xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
To: "Liu, Jinsong" <jinsong.liu@intel.com>,
	Daniel Kiper <daniel.kiper@oracle.com>
Cc: "xen-devel@lists.xensource.com" <xen-devel@lists.xensource.com>,
	"jeremy.fitzhardinge@citrix.com" <jeremy.fitzhardinge@citrix.com>,
	"Jiang, Yunhong" <yunhong.jiang@intel.com>
Subject: Re: [PATCH 06/10] Xen: Add memory hotadd to pvops dom0
Date: Wed, 3 Oct 2012 09:29:52 -0400	[thread overview]
Message-ID: <20121003132952.GA31173@phenom.dumpdata.com> (raw)
In-Reply-To: <DE8DF0795D48FD4CA783C40EC82923358A61@SHSMSX101.ccr.corp.intel.com>

On Thu, Dec 22, 2011 at 10:06:56AM +0000, Liu, Jinsong wrote:
> >From a5d21e2a01049eef6aa64406e71f6574e7a371c1 Mon Sep 17 00:00:00 2001
> From: Liu Jinsong <jinsong.liu@intel.com>
> Date: Sun, 11 Dec 2011 21:51:28 +0800
> Subject: [PATCH 06/10] Xen: Add memory hotadd to pvops dom0
> 
> This patch rebased from Jeremy's pvops commit
> fe3cb1ebf78399fd5176ff6aa1c9ab498f2d8bd7
> 
> When memory hotadd event happen, a Xen hook will be called, to notify
> hypervisor of the new added memory.

I recall posting a review asking whether it would make sense to use the
generic memory hotplug mechanism but never got any feedback. Is there
no interest in this anymore?

> 
> Because xen hypervisor will use the new memory to setup frametable/m2p
> table, so dom0 will always return success to acpi bios, and notify xen
> hypervisor later.
> 
> It add a hook in driver/acpi/acpi_memhotplug.c, but that change is
> quite small, not sure if it is acceptable. Other method is to provide
> a xen specific acpi_memory_device_driver, but I'm not sure if it worth
> to add so much changes, to simply avoid two hooks.
> 
> Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>
> Signed-off-by: Jiang, Yunhong <yunhong.jiang@intel.com>
> Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
> ---
>  drivers/acpi/acpi_memhotplug.c    |    4 +
>  drivers/xen/Makefile              |    1 +
>  drivers/xen/xen_acpi_memhotplug.c |  209 +++++++++++++++++++++++++++++++++++++
>  include/xen/acpi.h                |    5 +
>  include/xen/interface/platform.h  |    9 ++
>  5 files changed, 228 insertions(+), 0 deletions(-)
>  create mode 100644 drivers/xen/xen_acpi_memhotplug.c
> 
> diff --git a/drivers/acpi/acpi_memhotplug.c b/drivers/acpi/acpi_memhotplug.c
> index e28e64d..2f2169e 100644
> --- a/drivers/acpi/acpi_memhotplug.c
> +++ b/drivers/acpi/acpi_memhotplug.c
> @@ -32,6 +32,7 @@
>  #include <linux/memory_hotplug.h>
>  #include <linux/slab.h>
>  #include <acpi/acpi_drivers.h>
> +#include <xen/acpi.h>
>  
>  #define ACPI_MEMORY_DEVICE_CLASS		"memory"
>  #define ACPI_MEMORY_DEVICE_HID			"PNP0C80"
> @@ -214,6 +215,9 @@ static int acpi_memory_enable_device(struct acpi_memory_device *mem_device)
>  		return result;
>  	}
>  
> +	if (xen_initial_domain())
> +		return xen_hotadd_memory(mem_device);
> +
>  	node = acpi_get_node(mem_device->device->handle);
>  	/*
>  	 * Tell the VM there is more memory here...
> diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> index aedaf48..7dc3c0b 100644
> --- a/drivers/xen/Makefile
> +++ b/drivers/xen/Makefile
> @@ -20,6 +20,7 @@ obj-$(CONFIG_XEN_TMEM)			+= tmem.o
>  obj-$(CONFIG_SWIOTLB_XEN)		+= swiotlb-xen.o
>  obj-$(CONFIG_XEN_DOM0)			+= pci.o acpi.o
>  obj-$(CONFIG_XEN_PCIDEV_BACKEND)	+= xen-pciback/
> +obj-$(CONFIG_ACPI_HOTPLUG_MEMORY)	+= xen_acpi_memhotplug.o
>  ifdef CONFIG_ACPI_PROCESSOR_XEN
>  obj-$(CONFIG_ACPI_PROCESSOR)		+= acpi_processor.o
>  endif
> diff --git a/drivers/xen/xen_acpi_memhotplug.c b/drivers/xen/xen_acpi_memhotplug.c
> new file mode 100644
> index 0000000..0c4af99
> --- /dev/null
> +++ b/drivers/xen/xen_acpi_memhotplug.c
> @@ -0,0 +1,209 @@
> +/*
> + *  xen_acpi_memhotplug.c - interface to notify Xen on memory device hotadd
> + *
> + *  Copyright (C) 2008, Intel corporation
> + *
> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> + *
> + *  This program is free software; you can redistribute it and/or modify
> + *  it under the terms of the GNU General Public License as published by
> + *  the Free Software Foundation; either version 2 of the License, or (at
> + *  your option) any later version.
> + *
> + *  This program is distributed in the hope that it will be useful, but
> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> + *  General Public License for more details.
> + *
> + *  You should have received a copy of the GNU General Public License along
> + *  with this program; if not, write to the Free Software Foundation, Inc.,
> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
> + *
> + */
> +
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/types.h>
> +#include <linux/memory_hotplug.h>
> +#include <acpi/acpi_drivers.h>
> +#include <xen/interface/platform.h>
> +#include <linux/interrupt.h>
> +#include <linux/spinlock.h>
> +#include <asm/xen/hypervisor.h>
> +#include <asm/xen/hypercall.h>
> +#include <xen/acpi.h>
> +
> +struct xen_hotmem_entry {
> +	struct list_head hotmem_list;
> +	uint64_t start;
> +	uint64_t end;
> +	uint32_t flags;
> +	uint32_t pxm;
> +};
> +
> +struct xen_hotmem_list {
> +	struct list_head list;
> +	int entry_nr;
> +} xen_hotmem;
> +
> +DEFINE_SPINLOCK(xen_hotmem_lock);
> +
> +static int xen_hyper_addmem(struct xen_hotmem_entry *entry)
> +{
> +	int ret;
> +
> +	xen_platform_op_t op = {
> +		.cmd            = XENPF_mem_hotadd,
> +		.interface_version  = XENPF_INTERFACE_VERSION,
> +	};
> +	op.u.mem_add.spfn = entry->start >> PAGE_SHIFT;
> +	op.u.mem_add.epfn = entry->end >> PAGE_SHIFT;
> +	op.u.mem_add.flags = entry->flags;
> +	op.u.mem_add.pxm = entry->pxm;
> +
> +	ret = HYPERVISOR_dom0_op(&op);
> +	return ret;
> +}
> +
> +static int add_hotmem_entry(int pxm, uint64_t start,
> +			uint64_t length, uint32_t flags)
> +{
> +	struct xen_hotmem_entry *entry;
> +
> +	if (pxm < 0 || !length)
> +		return -EINVAL;
> +
> +	entry = kzalloc(sizeof(struct xen_hotmem_entry), GFP_ATOMIC);
> +	if (!entry)
> +		return -ENOMEM;
> +
> +	INIT_LIST_HEAD(&entry->hotmem_list);
> +	entry->start = start;
> +	entry->end = start + length;
> +	entry->flags = flags;
> +	entry->pxm = pxm;
> +
> +	spin_lock(&xen_hotmem_lock);
> +
> +	list_add_tail(&entry->hotmem_list, &xen_hotmem.list);
> +	xen_hotmem.entry_nr++;
> +
> +	spin_unlock(&xen_hotmem_lock);
> +
> +	return 0;
> +}
> +
> +static int free_hotmem_entry(struct xen_hotmem_entry *entry)
> +{
> +	list_del(&entry->hotmem_list);
> +	kfree(entry);
> +
> +	return 0;
> +}
> +
> +static void xen_hotadd_mem_dpc(struct work_struct *work)
> +{
> +	struct list_head *elem, *tmp;
> +	struct xen_hotmem_entry *entry;
> +	unsigned long flags;
> +	int ret;
> +
> +	spin_lock_irqsave(&xen_hotmem_lock, flags);
> +	list_for_each_safe(elem, tmp, &xen_hotmem.list) {
> +		entry = list_entry(elem, struct xen_hotmem_entry, hotmem_list);
> +		ret = xen_hyper_addmem(entry);
> +		if (ret)
> +			printk(KERN_WARNING "xen addmem failed with %x\n", ret);
> +		free_hotmem_entry(entry);
> +		xen_hotmem.entry_nr--;
> +	}
> +	spin_unlock_irqrestore(&xen_hotmem_lock, flags);
> +}
> +
> +static DECLARE_WORK(xen_hotadd_mem_work, xen_hotadd_mem_dpc);
> +
> +static int xen_acpi_get_pxm(acpi_handle h)
> +{
> +	unsigned long long pxm;
> +	acpi_status status;
> +	acpi_handle handle;
> +	acpi_handle phandle = h;
> +
> +	do {
> +		handle = phandle;
> +		status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
> +		if (ACPI_SUCCESS(status))
> +			return pxm;
> +		status = acpi_get_parent(handle, &phandle);
> +	} while (ACPI_SUCCESS(status));
> +
> +	return -1;
> +}
> +
> +int xen_hotadd_memory(struct acpi_memory_device *mem_device)
> +{
> +	int pxm, result;
> +	int num_enabled = 0;
> +	struct acpi_memory_info *info;
> +
> +	if (!mem_device)
> +		return -EINVAL;
> +
> +	pxm = xen_acpi_get_pxm(mem_device->device->handle);
> +
> +	if (pxm < 0)
> +		return -EINVAL;
> +
> +	/*
> +	 * Always return success to ACPI driver, and notify hypervisor later
> +	 * because hypervisor will utilize the memory in memory hotadd hypercall
> +	 */
> +	list_for_each_entry(info, &mem_device->res_list, list) {
> +		if (info->enabled) { /* just sanity check...*/
> +			num_enabled++;
> +			continue;
> +		}
> +		/*
> +		 * If the memory block size is zero, please ignore it.
> +		 * Don't try to do the following memory hotplug flowchart.
> +		 */
> +		if (!info->length)
> +			continue;
> +
> +		result = add_hotmem_entry(pxm, info->start_addr,
> +					info->length, 0);
> +		if (result)
> +			continue;
> +		info->enabled = 1;
> +		num_enabled++;
> +	}
> +
> +	if (!num_enabled)
> +		return -EINVAL;
> +
> +	schedule_work(&xen_hotadd_mem_work);
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL(xen_hotadd_memory);
> +
> +static int xen_hotadd_mem_init(void)
> +{
> +	if (!xen_initial_domain())
> +		return -ENODEV;
> +
> +	INIT_LIST_HEAD(&xen_hotmem.list);
> +	xen_hotmem.entry_nr = 0;
> +
> +	return 0;
> +}
> +
> +static void xen_hotadd_mem_exit(void)
> +{
> +	flush_scheduled_work();
> +}
> +
> +module_init(xen_hotadd_mem_init);
> +module_exit(xen_hotadd_mem_exit);
> +MODULE_LICENSE("GPL");
> diff --git a/include/xen/acpi.h b/include/xen/acpi.h
> index 7aa282d..8b3462e 100644
> --- a/include/xen/acpi.h
> +++ b/include/xen/acpi.h
> @@ -107,6 +107,11 @@ static inline int xen_acpi_processor_get_performance(struct acpi_processor *pr)
>  }
>  #endif
>  
> +#if defined(CONFIG_ACPI_HOTPLUG_MEMORY) || \
> +defined(CONFIG_ACPI_HOTPLUG_MEMORY_MODULE)
> +int xen_hotadd_memory(struct acpi_memory_device *mem_device);
> +#endif
> +
>  #if defined(CONFIG_ACPI_PROCESSOR_XEN) || \
>  defined(CONFIG_ACPI_PROCESSOR_XEN_MODULE)
>  
> diff --git a/include/xen/interface/platform.h b/include/xen/interface/platform.h
> index 9fd6b07..0787c68 100644
> --- a/include/xen/interface/platform.h
> +++ b/include/xen/interface/platform.h
> @@ -329,6 +329,14 @@ struct xenpf_cpu_hotadd {
>  	uint32_t pxm;
>  };
>  
> +#define XENPF_mem_hotadd    59
> +struct xenpf_mem_hotadd {
> +	uint64_t spfn;
> +	uint64_t epfn;
> +	uint32_t pxm;
> +	uint32_t flags;
> +};
> +
>  struct xen_platform_op {
>  	uint32_t cmd;
>  	uint32_t interface_version; /* XENPF_INTERFACE_VERSION */
> @@ -347,6 +355,7 @@ struct xen_platform_op {
>  		struct xenpf_pcpuinfo          pcpu_info;
>  		struct xenpf_cpu_ol            cpu_ol;
>  		struct xenpf_cpu_hotadd        cpu_add;
> +		struct xenpf_mem_hotadd        mem_add;
>  		uint8_t                        pad[128];
>  	} u;
>  };
> -- 
> 1.6.5.6

  reply	other threads:[~2012-10-03 13:29 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-12-22 10:06 [PATCH 06/10] Xen: Add memory hotadd to pvops dom0 Liu, Jinsong
2012-10-03 13:29 ` Konrad Rzeszutek Wilk [this message]
  -- strict thread matches above, loose matches on Subject: below --
2011-12-23 10:28 Liu, Jinsong

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=20121003132952.GA31173@phenom.dumpdata.com \
    --to=konrad.wilk@oracle.com \
    --cc=daniel.kiper@oracle.com \
    --cc=jeremy.fitzhardinge@citrix.com \
    --cc=jinsong.liu@intel.com \
    --cc=xen-devel@lists.xensource.com \
    --cc=yunhong.jiang@intel.com \
    /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).