                       ` Tejun Heo
2013-10-10 22:19                                         ` Tejun Heo
2013-10-10 23:00                                         ` Toshi Kani
2013-10-10 23:00                                           ` Toshi Kani
2013-10-09 21:19             ` Zhang Yanfei
2013-10-09 21:19               ` Zhang Yanfei
2013-10-09 21:22               ` H. Peter Anvin
2013-10-09 21:22                 ` H. Peter Anvin
2013-10-09 23:30                 ` Zhang Yanfei
2013-10-09 23:30                   ` Zhang Yanfei
2013-10-09 19:10         ` Yinghai Lu
2013-10-09 19:10           ` Yinghai Lu
2013-10-09 19:23           ` Tejun Heo
2013-10-09 19:23             ` Tejun Heo
2013-10-11  5:27             ` Yinghai Lu
2013-10-11  5:27               ` Yinghai Lu
2013-10-11  5:47               ` Zhang Yanfei
2013-10-11  5:47                 ` Zhang Yanfei
2013-10-11  6:33                 ` Ingo Molnar
2013-10-11  6:33                   ` Ingo Molnar
2013-10-11  6:46                   ` Zhang Yanfei
2013-10-11  6:46                     ` Zhang Yanfei
2013-10-04  2:01 ` [PATCH part1 v6 5/6] x86, acpi, crash, kdump: Do reserve_crashkernel() after SRAT is parsed Zhang Yanfei
2013-10-04  2:01   ` Zhang Yanfei
2013-10-05 22:10   ` Toshi Kani
2013-10-05 22:10     ` Toshi Kani
2013-10-04  2:02 ` [PATCH part1 v6 6/6] mem-hotplug: Introduce movable_node boot option Zhang Yanfei
2013-10-04  2:02   ` Zhang Yanfei
2013-10-05 22:28   ` Toshi Kani
2013-10-05 22:28     ` Toshi Kani
2013-10-06 14:43     ` [PATCH part1 v6 update " Zhang Yanfei
2013-10-06 14:43       ` Zhang Yanfei
2013-10-06 14:43       ` Zhang Yanfei
2013-10-06 23:03       ` Toshi Kani
2013-10-06 23:03         ` Toshi Kani
2013-10-08  4:23 ` [PATCH part1 v6 0/6] x86, memblock: Allocate memory near kernel image before SRAT parsed Ingo Molnar
2013-10-08  4:23   ` Ingo Molnar
2013-10-08 15:28   ` Zhang Yanfei
2013-10-08 15:28     ` Zhang Yanfei

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=5252C27C.4030506@gmail.com \
    --to=zhangyanfei.yes@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=gong.chen@linux.intel.com \
    --cc=hpa@zytor.com \
    --cc=isimatu.yasuaki@jp.fujitsu.com \
    --cc=izumi.taku@jp.fujitsu.com \
    --cc=jiang.liu@huawei.com \
    --cc=jweiner@redhat.com \
    --cc=laijs@cn.fujitsu.com \
    --cc=lenb@kernel.org \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kv \
    --cc=liwanp@linux.vnet.ibm.com \
    --cc=lwoodman@redhat.com \
    --cc=mgorman@suse.de \
    --cc=mina86@mina86.com \
    --cc=minchan@kernel.org \
    --cc=mingo@elte.hu \
    --cc=prarit@redhat.com \
    --cc=riel@redhat.com \
    --cc=rjw@sisk.pl \
    --cc=tglx@linutronix.de \
    --cc=tj@kernel.org \
    --cc=toshi.kani@hp.com \
    --cc=trenn@suse.de \
    --cc=vasilis.liaskovitis@profitbricks.com \
    --cc=wency@cn.fujitsu.com \
    --cc=x86@kernel.org \
    --cc=yinghai@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 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.