xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
To: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Cc: jeremy@goop.org, xen-devel@lists.xensource.com, Ian.Campbell@citrix.com
Subject: Re: [PATCH 5/6] xen-gntalloc: Userspace grant allocation driver
Date: Thu, 27 Jan 2011 13:52:10 -0500	[thread overview]
Message-ID: <20110127185210.GA27853@dumpdata.com> (raw)
In-Reply-To: <1295625548-22069-6-git-send-email-dgdegra@tycho.nsa.gov>

On Fri, Jan 21, 2011 at 10:59:07AM -0500, Daniel De Graaf wrote:
> This allows a userspace application to allocate a shared page for
> implementing inter-domain communication or device drivers. These
> shared pages can be mapped using the gntdev device or by the kernel
> in another domain.
> 
> Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
> ---
>  drivers/xen/Kconfig    |    8 +
>  drivers/xen/Makefile   |    2 +
>  drivers/xen/gntalloc.c |  477 ++++++++++++++++++++++++++++++++++++++++++++++++
>  include/xen/gntalloc.h |   70 +++++++
>  4 files changed, 557 insertions(+), 0 deletions(-)
>  create mode 100644 drivers/xen/gntalloc.c
>  create mode 100644 include/xen/gntalloc.h
> 
> diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig
> index 07bec09..a3d7afb 100644
> --- a/drivers/xen/Kconfig
> +++ b/drivers/xen/Kconfig
> @@ -80,6 +80,14 @@ config XEN_GNTDEV
>  	help
>  	  Allows userspace processes to use grants.
>  
> +config XEN_GRANT_DEV_ALLOC
> +	tristate "User-space grant reference allocator driver"
> +	depends on XEN
> +	help
> +	  Allows userspace processes to create pages with access granted
> +	  to other domains. This can be used to implement frontend drivers
> +	  or as part of an inter-domain shared memory channel.
> +
>  config XEN_PLATFORM_PCI
>  	tristate "xen platform pci device driver"
>  	depends on XEN_PVHVM && PCI
> diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> index 5088cc2..9585a1d 100644
> --- a/drivers/xen/Makefile
> +++ b/drivers/xen/Makefile
> @@ -10,6 +10,7 @@ obj-$(CONFIG_XEN_XENCOMM)	+= xencomm.o
>  obj-$(CONFIG_XEN_BALLOON)	+= balloon.o
>  obj-$(CONFIG_XEN_DEV_EVTCHN)	+= xen-evtchn.o
>  obj-$(CONFIG_XEN_GNTDEV)	+= xen-gntdev.o
> +obj-$(CONFIG_XEN_GRANT_DEV_ALLOC)	+= xen-gntalloc.o
>  obj-$(CONFIG_XENFS)		+= xenfs/
>  obj-$(CONFIG_XEN_SYS_HYPERVISOR)	+= sys-hypervisor.o
>  obj-$(CONFIG_XEN_PLATFORM_PCI)	+= xen-platform-pci.o
> @@ -18,5 +19,6 @@ obj-$(CONFIG_XEN_DOM0)		+= pci.o
>  
>  xen-evtchn-y			:= evtchn.o
>  xen-gntdev-y				:= gntdev.o
> +xen-gntalloc-y				:= gntalloc.o
>  
>  xen-platform-pci-y		:= platform-pci.o
> diff --git a/drivers/xen/gntalloc.c b/drivers/xen/gntalloc.c
> new file mode 100644
> index 0000000..a230dc4
> --- /dev/null
> +++ b/drivers/xen/gntalloc.c
> @@ -0,0 +1,477 @@
> +/******************************************************************************
> + * gntalloc.c
> + *
> + * Device for creating grant references (in user-space) that may be shared
> + * with other domains.
> + *
> + * 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
> + */
> +
> +/*
> + * This driver exists to allow userspace programs in Linux to allocate kernel
> + * memory that will later be shared with another domain.  Without this device,
> + * Linux userspace programs cannot create grant references.
> + *
> + * How this stuff works:
> + *   X -> granting a page to Y
> + *   Y -> mapping the grant from X
> + *
> + *   1. X uses the gntalloc device to allocate a page of kernel memory, P.
> + *   2. X creates an entry in the grant table that says domid(Y) can access P.
> + *      This is done without a hypercall unless the grant table needs expansion.
> + *   3. X gives the grant reference identifier, GREF, to Y.
> + *   4. Y maps the page, either directly into kernel memory for use in a backend
> + *      driver, or via a the gntdev device to map into the address space of an
> + *      application running in Y. This is the first point at which Xen does any
> + *      tracking of the page.
> + *   5. A program in X mmap()s a segment of the gntalloc device that corresponds
> + *      to the shared page, and can now communicate with Y over the shared page.
> + *
> + *
> + * NOTE TO USERSPACE LIBRARIES:
> + *   The grant allocation and mmap()ing are, naturally, two separate operations.
> + *   You set up the sharing by calling the create ioctl() and then the mmap().
> + *   Teardown requires munmap() and either close() or ioctl().
> + *
> + * WARNING: Since Xen does not allow a guest to forcibly end the use of a grant
> + * reference, this device can be used to consume kernel memory by leaving grant
> + * references mapped by another domain when an application exits. Therefore,
> + * there is a global limit on the number of pages that can be allocated. When
> + * all references to the page are unmapped, it will be freed during the next
> + * grant operation.
> + */
> +
> +#include <linux/atomic.h>
> +#include <linux/module.h>
> +#include <linux/miscdevice.h>
> +#include <linux/kernel.h>
> +#include <linux/init.h>
> +#include <linux/slab.h>
> +#include <linux/fs.h>
> +#include <linux/device.h>
> +#include <linux/mm.h>
> +#include <linux/uaccess.h>
> +#include <linux/types.h>
> +#include <linux/list.h>
> +
> +#include <xen/xen.h>
> +#include <xen/page.h>
> +#include <xen/grant_table.h>
> +#include <xen/gntalloc.h>
> +
> +static int limit = 1024;
> +module_param(limit, int, 0644);
> +MODULE_PARM_DESC(limit, "Maximum number of grants that may be allocated by "
> +		"the gntalloc device");
> +
> +static LIST_HEAD(gref_list);
> +static DEFINE_SPINLOCK(gref_lock);
> +static int gref_size;
> +
> +/* Metadata on a grant reference. */
> +struct gntalloc_gref {
> +	struct list_head next_all;   /* list entry gref_list */

How about 'next_gref' ?
> +	struct list_head next_file;  /* list entry file->list, if open */
> +	struct page *page;	     /* The shared page */
> +	uint64_t file_index;         /* File offset for mmap() */
> +	unsigned int users;          /* Use count - when zero, waiting on Xen */
> +	grant_ref_t gref_id;         /* The grant reference number */
> +};
> +
> +struct gntalloc_file_private_data {
> +	struct list_head list;
> +	uint64_t index;
> +};
> +
> +static void __del_gref(struct gntalloc_gref *gref);
> +
> +static void do_cleanup(void)
> +{
> +	struct gntalloc_gref *gref, *n;
> +	list_for_each_entry_safe(gref, n, &gref_list, next_all) {
> +		if (!gref->users)
> +			__del_gref(gref);
> +	}
> +}
> +
> +static int add_grefs(struct ioctl_gntalloc_alloc_gref *op,
> +	uint32_t *gref_ids, struct gntalloc_file_private_data *priv)
> +{
> +	int i, rc, readonly;
> +	LIST_HEAD(queue_all);
> +	LIST_HEAD(queue_file);
> +	struct gntalloc_gref *gref;
> +
> +	readonly = !(op->flags & GNTALLOC_FLAG_WRITABLE);
> +	rc = -ENOMEM;
> +	for (i = 0; i < op->count; i++) {
> +		gref = kzalloc(sizeof(*gref), GFP_KERNEL);
> +		if (!gref)
> +			goto undo;
> +		list_add_tail(&gref->next_all, &queue_all);
> +		list_add_tail(&gref->next_file, &queue_file);
> +		gref->users = 1;
> +		gref->file_index = op->index + i * PAGE_SIZE;
> +		gref->page = alloc_page(GFP_KERNEL|__GFP_ZERO);
> +		if (!gref->page)
> +			goto undo;
> +
> +		/* Grant foreign access to the page. */
> +		gref->gref_id = gnttab_grant_foreign_access(op->domid,
> +			pfn_to_mfn(page_to_pfn(gref->page)), readonly);
> +		if (gref->gref_id < 0) {
> +			rc = gref->gref_id;
> +			goto undo;
> +		}
> +		gref_ids[i] = gref->gref_id;
> +	}
> +
> +	/* Add to gref lists. */
> +	spin_lock(&gref_lock);
> +	list_splice_tail(&queue_all, &gref_list);
> +	list_splice_tail(&queue_file, &priv->list);
> +	spin_unlock(&gref_lock);
> +
> +	return 0;
> +
> +undo:
> +	list_for_each_entry(gref, &queue_file, next_file) {
> +		if (gref->page && gref->gref_id >= 0) {
> +			__del_gref(gref);
> +		} else {
> +			if (gref->page)
> +				__free_page(gref->page);
> +			list_del(&gref->next_all);

Ohh, you are actually deleting them from queue_all, ah.
Can you add a comment about that since the first time I read this
I thought you were iterating over the queue_file and deleting
from _that_ list. Tricky..


Can collapse this deleting into __del_gref? Looks like
you should just preface the gnttab with "if gref->page && gref->gref_id"
and also the gref_size, and the "if (gref->page)" before deleting?


> +			kfree(gref);
> +		}
> +	}
> +
> +	/* It's possible for the target domain to map the just-allocated grant
> +	 * references by blindly guessing their IDs; if this is done, then
> +	 * __del_gref will leave them in the queue_all list. They need to be
> +	 * added to the global list so that we can free them when they are no
> +	 * longer referenced.
> +	 */
> +	if (unlikely(!list_empty(&queue_all))) {
> +		spin_lock(&gref_lock);
> +		list_splice_tail(&queue_all, &gref_list);
> +		spin_unlock(&gref_lock);
> +	}
> +	return rc;
> +}
> +
> +static void __del_gref(struct gntalloc_gref *gref)
> +{
> +	if (gnttab_query_foreign_access(gref->gref_id))
> +		return;
> +
> +	if (!gnttab_end_foreign_access_ref(gref->gref_id, 0))
> +		return;
> +
> +	gref_size--;
> +	list_del(&gref->next_all);
> +
> +	__free_page(gref->page);
> +	kfree(gref);
> +}
> +
> +/* finds contiguous grant references in a file, returns the first */
> +static struct gntalloc_gref *find_grefs(struct gntalloc_file_private_data *priv,
> +		uint64_t index, uint32_t count)
> +{
> +	struct gntalloc_gref *rv = NULL, *gref;
> +	list_for_each_entry(gref, &priv->list, next_file) {
> +		if (gref->file_index == index && !rv)
> +			rv = gref;
> +		if (rv) {
> +			if (gref->file_index != index)
> +				return NULL;
> +			index += PAGE_SIZE;
> +			count--;
> +			if (count == 0)
> +				return rv;
> +		}
> +	}
> +	return NULL;
> +}
> +
> +/*
> + * -------------------------------------
> + *  File operations.
> + * -------------------------------------
> + */
> +static int gntalloc_open(struct inode *inode, struct file *filp)
> +{
> +	struct gntalloc_file_private_data *priv;
> +
> +	try_module_get(THIS_MODULE);

No checking if it fails?
> +
> +	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
> +	if (!priv)
> +		goto out_nomem;
> +	INIT_LIST_HEAD(&priv->list);
> +
> +	filp->private_data = priv;
> +
> +	pr_debug("%s: priv %p\n", __func__, priv);
> +
> +	return 0;
> +
> +out_nomem:

You should also do:

	module_put(THIS_MODULE);

> +	return -ENOMEM;
> +}
> +
> +static int gntalloc_release(struct inode *inode, struct file *filp)
> +{
> +	struct gntalloc_file_private_data *priv = filp->private_data;
> +	struct gntalloc_gref *gref;
> +
> +	pr_debug("%s: priv %p\n", __func__, priv);
> +
> +	spin_lock(&gref_lock);
> +	while (!list_empty(&priv->list)) {
> +		gref = list_entry(priv->list.next,
> +			struct gntalloc_gref, next_file);


> +		list_del(&gref->next_file);
> +		gref->users--;
> +		if (gref->users == 0)
> +			__del_gref(gref);
> +	}
> +	kfree(priv);
> +	spin_unlock(&gref_lock);
> +
> +	module_put(THIS_MODULE);
> +
> +	return 0;
> +}
> +
> +static long gntalloc_ioctl_alloc(struct gntalloc_file_private_data *priv,
> +		struct ioctl_gntalloc_alloc_gref __user *arg)
> +{
> +	int rc = 0;
> +	struct ioctl_gntalloc_alloc_gref op;
> +	uint32_t *gref_ids;
> +
> +	pr_debug("%s: priv %p\n", __func__, priv);
> +
> +	if (copy_from_user(&op, arg, sizeof(op))) {
> +		rc = -EFAULT;
> +		goto out;
> +	}
> +
> +	gref_ids = kzalloc(sizeof(gref_ids[0]) * op.count, GFP_TEMPORARY);
> +	if (!gref_ids) {
> +		rc = -ENOMEM;
> +		goto out;
> +	}
> +
> +	spin_lock(&gref_lock);
> +	do_cleanup();

Hmm, why the cleanup here? I see it  gntalloc_ioctl_dealloc
which makes sense since the users-- might have been decremented to zero.
But here?

> +	if (gref_size + op.count > limit) {
> +		spin_unlock(&gref_lock);
> +		rc = -ENOSPC;
> +		goto out_free;
> +	}
> +	gref_size += op.count;
> +	op.index = priv->index;
> +	priv->index += op.count * PAGE_SIZE;
> +	spin_unlock(&gref_lock);
> +
> +	rc = add_grefs(&op, gref_ids, priv);
> +	if (rc < 0)
> +		goto out_free;
> +
> +	if (copy_to_user(arg, &op, sizeof(op))) {
> +		rc = -EFAULT;
> +		goto out_free;

Not something that would clean the newly added grant? Say
the code I suggested below the 'out' label.

> +	}
> +	if (copy_to_user(arg->gref_ids, gref_ids,
> +			sizeof(gref_ids[0]) * op.count)) {
> +		rc = -EFAULT;
> +		goto out_free;

Ditto.
> +	}
> +
> +out_free:
> +	kfree(gref_ids);
> +out:
> +	return rc;

out_del_gref:
	gref->users--;
	__del_gref(gref);

	goto out_free;
> +}
> +
> +static long gntalloc_ioctl_dealloc(struct gntalloc_file_private_data *priv,
> +		void __user *arg)
> +{
> +	int i, rc = 0;
> +	struct ioctl_gntalloc_dealloc_gref op;
> +	struct gntalloc_gref *gref, *n;
> +
> +	pr_debug("%s: priv %p\n", __func__, priv);
> +
> +	if (copy_from_user(&op, arg, sizeof(op))) {
> +		rc = -EFAULT;
> +		goto dealloc_grant_out;
> +	}
> +
> +	spin_lock(&gref_lock);
> +	gref = find_grefs(priv, op.index, op.count);
> +	if (gref) {
> +		for (i = 0; i < op.count; i++) {
> +			n = list_entry(gref->next_file.next,
> +				struct gntalloc_gref, next_file);

Might want to add a comment that you are deleting only from ->next_file
b/c the do_cleanpup will take of deleting from the other list (that is
if the users is at zero).

> +			list_del(&gref->next_file);
> +			gref->users--;
> +			gref = n;
> +		}
> +	} else {
> +		rc = -EINVAL;
> +	}
> +
> +	do_cleanup();
> +
> +	spin_unlock(&gref_lock);
> +dealloc_grant_out:
> +	return rc;
> +}
> +
> +static long gntalloc_ioctl(struct file *filp, unsigned int cmd,
> +		unsigned long arg)
> +{
> +	struct gntalloc_file_private_data *priv = filp->private_data;
> +
> +	switch (cmd) {
> +	case IOCTL_GNTALLOC_ALLOC_GREF:
> +		return gntalloc_ioctl_alloc(priv, (void __user *)arg);
> +
> +	case IOCTL_GNTALLOC_DEALLOC_GREF:
> +		return gntalloc_ioctl_dealloc(priv, (void __user *)arg);
> +
> +	default:
> +		return -ENOIOCTLCMD;
> +	}
> +
> +	return 0;
> +}
> +
> +static void gntalloc_vma_close(struct vm_area_struct *vma)
> +{
> +	struct gntalloc_gref *gref = vma->vm_private_data;
> +	if (!gref)
> +		return;
> +
> +	spin_lock(&gref_lock);
> +	gref->users--;
> +	if (gref->users == 0)
> +		__del_gref(gref);

I just want to be convienced here that I am wrong.

If the 'ioctl_deallo' has not been done, what will if unmap this VMA?
Will it be OK to  yank the gref from gref_list while (and kfree it) while
it is still referenced in the filp->private_data? Or would end up trying
to derefence the *priv and go BOOM?


> +	spin_unlock(&gref_lock);
> +}
> +
> +static struct vm_operations_struct gntalloc_vmops = {
> +	.close = gntalloc_vma_close,
> +};
> +
> +static int gntalloc_mmap(struct file *filp, struct vm_area_struct *vma)
> +{
> +	struct gntalloc_file_private_data *priv = filp->private_data;
> +	struct gntalloc_gref *gref;
> +	int count = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
> +	int rv, i;
> +
> +	pr_debug("%s: priv %p, page %lu+%d\n", __func__,
> +		       priv, vma->vm_pgoff, count);
> +
> +	if (!(vma->vm_flags & VM_SHARED)) {
> +		printk(KERN_ERR "%s: Mapping must be shared.\n", __func__);
> +		return -EINVAL;
> +	}
> +
> +	spin_lock(&gref_lock);
> +	gref = find_grefs(priv, vma->vm_pgoff << PAGE_SHIFT, count);
> +	if (gref == NULL) {
> +		rv = -ENOENT;
> +		pr_debug("%s: Could not find grant reference",
> +				__func__);
> +		goto out_unlock;
> +	}
> +
> +	vma->vm_private_data = gref;
> +
> +	vma->vm_flags |= VM_RESERVED;
> +	vma->vm_flags |= VM_DONTCOPY;
> +	vma->vm_flags |= VM_IO;

Hmm, I don't think you need the VM_IO.

> +	vma->vm_flags |= VM_PFNMAP | VM_PFN_AT_MMAP;
> +
> +	vma->vm_ops = &gntalloc_vmops;
> +
> +	for (i = 0; i < count; i++) {
> +		gref->users++;
> +		rv = vm_insert_page(vma, vma->vm_start + i * PAGE_SIZE,
> +				gref->page);
> +		if (rv)
> +			goto out_unlock;
> +
> +		gref = list_entry(gref->next_file.next,
> +				struct gntalloc_gref, next_file);
> +	}
> +	rv = 0;
> +
> +out_unlock:
> +	spin_unlock(&gref_lock);
> +	return rv;
> +}
> +
> +static const struct file_operations gntalloc_fops = {
> +	.owner = THIS_MODULE,
> +	.open = gntalloc_open,
> +	.release = gntalloc_release,
> +	.unlocked_ioctl = gntalloc_ioctl,
> +	.mmap = gntalloc_mmap
> +};
> +
> +/*
> + * -------------------------------------
> + * Module creation/destruction.
> + * -------------------------------------
> + */
> +static struct miscdevice gntalloc_miscdev = {
> +	.minor	= MISC_DYNAMIC_MINOR,
> +	.name	= "xen/gntalloc",
> +	.fops	= &gntalloc_fops,
> +};
> +
> +static int __init gntalloc_init(void)
> +{
> +	int err;
> +
> +	if (!xen_domain())
> +		return -ENODEV;
> +
> +	err = misc_register(&gntalloc_miscdev);
> +	if (err != 0) {
> +		printk(KERN_ERR "Could not register misc gntalloc device\n");
> +		return err;
> +	}
> +
> +	pr_debug("Created grant allocation device at %d,%d\n",
> +			MISC_MAJOR, gntalloc_miscdev.minor);
> +
> +	return 0;
> +}
> +
> +static void __exit gntalloc_exit(void)
> +{
> +	misc_deregister(&gntalloc_miscdev);
> +}
> +
> +module_init(gntalloc_init);
> +module_exit(gntalloc_exit);
> +
> +MODULE_LICENSE("GPL");
> +MODULE_AUTHOR("Carter Weatherly <carter.weatherly@jhuapl.edu>, "
> +		"Daniel De Graaf <dgdegra@tycho.nsa.gov>");
> +MODULE_DESCRIPTION("User-space grant reference allocator driver");
> diff --git a/include/xen/gntalloc.h b/include/xen/gntalloc.h
> new file mode 100644
> index 0000000..e1d6d0f
> --- /dev/null
> +++ b/include/xen/gntalloc.h
> @@ -0,0 +1,70 @@
> +/******************************************************************************
> + * gntalloc.h
> + *

Copyright?

> + * Interface to /dev/xen/gntalloc.
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License version 2
> + * as published by the Free Software Foundation; or, when distributed
> + * separately from the Linux kernel or incorporated into other
> + * software packages, subject to the following license:
> + *
> + * Permission is hereby granted, free of charge, to any person obtaining a copy
> + * of this source file (the "Software"), to deal in the Software without
> + * restriction, including without limitation the rights to use, copy, modify,
> + * merge, publish, distribute, sublicense, and/or sell copies of the Software,
> + * and to permit persons to whom the Software is furnished to do so, subject to
> + * the following conditions:
> + *
> + * The above copyright notice and this permission notice shall be included in
> + * all copies or substantial portions of the Software.
> + *
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
> + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
> + * IN THE SOFTWARE.
> + */
> +
> +#ifndef __LINUX_PUBLIC_GNTALLOC_H__
> +#define __LINUX_PUBLIC_GNTALLOC_H__
> +
> +/*
> + * Allocates a new page and creates a new grant reference.
> + */
> +#define IOCTL_GNTALLOC_ALLOC_GREF \
> +_IOC(_IOC_NONE, 'G', 5, sizeof(struct ioctl_gntalloc_alloc_gref))
> +struct ioctl_gntalloc_alloc_gref {
> +	/* IN parameters */
> +	/* The ID of the domain to be given access to the grants. */
> +	uint16_t domid;
> +	/* Flags for this mapping */
> +	uint16_t flags;
> +	/* Number of pages to map */
> +	uint32_t count;
> +	/* OUT parameters */
> +	/* The offset to be used on a subsequent call to mmap(). */
> +	uint64_t index;
> +	/* The grant references of the newly created grant, one per page */
> +	/* Variable size, depending on count */
> +	uint32_t gref_ids[1];
> +};
> +
> +#define GNTALLOC_FLAG_WRITABLE 1
> +
> +/*
> + * Deallocates the grant reference, allowing the associated page to be freed if
> + * no other domains are using it.
> + */
> +#define IOCTL_GNTALLOC_DEALLOC_GREF \
> +_IOC(_IOC_NONE, 'G', 6, sizeof(struct ioctl_gntalloc_dealloc_gref))
> +struct ioctl_gntalloc_dealloc_gref {
> +	/* IN parameters */
> +	/* The offset returned in the map operation */
> +	uint64_t index;
> +	/* Number of references to unmap */
> +	uint32_t count;
> +};
> +#endif /* __LINUX_PUBLIC_GNTALLOC_H__ */
> -- 
> 1.7.3.4
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

  reply	other threads:[~2011-01-27 18:52 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-01-21 15:59 [SPAM] [PATCH v5] Userspace grant communication Daniel De Graaf
2011-01-21 15:59 ` [PATCH 1/6] xen-gntdev: Change page limit to be global instead of per-open Daniel De Graaf
2011-01-21 15:59 ` [PATCH 2/6] xen-gntdev: Use find_vma rather than iterating our vma list manually Daniel De Graaf
2011-01-21 15:59 ` [PATCH 3/6] xen-gntdev: Add reference counting to maps Daniel De Graaf
2011-01-21 15:59 ` [PATCH 4/6] xen-gntdev: Support mapping in HVM domains Daniel De Graaf
2011-01-27 18:52   ` Konrad Rzeszutek Wilk
2011-01-27 19:26     ` Daniel De Graaf
2011-03-04 15:57     ` Ian Campbell
2011-03-04 16:34       ` Daniel De Graaf
2011-01-21 15:59 ` [PATCH 5/6] xen-gntalloc: Userspace grant allocation driver Daniel De Graaf
2011-01-27 18:52   ` Konrad Rzeszutek Wilk [this message]
2011-01-27 19:23     ` Konrad Rzeszutek Wilk
2011-01-27 19:51       ` Daniel De Graaf
2011-01-27 20:55     ` Daniel De Graaf
2011-01-27 21:29       ` Konrad Rzeszutek Wilk
2011-01-21 15:59 ` [PATCH 6/6] xen/gntalloc, gntdev: Add unmap notify ioctl Daniel De Graaf
2011-01-27 19:20   ` Konrad Rzeszutek Wilk
2011-01-27 20:09     ` Daniel De Graaf
  -- strict thread matches above, loose matches on Subject: below --
2011-02-03 17:18 [PATCH v6] Userspace grant communication Daniel De Graaf
2011-02-03 17:19 ` [PATCH 5/6] xen-gntalloc: Userspace grant allocation driver Daniel De Graaf
2011-02-08 22:48   ` Konrad Rzeszutek Wilk
2011-02-09 18:52     ` Daniel De Graaf
2010-12-14 14:55 [PATCH v2] Userspace grant communication Daniel De Graaf
2010-12-14 14:55 ` [PATCH 5/6] xen-gntalloc: Userspace grant allocation driver Daniel De Graaf
2010-12-14 21:42   ` Jeremy Fitzhardinge
2010-12-14 22:06     ` Daniel De Graaf
2010-12-14 22:40       ` Jeremy Fitzhardinge
2010-12-15 14:18         ` Daniel De Graaf
2010-12-16  1:05           ` Jeremy Fitzhardinge
2010-12-16 15:22             ` Daniel De Graaf
2010-12-16 19:14               ` Jeremy Fitzhardinge

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=20110127185210.GA27853@dumpdata.com \
    --to=konrad.wilk@oracle.com \
    --cc=Ian.Campbell@citrix.com \
    --cc=dgdegra@tycho.nsa.gov \
    --cc=jeremy@goop.org \
    --cc=xen-devel@lists.xensource.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).