linux/drivers/gpu/drm/drm_gem.c

1044 lines
30 KiB
C
Raw Normal View History

/*
* Copyright © 2008 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (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 (including the next
* paragraph) 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.
*
* Authors:
* Eric Anholt <eric@anholt.net>
*
*/
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/uaccess.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/module.h>
#include <linux/mman.h>
#include <linux/pagemap.h>
#include <linux/shmem_fs.h>
#include <linux/dma-buf.h>
#include <drm/drmP.h>
#include <drm/drm_vma_manager.h>
#include <drm/drm_gem.h>
#include "drm_internal.h"
/** @file drm_gem.c
*
* This file provides some of the base ioctls and library routines for
* the graphics memory manager implemented by each device driver.
*
* Because various devices have different requirements in terms of
* synchronization and migration strategies, implementing that is left up to
* the driver, and all that the general API provides should be generic --
* allocating objects, reading/writing data with the cpu, freeing objects.
* Even there, platform-dependent optimizations for reading/writing data with
* the CPU mean we'll likely hook those out to driver-specific calls. However,
* the DRI2 implementation wants to have at least allocate/mmap be generic.
*
* The goal was to have swap-backed object allocation managed through
* struct file. However, file descriptors as handles to a struct file have
* two major failings:
* - Process limits prevent more than 1024 or so being used at a time by
* default.
* - Inability to allocate high fds will aggravate the X Server's select()
* handling, and likely that of many GL client applications as well.
*
* This led to a plan of using our own integer IDs (called handles, following
* DRM terminology) to mimic fds, and implement the fd syscalls we need as
* ioctls. The objects themselves will still include the struct file so
* that we can transition to fds if the required kernel infrastructure shows
* up at a later date, and as our interface with shmfs for memory allocation.
*/
/*
* We make up offsets for buffer objects so we can recognize them at
* mmap time.
*/
/* pgoff in mmap is an unsigned long, so we need to make sure that
* the faked up offset will fit
*/
#if BITS_PER_LONG == 64
#define DRM_FILE_PAGE_OFFSET_START ((0xFFFFFFFFUL >> PAGE_SHIFT) + 1)
#define DRM_FILE_PAGE_OFFSET_SIZE ((0xFFFFFFFFUL >> PAGE_SHIFT) * 16)
#else
#define DRM_FILE_PAGE_OFFSET_START ((0xFFFFFFFUL >> PAGE_SHIFT) + 1)
#define DRM_FILE_PAGE_OFFSET_SIZE ((0xFFFFFFFUL >> PAGE_SHIFT) * 16)
#endif
/**
* drm_gem_init - Initialize the GEM device fields
* @dev: drm_devic structure to initialize
*/
int
drm_gem_init(struct drm_device *dev)
{
struct drm_vma_offset_manager *vma_offset_manager;
mutex_init(&dev->object_name_lock);
idr_init(&dev->object_name_idr);
vma_offset_manager = kzalloc(sizeof(*vma_offset_manager), GFP_KERNEL);
if (!vma_offset_manager) {
DRM_ERROR("out of memory\n");
return -ENOMEM;
}
dev->vma_offset_manager = vma_offset_manager;
drm_vma_offset_manager_init(vma_offset_manager,
DRM_FILE_PAGE_OFFSET_START,
DRM_FILE_PAGE_OFFSET_SIZE);
return 0;
}
void
drm_gem_destroy(struct drm_device *dev)
{
drm_vma_offset_manager_destroy(dev->vma_offset_manager);
kfree(dev->vma_offset_manager);
dev->vma_offset_manager = NULL;
}
/**
* drm_gem_object_init - initialize an allocated shmem-backed GEM object
* @dev: drm_device the object should be initialized for
* @obj: drm_gem_object to initialize
* @size: object size
*
* Initialize an already allocated GEM object of the specified size with
* shmfs backing store.
*/
int drm_gem_object_init(struct drm_device *dev,
struct drm_gem_object *obj, size_t size)
{
struct file *filp;
drm_gem_private_object_init(dev, obj, size);
filp = shmem_file_setup("drm mm object", size, VM_NORESERVE);
if (IS_ERR(filp))
return PTR_ERR(filp);
obj->filp = filp;
return 0;
}
EXPORT_SYMBOL(drm_gem_object_init);
/**
* drm_gem_private_object_init - initialize an allocated private GEM object
* @dev: drm_device the object should be initialized for
* @obj: drm_gem_object to initialize
* @size: object size
*
* Initialize an already allocated GEM object of the specified size with
* no GEM provided backing store. Instead the caller is responsible for
* backing the object and handling it.
*/
void drm_gem_private_object_init(struct drm_device *dev,
struct drm_gem_object *obj, size_t size)
{
BUG_ON((size & (PAGE_SIZE - 1)) != 0);
obj->dev = dev;
obj->filp = NULL;
kref_init(&obj->refcount);
drm/gem: fix up flink name create race This is the 2nd attempt, I've always been a bit dissatisified with the tricky nature of the first one: http://lists.freedesktop.org/archives/dri-devel/2012-July/025451.html The issue is that the flink ioctl can race with calling gem_close on the last gem handle. In that case we'll end up with a zero handle count, but an flink name (and it's corresponding reference). Which results in a neat space leak. In my first attempt I've solved this by rechecking the handle count. But fundamentally the issue is that ->handle_count isn't your usual refcount - it can be resurrected from 0 among other things. For those special beasts atomic_t often suggest way more ordering that it actually guarantees. To prevent being tricked by those hairy semantics take the easy way out and simply protect the handle with the existing dev->object_name_lock. With that change implemented it's dead easy to fix the flink vs. gem close reace: When we try to create the name we simply have to check whether there's still officially a gem handle around and if not refuse to create the flink name. Since the handle count decrement and flink name destruction is now also protected by that lock the reace is gone and we can't ever leak the flink reference again. Outside of the drm core only the exynos driver looks at the handle count, and tbh I have no idea why (it's just for debug dmesg output luckily). I've considered inlining the drm_gem_object_handle_free, but I plan to add more name-like things (like the exported dma_buf) to this scheme, so it's clearer to leave the handle freeing in its own function. This is exercised by the new gem_flink_race i-g-t testcase, which on my snb leaks gem objects at a rate of roughly 1k objects/s. v2: Fix up the error path handling in handle_create and make it more robust by simply calling object_handle_unreference. v3: Fix up the handle_unreference logic bug - atomic_dec_and_test retursn 1 for 0. Oops. v4: Squash in inlining of drm_gem_object_handle_reference as suggested by Dave Airlie and add a note that we now have a testcase. Cc: Dave Airlie <airlied@gmail.com> Cc: Inki Dae <inki.dae@samsung.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:37 +00:00
obj->handle_count = 0;
obj->size = size;
drm_vma_node_reset(&obj->vma_node);
}
EXPORT_SYMBOL(drm_gem_private_object_init);
static void
drm_gem_remove_prime_handles(struct drm_gem_object *obj, struct drm_file *filp)
{
drm/prime: proper locking+refcounting for obj->dma_buf link The export dma-buf cache is semantically similar to an flink name. So semantically it makes sense to treat it the same and remove the name (i.e. the dma_buf pointer) and its references when the last gem handle disappears. Again we need to be careful, but double so: Not just could someone race and export with a gem close ioctl (so we need to recheck obj->handle_count again when assigning the new name), but multiple exports can also race against each another. This is prevented by holding the dev->object_name_lock across the entire section which touches obj->dma_buf. With the new scheme we also need to reinstate the obj->dma_buf link at import time (in case the only reference userspace has held in-between was through the dma-buf fd and not through any native gem handle). For simplicity we don't check whether it's a native object but unconditionally set up that link - with the new scheme of removing the obj->dma_buf reference when the last handle disappears we can do that. To make it clear that this is not just for exported buffers anymore als rename it from export_dma_buf to dma_buf. To make sure that now one can race a fd_to_handle or handle_to_fd with gem_close we use the same tricks as in flink of extending the dev->object_name_locking critical section. With this change we finally have a guaranteed 1:1 relationship (at least for native objects) between gem objects and dma-bufs, even accounting for races (which can happen since the dma-buf itself holds a reference while in-flight). This prevent igt/prime_self_import/export-vs-gem_close-race from Oopsing the kernel. There is still a leak though since the per-file priv dma-buf/handle cache handling is racy. That will be fixed in a later patch. v2: Remove the bogus dma_buf_put from the export_and_register_object failure path if we've raced with the handle count dropping to 0. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:46 +00:00
/*
* Note: obj->dma_buf can't disappear as long as we still hold a
* handle reference in obj->handle_count.
*/
drm/prime: Always add exported buffers to the handle cache ... not only when the dma-buf is freshly created. In contrived examples someone else could have exported/imported the dma-buf already and handed us the gem object with a flink name. If such on object gets reexported as a dma_buf we won't have it in the handle cache already, which breaks the guarantee that for dma-buf imports we always hand back an existing handle if there is one. This is exercised by igt/prime_self_import/with_one_bo_two_files Now if we extend the locked sections just a notch more we can also plug th racy buf/handle cache setup in handle_to_fd: If evil userspace races a concurrent gem close against a prime export operation we can end up tearing down the gem handle before the dma buf handle cache is set up. When handle_to_fd gets around to adding the handle to the cache there will be no one left to clean it up, effectily leaking the bo (and the dma-buf, since the handle cache holds a ref on the dma-buf): Thread A Thread B handle_to_fd: lookup gem object from handle creates new dma_buf gem_close on the same handle obj->dma_buf is set, but file priv buf handle cache has no entry obj->handle_count drops to 0 drm_prime_add_buf_handle sets up the handle cache -> We have a dma-buf reference in the handle cache, but since the handle_count of the gem object already dropped to 0 no on will clean it up. When closing the drm device fd we'll hit the WARN_ON in drm_prime_destroy_file_private. The important change is to extend the critical section of the filp->prime.lock to cover the gem handle lookup. This serializes with a concurrent gem handle close. This leak is exercised by igt/prime_self_import/export-vs-gem_close-race Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:49 +00:00
mutex_lock(&filp->prime.lock);
drm/prime: proper locking+refcounting for obj->dma_buf link The export dma-buf cache is semantically similar to an flink name. So semantically it makes sense to treat it the same and remove the name (i.e. the dma_buf pointer) and its references when the last gem handle disappears. Again we need to be careful, but double so: Not just could someone race and export with a gem close ioctl (so we need to recheck obj->handle_count again when assigning the new name), but multiple exports can also race against each another. This is prevented by holding the dev->object_name_lock across the entire section which touches obj->dma_buf. With the new scheme we also need to reinstate the obj->dma_buf link at import time (in case the only reference userspace has held in-between was through the dma-buf fd and not through any native gem handle). For simplicity we don't check whether it's a native object but unconditionally set up that link - with the new scheme of removing the obj->dma_buf reference when the last handle disappears we can do that. To make it clear that this is not just for exported buffers anymore als rename it from export_dma_buf to dma_buf. To make sure that now one can race a fd_to_handle or handle_to_fd with gem_close we use the same tricks as in flink of extending the dev->object_name_locking critical section. With this change we finally have a guaranteed 1:1 relationship (at least for native objects) between gem objects and dma-bufs, even accounting for races (which can happen since the dma-buf itself holds a reference while in-flight). This prevent igt/prime_self_import/export-vs-gem_close-race from Oopsing the kernel. There is still a leak though since the per-file priv dma-buf/handle cache handling is racy. That will be fixed in a later patch. v2: Remove the bogus dma_buf_put from the export_and_register_object failure path if we've raced with the handle count dropping to 0. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:46 +00:00
if (obj->dma_buf) {
drm/prime: Always add exported buffers to the handle cache ... not only when the dma-buf is freshly created. In contrived examples someone else could have exported/imported the dma-buf already and handed us the gem object with a flink name. If such on object gets reexported as a dma_buf we won't have it in the handle cache already, which breaks the guarantee that for dma-buf imports we always hand back an existing handle if there is one. This is exercised by igt/prime_self_import/with_one_bo_two_files Now if we extend the locked sections just a notch more we can also plug th racy buf/handle cache setup in handle_to_fd: If evil userspace races a concurrent gem close against a prime export operation we can end up tearing down the gem handle before the dma buf handle cache is set up. When handle_to_fd gets around to adding the handle to the cache there will be no one left to clean it up, effectily leaking the bo (and the dma-buf, since the handle cache holds a ref on the dma-buf): Thread A Thread B handle_to_fd: lookup gem object from handle creates new dma_buf gem_close on the same handle obj->dma_buf is set, but file priv buf handle cache has no entry obj->handle_count drops to 0 drm_prime_add_buf_handle sets up the handle cache -> We have a dma-buf reference in the handle cache, but since the handle_count of the gem object already dropped to 0 no on will clean it up. When closing the drm device fd we'll hit the WARN_ON in drm_prime_destroy_file_private. The important change is to extend the critical section of the filp->prime.lock to cover the gem handle lookup. This serializes with a concurrent gem handle close. This leak is exercised by igt/prime_self_import/export-vs-gem_close-race Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:49 +00:00
drm_prime_remove_buf_handle_locked(&filp->prime,
obj->dma_buf);
}
drm/prime: Always add exported buffers to the handle cache ... not only when the dma-buf is freshly created. In contrived examples someone else could have exported/imported the dma-buf already and handed us the gem object with a flink name. If such on object gets reexported as a dma_buf we won't have it in the handle cache already, which breaks the guarantee that for dma-buf imports we always hand back an existing handle if there is one. This is exercised by igt/prime_self_import/with_one_bo_two_files Now if we extend the locked sections just a notch more we can also plug th racy buf/handle cache setup in handle_to_fd: If evil userspace races a concurrent gem close against a prime export operation we can end up tearing down the gem handle before the dma buf handle cache is set up. When handle_to_fd gets around to adding the handle to the cache there will be no one left to clean it up, effectily leaking the bo (and the dma-buf, since the handle cache holds a ref on the dma-buf): Thread A Thread B handle_to_fd: lookup gem object from handle creates new dma_buf gem_close on the same handle obj->dma_buf is set, but file priv buf handle cache has no entry obj->handle_count drops to 0 drm_prime_add_buf_handle sets up the handle cache -> We have a dma-buf reference in the handle cache, but since the handle_count of the gem object already dropped to 0 no on will clean it up. When closing the drm device fd we'll hit the WARN_ON in drm_prime_destroy_file_private. The important change is to extend the critical section of the filp->prime.lock to cover the gem handle lookup. This serializes with a concurrent gem handle close. This leak is exercised by igt/prime_self_import/export-vs-gem_close-race Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:49 +00:00
mutex_unlock(&filp->prime.lock);
}
/**
* drm_gem_object_handle_free - release resources bound to userspace handles
* @obj: GEM object to clean up.
*
* Called after the last handle to the object has been closed
*
* Removes any name for the object. Note that this must be
* called before drm_gem_object_free or we'll be touching
* freed memory
*/
static void drm_gem_object_handle_free(struct drm_gem_object *obj)
{
struct drm_device *dev = obj->dev;
/* Remove any name for this object */
if (obj->name) {
idr_remove(&dev->object_name_idr, obj->name);
obj->name = 0;
drm/gem: fix up flink name create race This is the 2nd attempt, I've always been a bit dissatisified with the tricky nature of the first one: http://lists.freedesktop.org/archives/dri-devel/2012-July/025451.html The issue is that the flink ioctl can race with calling gem_close on the last gem handle. In that case we'll end up with a zero handle count, but an flink name (and it's corresponding reference). Which results in a neat space leak. In my first attempt I've solved this by rechecking the handle count. But fundamentally the issue is that ->handle_count isn't your usual refcount - it can be resurrected from 0 among other things. For those special beasts atomic_t often suggest way more ordering that it actually guarantees. To prevent being tricked by those hairy semantics take the easy way out and simply protect the handle with the existing dev->object_name_lock. With that change implemented it's dead easy to fix the flink vs. gem close reace: When we try to create the name we simply have to check whether there's still officially a gem handle around and if not refuse to create the flink name. Since the handle count decrement and flink name destruction is now also protected by that lock the reace is gone and we can't ever leak the flink reference again. Outside of the drm core only the exynos driver looks at the handle count, and tbh I have no idea why (it's just for debug dmesg output luckily). I've considered inlining the drm_gem_object_handle_free, but I plan to add more name-like things (like the exported dma_buf) to this scheme, so it's clearer to leave the handle freeing in its own function. This is exercised by the new gem_flink_race i-g-t testcase, which on my snb leaks gem objects at a rate of roughly 1k objects/s. v2: Fix up the error path handling in handle_create and make it more robust by simply calling object_handle_unreference. v3: Fix up the handle_unreference logic bug - atomic_dec_and_test retursn 1 for 0. Oops. v4: Squash in inlining of drm_gem_object_handle_reference as suggested by Dave Airlie and add a note that we now have a testcase. Cc: Dave Airlie <airlied@gmail.com> Cc: Inki Dae <inki.dae@samsung.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:37 +00:00
}
}
drm/prime: proper locking+refcounting for obj->dma_buf link The export dma-buf cache is semantically similar to an flink name. So semantically it makes sense to treat it the same and remove the name (i.e. the dma_buf pointer) and its references when the last gem handle disappears. Again we need to be careful, but double so: Not just could someone race and export with a gem close ioctl (so we need to recheck obj->handle_count again when assigning the new name), but multiple exports can also race against each another. This is prevented by holding the dev->object_name_lock across the entire section which touches obj->dma_buf. With the new scheme we also need to reinstate the obj->dma_buf link at import time (in case the only reference userspace has held in-between was through the dma-buf fd and not through any native gem handle). For simplicity we don't check whether it's a native object but unconditionally set up that link - with the new scheme of removing the obj->dma_buf reference when the last handle disappears we can do that. To make it clear that this is not just for exported buffers anymore als rename it from export_dma_buf to dma_buf. To make sure that now one can race a fd_to_handle or handle_to_fd with gem_close we use the same tricks as in flink of extending the dev->object_name_locking critical section. With this change we finally have a guaranteed 1:1 relationship (at least for native objects) between gem objects and dma-bufs, even accounting for races (which can happen since the dma-buf itself holds a reference while in-flight). This prevent igt/prime_self_import/export-vs-gem_close-race from Oopsing the kernel. There is still a leak though since the per-file priv dma-buf/handle cache handling is racy. That will be fixed in a later patch. v2: Remove the bogus dma_buf_put from the export_and_register_object failure path if we've raced with the handle count dropping to 0. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:46 +00:00
static void drm_gem_object_exported_dma_buf_free(struct drm_gem_object *obj)
{
/* Unbreak the reference cycle if we have an exported dma_buf. */
if (obj->dma_buf) {
dma_buf_put(obj->dma_buf);
obj->dma_buf = NULL;
}
}
static void
drm_gem_object_handle_put_unlocked(struct drm_gem_object *obj)
{
struct drm_device *dev = obj->dev;
bool final = false;
drm/gem: fix up flink name create race This is the 2nd attempt, I've always been a bit dissatisified with the tricky nature of the first one: http://lists.freedesktop.org/archives/dri-devel/2012-July/025451.html The issue is that the flink ioctl can race with calling gem_close on the last gem handle. In that case we'll end up with a zero handle count, but an flink name (and it's corresponding reference). Which results in a neat space leak. In my first attempt I've solved this by rechecking the handle count. But fundamentally the issue is that ->handle_count isn't your usual refcount - it can be resurrected from 0 among other things. For those special beasts atomic_t often suggest way more ordering that it actually guarantees. To prevent being tricked by those hairy semantics take the easy way out and simply protect the handle with the existing dev->object_name_lock. With that change implemented it's dead easy to fix the flink vs. gem close reace: When we try to create the name we simply have to check whether there's still officially a gem handle around and if not refuse to create the flink name. Since the handle count decrement and flink name destruction is now also protected by that lock the reace is gone and we can't ever leak the flink reference again. Outside of the drm core only the exynos driver looks at the handle count, and tbh I have no idea why (it's just for debug dmesg output luckily). I've considered inlining the drm_gem_object_handle_free, but I plan to add more name-like things (like the exported dma_buf) to this scheme, so it's clearer to leave the handle freeing in its own function. This is exercised by the new gem_flink_race i-g-t testcase, which on my snb leaks gem objects at a rate of roughly 1k objects/s. v2: Fix up the error path handling in handle_create and make it more robust by simply calling object_handle_unreference. v3: Fix up the handle_unreference logic bug - atomic_dec_and_test retursn 1 for 0. Oops. v4: Squash in inlining of drm_gem_object_handle_reference as suggested by Dave Airlie and add a note that we now have a testcase. Cc: Dave Airlie <airlied@gmail.com> Cc: Inki Dae <inki.dae@samsung.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:37 +00:00
if (WARN_ON(obj->handle_count == 0))
return;
/*
* Must bump handle count first as this may be the last
* ref, in which case the object would disappear before we
* checked for a name
*/
mutex_lock(&dev->object_name_lock);
drm/prime: proper locking+refcounting for obj->dma_buf link The export dma-buf cache is semantically similar to an flink name. So semantically it makes sense to treat it the same and remove the name (i.e. the dma_buf pointer) and its references when the last gem handle disappears. Again we need to be careful, but double so: Not just could someone race and export with a gem close ioctl (so we need to recheck obj->handle_count again when assigning the new name), but multiple exports can also race against each another. This is prevented by holding the dev->object_name_lock across the entire section which touches obj->dma_buf. With the new scheme we also need to reinstate the obj->dma_buf link at import time (in case the only reference userspace has held in-between was through the dma-buf fd and not through any native gem handle). For simplicity we don't check whether it's a native object but unconditionally set up that link - with the new scheme of removing the obj->dma_buf reference when the last handle disappears we can do that. To make it clear that this is not just for exported buffers anymore als rename it from export_dma_buf to dma_buf. To make sure that now one can race a fd_to_handle or handle_to_fd with gem_close we use the same tricks as in flink of extending the dev->object_name_locking critical section. With this change we finally have a guaranteed 1:1 relationship (at least for native objects) between gem objects and dma-bufs, even accounting for races (which can happen since the dma-buf itself holds a reference while in-flight). This prevent igt/prime_self_import/export-vs-gem_close-race from Oopsing the kernel. There is still a leak though since the per-file priv dma-buf/handle cache handling is racy. That will be fixed in a later patch. v2: Remove the bogus dma_buf_put from the export_and_register_object failure path if we've raced with the handle count dropping to 0. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:46 +00:00
if (--obj->handle_count == 0) {
drm_gem_object_handle_free(obj);
drm/prime: proper locking+refcounting for obj->dma_buf link The export dma-buf cache is semantically similar to an flink name. So semantically it makes sense to treat it the same and remove the name (i.e. the dma_buf pointer) and its references when the last gem handle disappears. Again we need to be careful, but double so: Not just could someone race and export with a gem close ioctl (so we need to recheck obj->handle_count again when assigning the new name), but multiple exports can also race against each another. This is prevented by holding the dev->object_name_lock across the entire section which touches obj->dma_buf. With the new scheme we also need to reinstate the obj->dma_buf link at import time (in case the only reference userspace has held in-between was through the dma-buf fd and not through any native gem handle). For simplicity we don't check whether it's a native object but unconditionally set up that link - with the new scheme of removing the obj->dma_buf reference when the last handle disappears we can do that. To make it clear that this is not just for exported buffers anymore als rename it from export_dma_buf to dma_buf. To make sure that now one can race a fd_to_handle or handle_to_fd with gem_close we use the same tricks as in flink of extending the dev->object_name_locking critical section. With this change we finally have a guaranteed 1:1 relationship (at least for native objects) between gem objects and dma-bufs, even accounting for races (which can happen since the dma-buf itself holds a reference while in-flight). This prevent igt/prime_self_import/export-vs-gem_close-race from Oopsing the kernel. There is still a leak though since the per-file priv dma-buf/handle cache handling is racy. That will be fixed in a later patch. v2: Remove the bogus dma_buf_put from the export_and_register_object failure path if we've raced with the handle count dropping to 0. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:46 +00:00
drm_gem_object_exported_dma_buf_free(obj);
final = true;
drm/prime: proper locking+refcounting for obj->dma_buf link The export dma-buf cache is semantically similar to an flink name. So semantically it makes sense to treat it the same and remove the name (i.e. the dma_buf pointer) and its references when the last gem handle disappears. Again we need to be careful, but double so: Not just could someone race and export with a gem close ioctl (so we need to recheck obj->handle_count again when assigning the new name), but multiple exports can also race against each another. This is prevented by holding the dev->object_name_lock across the entire section which touches obj->dma_buf. With the new scheme we also need to reinstate the obj->dma_buf link at import time (in case the only reference userspace has held in-between was through the dma-buf fd and not through any native gem handle). For simplicity we don't check whether it's a native object but unconditionally set up that link - with the new scheme of removing the obj->dma_buf reference when the last handle disappears we can do that. To make it clear that this is not just for exported buffers anymore als rename it from export_dma_buf to dma_buf. To make sure that now one can race a fd_to_handle or handle_to_fd with gem_close we use the same tricks as in flink of extending the dev->object_name_locking critical section. With this change we finally have a guaranteed 1:1 relationship (at least for native objects) between gem objects and dma-bufs, even accounting for races (which can happen since the dma-buf itself holds a reference while in-flight). This prevent igt/prime_self_import/export-vs-gem_close-race from Oopsing the kernel. There is still a leak though since the per-file priv dma-buf/handle cache handling is racy. That will be fixed in a later patch. v2: Remove the bogus dma_buf_put from the export_and_register_object failure path if we've raced with the handle count dropping to 0. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:46 +00:00
}
mutex_unlock(&dev->object_name_lock);
drm/gem: fix up flink name create race This is the 2nd attempt, I've always been a bit dissatisified with the tricky nature of the first one: http://lists.freedesktop.org/archives/dri-devel/2012-July/025451.html The issue is that the flink ioctl can race with calling gem_close on the last gem handle. In that case we'll end up with a zero handle count, but an flink name (and it's corresponding reference). Which results in a neat space leak. In my first attempt I've solved this by rechecking the handle count. But fundamentally the issue is that ->handle_count isn't your usual refcount - it can be resurrected from 0 among other things. For those special beasts atomic_t often suggest way more ordering that it actually guarantees. To prevent being tricked by those hairy semantics take the easy way out and simply protect the handle with the existing dev->object_name_lock. With that change implemented it's dead easy to fix the flink vs. gem close reace: When we try to create the name we simply have to check whether there's still officially a gem handle around and if not refuse to create the flink name. Since the handle count decrement and flink name destruction is now also protected by that lock the reace is gone and we can't ever leak the flink reference again. Outside of the drm core only the exynos driver looks at the handle count, and tbh I have no idea why (it's just for debug dmesg output luckily). I've considered inlining the drm_gem_object_handle_free, but I plan to add more name-like things (like the exported dma_buf) to this scheme, so it's clearer to leave the handle freeing in its own function. This is exercised by the new gem_flink_race i-g-t testcase, which on my snb leaks gem objects at a rate of roughly 1k objects/s. v2: Fix up the error path handling in handle_create and make it more robust by simply calling object_handle_unreference. v3: Fix up the handle_unreference logic bug - atomic_dec_and_test retursn 1 for 0. Oops. v4: Squash in inlining of drm_gem_object_handle_reference as suggested by Dave Airlie and add a note that we now have a testcase. Cc: Dave Airlie <airlied@gmail.com> Cc: Inki Dae <inki.dae@samsung.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:37 +00:00
if (final)
drm_gem_object_put_unlocked(obj);
}
/*
* Called at device or object close to release the file's
* handle references on objects.
*/
static int
drm_gem_object_release_handle(int id, void *ptr, void *data)
{
struct drm_file *file_priv = data;
struct drm_gem_object *obj = ptr;
struct drm_device *dev = obj->dev;
if (drm_core_check_feature(dev, DRIVER_PRIME))
drm_gem_remove_prime_handles(obj, file_priv);
drm_vma_node_revoke(&obj->vma_node, file_priv);
if (dev->driver->gem_close_object)
dev->driver->gem_close_object(obj, file_priv);
drm_gem_object_handle_put_unlocked(obj);
return 0;
}
/**
* drm_gem_handle_delete - deletes the given file-private handle
* @filp: drm file-private structure to use for the handle look up
* @handle: userspace handle to delete
*
* Removes the GEM handle from the @filp lookup table which has been added with
* drm_gem_handle_create(). If this is the last handle also cleans up linked
* resources like GEM names.
*/
int
drm_gem_handle_delete(struct drm_file *filp, u32 handle)
{
struct drm_gem_object *obj;
/* This is gross. The idr system doesn't let us try a delete and
* return an error code. It just spews if you fail at deleting.
* So, we have to grab a lock around finding the object and then
* doing the delete on it and dropping the refcount, or the user
* could race us to double-decrement the refcount and cause a
* use-after-free later. Given the frequency of our handle lookups,
* we may want to use ida for number allocation and a hash table
* for the pointers, anyway.
*/
spin_lock(&filp->table_lock);
/* Check if we currently have a reference on the object */
drm: Release driver references to handle before making it available again When userspace closes a handle, we remove it from the file->object_idr and then tell the driver to drop its references to that file/handle. However, as the file/handle is already available again for reuse, it may be reallocated back to userspace and active on a new object before the driver has had a chance to drop the old file/handle references. Whilst calling back into the driver, we have to drop the file->table_lock spinlock and so to prevent reusing the closed handle we mark that handle as stale in the idr, perform the callback and then remove the handle. We set the stale handle to point to the NULL object, then any idr_find() whilst the driver is removing the handle will return NULL, just as if the handle is already removed from idr. Note: This will be used to have a direct handle -> vma lookup table, instead of first a handle -> obj lookup, and then an (obj, vm) -> vma lookup. v2: Use NULL rather than an ERR_PTR to avoid having to adjust callers. idr_alloc() tracks existing handles using an internal bitmap, so we are free to use the NULL object as our stale identifier. v3: Needed to update the return value check after changing from using the stale error pointer to NULL. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: dri-devel@lists.freedesktop.org Cc: David Airlie <airlied@linux.ie> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Thierry Reding <treding@nvidia.com> [danvet: Add note about the use-case.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1460721308-32405-1-git-send-email-chris@chris-wilson.co.uk
2016-04-15 11:55:08 +00:00
obj = idr_replace(&filp->object_idr, NULL, handle);
spin_unlock(&filp->table_lock);
if (IS_ERR_OR_NULL(obj))
return -EINVAL;
drm: Release driver references to handle before making it available again When userspace closes a handle, we remove it from the file->object_idr and then tell the driver to drop its references to that file/handle. However, as the file/handle is already available again for reuse, it may be reallocated back to userspace and active on a new object before the driver has had a chance to drop the old file/handle references. Whilst calling back into the driver, we have to drop the file->table_lock spinlock and so to prevent reusing the closed handle we mark that handle as stale in the idr, perform the callback and then remove the handle. We set the stale handle to point to the NULL object, then any idr_find() whilst the driver is removing the handle will return NULL, just as if the handle is already removed from idr. Note: This will be used to have a direct handle -> vma lookup table, instead of first a handle -> obj lookup, and then an (obj, vm) -> vma lookup. v2: Use NULL rather than an ERR_PTR to avoid having to adjust callers. idr_alloc() tracks existing handles using an internal bitmap, so we are free to use the NULL object as our stale identifier. v3: Needed to update the return value check after changing from using the stale error pointer to NULL. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: dri-devel@lists.freedesktop.org Cc: David Airlie <airlied@linux.ie> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Thierry Reding <treding@nvidia.com> [danvet: Add note about the use-case.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1460721308-32405-1-git-send-email-chris@chris-wilson.co.uk
2016-04-15 11:55:08 +00:00
/* Release driver's reference and decrement refcount. */
drm_gem_object_release_handle(handle, obj, filp);
/* And finally make the handle available for future allocations. */
spin_lock(&filp->table_lock);
idr_remove(&filp->object_idr, handle);
spin_unlock(&filp->table_lock);
return 0;
}
EXPORT_SYMBOL(drm_gem_handle_delete);
/**
* drm_gem_dumb_map_offset - return the fake mmap offset for a gem object
* @file: drm file-private structure containing the gem object
* @dev: corresponding drm_device
* @handle: gem object handle
* @offset: return location for the fake mmap offset
*
* This implements the &drm_driver.dumb_map_offset kms driver callback for
* drivers which use gem to manage their backing storage.
*
* Returns:
* 0 on success or a negative error code on failure.
*/
int drm_gem_dumb_map_offset(struct drm_file *file, struct drm_device *dev,
u32 handle, u64 *offset)
{
struct drm_gem_object *obj;
int ret;
obj = drm_gem_object_lookup(file, handle);
if (!obj)
return -ENOENT;
ret = drm_gem_create_mmap_offset(obj);
if (ret)
goto out;
*offset = drm_vma_node_offset_addr(&obj->vma_node);
out:
drm_gem_object_put_unlocked(obj);
return ret;
}
EXPORT_SYMBOL_GPL(drm_gem_dumb_map_offset);
/**
* drm_gem_dumb_destroy - dumb fb callback helper for gem based drivers
* @file: drm file-private structure to remove the dumb handle from
* @dev: corresponding drm_device
* @handle: the dumb handle to remove
*
* This implements the &drm_driver.dumb_destroy kms driver callback for drivers
* which use gem to manage their backing storage.
*/
int drm_gem_dumb_destroy(struct drm_file *file,
struct drm_device *dev,
uint32_t handle)
{
return drm_gem_handle_delete(file, handle);
}
EXPORT_SYMBOL(drm_gem_dumb_destroy);
/**
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
* drm_gem_handle_create_tail - internal functions to create a handle
* @file_priv: drm file-private structure to register the handle for
* @obj: object to register
* @handlep: pointer to return the created handle to the caller
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
*
* This expects the &drm_device.object_name_lock to be held already and will
* drop it before returning. Used to avoid races in establishing new handles
* when importing an object from either an flink name or a dma-buf.
*
* Handles must be release again through drm_gem_handle_delete(). This is done
* when userspace closes @file_priv for all attached handles, or through the
* GEM_CLOSE ioctl for individual handles.
*/
int
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
drm_gem_handle_create_tail(struct drm_file *file_priv,
struct drm_gem_object *obj,
u32 *handlep)
{
struct drm_device *dev = obj->dev;
u32 handle;
int ret;
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
WARN_ON(!mutex_is_locked(&dev->object_name_lock));
if (obj->handle_count++ == 0)
drm_gem_object_get(obj);
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
/*
* Get the user-visible handle using idr. Preload and perform
* allocation under our spinlock.
*/
idr_preload(GFP_KERNEL);
spin_lock(&file_priv->table_lock);
ret = idr_alloc(&file_priv->object_idr, obj, 1, 0, GFP_NOWAIT);
spin_unlock(&file_priv->table_lock);
idr_preload_end();
mutex_unlock(&dev->object_name_lock);
if (ret < 0)
goto err_unref;
handle = ret;
ret = drm_vma_node_allow(&obj->vma_node, file_priv);
if (ret)
goto err_remove;
if (dev->driver->gem_open_object) {
ret = dev->driver->gem_open_object(obj, file_priv);
if (ret)
goto err_revoke;
}
*handlep = handle;
return 0;
err_revoke:
drm_vma_node_revoke(&obj->vma_node, file_priv);
err_remove:
spin_lock(&file_priv->table_lock);
idr_remove(&file_priv->object_idr, handle);
spin_unlock(&file_priv->table_lock);
err_unref:
drm_gem_object_handle_put_unlocked(obj);
return ret;
}
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
/**
* drm_gem_handle_create - create a gem handle for an object
* @file_priv: drm file-private structure to register the handle for
* @obj: object to register
* @handlep: pionter to return the created handle to the caller
*
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
* Create a handle for this object. This adds a handle reference
* to the object, which includes a regular reference count. Callers
* will likely want to dereference the object afterwards.
*/
int drm_gem_handle_create(struct drm_file *file_priv,
struct drm_gem_object *obj,
u32 *handlep)
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
{
mutex_lock(&obj->dev->object_name_lock);
return drm_gem_handle_create_tail(file_priv, obj, handlep);
}
EXPORT_SYMBOL(drm_gem_handle_create);
/**
* drm_gem_free_mmap_offset - release a fake mmap offset for an object
* @obj: obj in question
*
* This routine frees fake offsets allocated by drm_gem_create_mmap_offset().
*
* Note that drm_gem_object_release() already calls this function, so drivers
* don't have to take care of releasing the mmap offset themselves when freeing
* the GEM object.
*/
void
drm_gem_free_mmap_offset(struct drm_gem_object *obj)
{
struct drm_device *dev = obj->dev;
drm_vma_offset_remove(dev->vma_offset_manager, &obj->vma_node);
}
EXPORT_SYMBOL(drm_gem_free_mmap_offset);
/**
* drm_gem_create_mmap_offset_size - create a fake mmap offset for an object
* @obj: obj in question
* @size: the virtual size
*
* GEM memory mapping works by handing back to userspace a fake mmap offset
* it can use in a subsequent mmap(2) call. The DRM core code then looks
* up the object based on the offset and sets up the various memory mapping
* structures.
*
* This routine allocates and attaches a fake offset for @obj, in cases where
* the virtual size differs from the physical size (ie. &drm_gem_object.size).
* Otherwise just use drm_gem_create_mmap_offset().
*
* This function is idempotent and handles an already allocated mmap offset
* transparently. Drivers do not need to check for this case.
*/
int
drm_gem_create_mmap_offset_size(struct drm_gem_object *obj, size_t size)
{
struct drm_device *dev = obj->dev;
return drm_vma_offset_add(dev->vma_offset_manager, &obj->vma_node,
size / PAGE_SIZE);
}
EXPORT_SYMBOL(drm_gem_create_mmap_offset_size);
/**
* drm_gem_create_mmap_offset - create a fake mmap offset for an object
* @obj: obj in question
*
* GEM memory mapping works by handing back to userspace a fake mmap offset
* it can use in a subsequent mmap(2) call. The DRM core code then looks
* up the object based on the offset and sets up the various memory mapping
* structures.
*
* This routine allocates and attaches a fake offset for @obj.
*
* Drivers can call drm_gem_free_mmap_offset() before freeing @obj to release
* the fake offset again.
*/
int drm_gem_create_mmap_offset(struct drm_gem_object *obj)
{
return drm_gem_create_mmap_offset_size(obj, obj->size);
}
EXPORT_SYMBOL(drm_gem_create_mmap_offset);
/**
* drm_gem_get_pages - helper to allocate backing pages for a GEM object
* from shmem
* @obj: obj in question
*
* This reads the page-array of the shmem-backing storage of the given gem
* object. An array of pages is returned. If a page is not allocated or
* swapped-out, this will allocate/swap-in the required pages. Note that the
* whole object is covered by the page-array and pinned in memory.
*
* Use drm_gem_put_pages() to release the array and unpin all pages.
*
* This uses the GFP-mask set on the shmem-mapping (see mapping_set_gfp_mask()).
* If you require other GFP-masks, you have to do those allocations yourself.
*
* Note that you are not allowed to change gfp-zones during runtime. That is,
* shmem_read_mapping_page_gfp() must be called with the same gfp_zone(gfp) as
* set during initialization. If you have special zone constraints, set them
* after drm_gem_init_object() via mapping_set_gfp_mask(). shmem-core takes care
* to keep pages in the required zone during swap-in.
*/
struct page **drm_gem_get_pages(struct drm_gem_object *obj)
{
struct address_space *mapping;
struct page *p, **pages;
int i, npages;
/* This is the shared memory object that backs the GEM resource */
mapping = obj->filp->f_mapping;
/* We already BUG_ON() for non-page-aligned sizes in
* drm_gem_object_init(), so we should never hit this unless
* driver author is doing something really wrong:
*/
WARN_ON((obj->size & (PAGE_SIZE - 1)) != 0);
npages = obj->size >> PAGE_SHIFT;
pages = kvmalloc_array(npages, sizeof(struct page *), GFP_KERNEL);
if (pages == NULL)
return ERR_PTR(-ENOMEM);
for (i = 0; i < npages; i++) {
p = shmem_read_mapping_page(mapping, i);
if (IS_ERR(p))
goto fail;
pages[i] = p;
/* Make sure shmem keeps __GFP_DMA32 allocated pages in the
* correct region during swapin. Note that this requires
* __GFP_DMA32 to be set in mapping_gfp_mask(inode->i_mapping)
* so shmem can relocate pages during swapin if required.
*/
BUG_ON(mapping_gfp_constraint(mapping, __GFP_DMA32) &&
(page_to_pfn(p) >= 0x00100000UL));
}
return pages;
fail:
while (i--)
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
put_page(pages[i]);
kvfree(pages);
return ERR_CAST(p);
}
EXPORT_SYMBOL(drm_gem_get_pages);
/**
* drm_gem_put_pages - helper to free backing pages for a GEM object
* @obj: obj in question
* @pages: pages to free
* @dirty: if true, pages will be marked as dirty
* @accessed: if true, the pages will be marked as accessed
*/
void drm_gem_put_pages(struct drm_gem_object *obj, struct page **pages,
bool dirty, bool accessed)
{
int i, npages;
/* We already BUG_ON() for non-page-aligned sizes in
* drm_gem_object_init(), so we should never hit this unless
* driver author is doing something really wrong:
*/
WARN_ON((obj->size & (PAGE_SIZE - 1)) != 0);
npages = obj->size >> PAGE_SHIFT;
for (i = 0; i < npages; i++) {
if (dirty)
set_page_dirty(pages[i]);
if (accessed)
mark_page_accessed(pages[i]);
/* Undo the reference we took when populating the table */
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
put_page(pages[i]);
}
kvfree(pages);
}
EXPORT_SYMBOL(drm_gem_put_pages);
/**
* drm_gem_object_lookup - look up a GEM object from it's handle
* @filp: DRM file private date
* @handle: userspace handle
*
* Returns:
*
* A reference to the object named by the handle if such exists on @filp, NULL
* otherwise.
*/
struct drm_gem_object *
drm_gem_object_lookup(struct drm_file *filp, u32 handle)
{
struct drm_gem_object *obj;
spin_lock(&filp->table_lock);
/* Check if we currently have a reference on the object */
obj = idr_find(&filp->object_idr, handle);
if (obj)
drm_gem_object_get(obj);
spin_unlock(&filp->table_lock);
return obj;
}
EXPORT_SYMBOL(drm_gem_object_lookup);
/**
* drm_gem_close_ioctl - implementation of the GEM_CLOSE ioctl
* @dev: drm_device
* @data: ioctl data
* @file_priv: drm file-private structure
*
* Releases the handle to an mm object.
*/
int
drm_gem_close_ioctl(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_gem_close *args = data;
int ret;
if (!drm_core_check_feature(dev, DRIVER_GEM))
return -ENODEV;
ret = drm_gem_handle_delete(file_priv, args->handle);
return ret;
}
/**
* drm_gem_flink_ioctl - implementation of the GEM_FLINK ioctl
* @dev: drm_device
* @data: ioctl data
* @file_priv: drm file-private structure
*
* Create a global name for an object, returning the name.
*
* Note that the name does not hold a reference; when the object
* is freed, the name goes away.
*/
int
drm_gem_flink_ioctl(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_gem_flink *args = data;
struct drm_gem_object *obj;
int ret;
if (!drm_core_check_feature(dev, DRIVER_GEM))
return -ENODEV;
obj = drm_gem_object_lookup(file_priv, args->handle);
if (obj == NULL)
return -ENOENT;
mutex_lock(&dev->object_name_lock);
drm/gem: fix up flink name create race This is the 2nd attempt, I've always been a bit dissatisified with the tricky nature of the first one: http://lists.freedesktop.org/archives/dri-devel/2012-July/025451.html The issue is that the flink ioctl can race with calling gem_close on the last gem handle. In that case we'll end up with a zero handle count, but an flink name (and it's corresponding reference). Which results in a neat space leak. In my first attempt I've solved this by rechecking the handle count. But fundamentally the issue is that ->handle_count isn't your usual refcount - it can be resurrected from 0 among other things. For those special beasts atomic_t often suggest way more ordering that it actually guarantees. To prevent being tricked by those hairy semantics take the easy way out and simply protect the handle with the existing dev->object_name_lock. With that change implemented it's dead easy to fix the flink vs. gem close reace: When we try to create the name we simply have to check whether there's still officially a gem handle around and if not refuse to create the flink name. Since the handle count decrement and flink name destruction is now also protected by that lock the reace is gone and we can't ever leak the flink reference again. Outside of the drm core only the exynos driver looks at the handle count, and tbh I have no idea why (it's just for debug dmesg output luckily). I've considered inlining the drm_gem_object_handle_free, but I plan to add more name-like things (like the exported dma_buf) to this scheme, so it's clearer to leave the handle freeing in its own function. This is exercised by the new gem_flink_race i-g-t testcase, which on my snb leaks gem objects at a rate of roughly 1k objects/s. v2: Fix up the error path handling in handle_create and make it more robust by simply calling object_handle_unreference. v3: Fix up the handle_unreference logic bug - atomic_dec_and_test retursn 1 for 0. Oops. v4: Squash in inlining of drm_gem_object_handle_reference as suggested by Dave Airlie and add a note that we now have a testcase. Cc: Dave Airlie <airlied@gmail.com> Cc: Inki Dae <inki.dae@samsung.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:37 +00:00
/* prevent races with concurrent gem_close. */
if (obj->handle_count == 0) {
ret = -ENOENT;
goto err;
}
if (!obj->name) {
ret = idr_alloc(&dev->object_name_idr, obj, 1, 0, GFP_KERNEL);
if (ret < 0)
goto err;
obj->name = ret;
}
args->name = (uint64_t) obj->name;
ret = 0;
err:
mutex_unlock(&dev->object_name_lock);
drm_gem_object_put_unlocked(obj);
return ret;
}
/**
* drm_gem_open - implementation of the GEM_OPEN ioctl
* @dev: drm_device
* @data: ioctl data
* @file_priv: drm file-private structure
*
* Open an object using the global name, returning a handle and the size.
*
* This handle (of course) holds a reference to the object, so the object
* will not go away until the handle is deleted.
*/
int
drm_gem_open_ioctl(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_gem_open *args = data;
struct drm_gem_object *obj;
int ret;
u32 handle;
if (!drm_core_check_feature(dev, DRIVER_GEM))
return -ENODEV;
mutex_lock(&dev->object_name_lock);
obj = idr_find(&dev->object_name_idr, (int) args->name);
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
if (obj) {
drm_gem_object_get(obj);
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
} else {
mutex_unlock(&dev->object_name_lock);
return -ENOENT;
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
}
drm/gem: completely close gem_open vs. gem_close races The gem flink name holds a reference onto the object itself, and this self-reference would prevent an flink'ed object from every being freed. To break that loop we remove the flink name when the last userspace handle disappears, i.e. when obj->handle_count reaches 0. Now in gem_open we drop the dev->object_name_lock between the flink name lookup and actually adding the handle. This means a concurrent gem_close of the last handle could result in the flink name getting reaped right inbetween, i.e. Thread 1 Thread 2 gem_open gem_close flink -> obj lookup handle_count drops to 0 remove flink name create_handle handle_count++ If someone now flinks this object again, we'll get a new flink name. We can close this race by removing the lock dropping and making the entire lookup+handle_create sequence atomic. Unfortunately to still be able to share the handle_create logic this requires a handle_create_tail function which drops the lock - we can't hold the object_name_lock while calling into a driver's ->gem_open callback. Note that for flink fixing this race isn't really important, since racing gem_open against gem_close is clearly a userspace bug. And no matter how the race ends, we won't leak any references. But with dma-buf where the userspace dma-buf fd itself is refcounted this is a valid sequence and hence we should fix it. Therefore this patch here is just a warm-up exercise (and for consistency between flink buffer sharing and dma-buf buffer sharing with self-imports). Also note that this extension of the critical section in gem_open protected by dev->object_name_lock only works because it's now a mutex: A spinlock would conflict with the potential memory allocation in idr_preload(). This is exercises by igt/gem_flink_race/flink_name. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:45 +00:00
/* drm_gem_handle_create_tail unlocks dev->object_name_lock. */
ret = drm_gem_handle_create_tail(file_priv, obj, &handle);
drm_gem_object_put_unlocked(obj);
if (ret)
return ret;
args->handle = handle;
args->size = obj->size;
return 0;
}
/**
* gem_gem_open - initalizes GEM file-private structures at devnode open time
* @dev: drm_device which is being opened by userspace
* @file_private: drm file-private structure to set up
*
* Called at device open time, sets up the structure for handling refcounting
* of mm objects.
*/
void
drm_gem_open(struct drm_device *dev, struct drm_file *file_private)
{
idr_init(&file_private->object_idr);
spin_lock_init(&file_private->table_lock);
}
/**
* drm_gem_release - release file-private GEM resources
* @dev: drm_device which is being closed by userspace
* @file_private: drm file-private structure to clean up
*
* Called at close time when the filp is going away.
*
* Releases any remaining references on objects by this filp.
*/
void
drm_gem_release(struct drm_device *dev, struct drm_file *file_private)
{
idr_for_each(&file_private->object_idr,
&drm_gem_object_release_handle, file_private);
idr_destroy(&file_private->object_idr);
}
/**
* drm_gem_object_release - release GEM buffer object resources
* @obj: GEM buffer object
*
* This releases any structures and resources used by @obj and is the invers of
* drm_gem_object_init().
*/
void
drm_gem_object_release(struct drm_gem_object *obj)
{
drm/prime: proper locking+refcounting for obj->dma_buf link The export dma-buf cache is semantically similar to an flink name. So semantically it makes sense to treat it the same and remove the name (i.e. the dma_buf pointer) and its references when the last gem handle disappears. Again we need to be careful, but double so: Not just could someone race and export with a gem close ioctl (so we need to recheck obj->handle_count again when assigning the new name), but multiple exports can also race against each another. This is prevented by holding the dev->object_name_lock across the entire section which touches obj->dma_buf. With the new scheme we also need to reinstate the obj->dma_buf link at import time (in case the only reference userspace has held in-between was through the dma-buf fd and not through any native gem handle). For simplicity we don't check whether it's a native object but unconditionally set up that link - with the new scheme of removing the obj->dma_buf reference when the last handle disappears we can do that. To make it clear that this is not just for exported buffers anymore als rename it from export_dma_buf to dma_buf. To make sure that now one can race a fd_to_handle or handle_to_fd with gem_close we use the same tricks as in flink of extending the dev->object_name_locking critical section. With this change we finally have a guaranteed 1:1 relationship (at least for native objects) between gem objects and dma-bufs, even accounting for races (which can happen since the dma-buf itself holds a reference while in-flight). This prevent igt/prime_self_import/export-vs-gem_close-race from Oopsing the kernel. There is still a leak though since the per-file priv dma-buf/handle cache handling is racy. That will be fixed in a later patch. v2: Remove the bogus dma_buf_put from the export_and_register_object failure path if we've raced with the handle count dropping to 0. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-08-14 22:02:46 +00:00
WARN_ON(obj->dma_buf);
if (obj->filp)
fput(obj->filp);
drm_gem_free_mmap_offset(obj);
}
EXPORT_SYMBOL(drm_gem_object_release);
/**
* drm_gem_object_free - free a GEM object
* @kref: kref of the object to free
*
* Called after the last reference to the object has been lost.
* Must be called holding &drm_device.struct_mutex.
*
* Frees the object
*/
void
drm_gem_object_free(struct kref *kref)
{
struct drm_gem_object *obj =
container_of(kref, struct drm_gem_object, refcount);
struct drm_device *dev = obj->dev;
if (dev->driver->gem_free_object_unlocked) {
dev->driver->gem_free_object_unlocked(obj);
} else if (dev->driver->gem_free_object) {
WARN_ON(!mutex_is_locked(&dev->struct_mutex));
dev->driver->gem_free_object(obj);
}
}
EXPORT_SYMBOL(drm_gem_object_free);
/**
* drm_gem_object_put_unlocked - drop a GEM buffer object reference
* @obj: GEM buffer object
*
* This releases a reference to @obj. Callers must not hold the
* &drm_device.struct_mutex lock when calling this function.
*
* See also __drm_gem_object_put().
*/
void
drm_gem_object_put_unlocked(struct drm_gem_object *obj)
{
struct drm_device *dev;
if (!obj)
return;
dev = obj->dev;
if (dev->driver->gem_free_object_unlocked) {
kref_put(&obj->refcount, drm_gem_object_free);
} else {
might_lock(&dev->struct_mutex);
if (kref_put_mutex(&obj->refcount, drm_gem_object_free,
&dev->struct_mutex))
mutex_unlock(&dev->struct_mutex);
}
}
EXPORT_SYMBOL(drm_gem_object_put_unlocked);
/**
* drm_gem_object_put - release a GEM buffer object reference
* @obj: GEM buffer object
*
* This releases a reference to @obj. Callers must hold the
* &drm_device.struct_mutex lock when calling this function, even when the
* driver doesn't use &drm_device.struct_mutex for anything.
*
* For drivers not encumbered with legacy locking use
* drm_gem_object_put_unlocked() instead.
*/
void
drm_gem_object_put(struct drm_gem_object *obj)
{
if (obj) {
WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex));
kref_put(&obj->refcount, drm_gem_object_free);
}
}
EXPORT_SYMBOL(drm_gem_object_put);
/**
* drm_gem_vm_open - vma->ops->open implementation for GEM
* @vma: VM area structure
*
* This function implements the #vm_operations_struct open() callback for GEM
* drivers. This must be used together with drm_gem_vm_close().
*/
void drm_gem_vm_open(struct vm_area_struct *vma)
{
struct drm_gem_object *obj = vma->vm_private_data;
drm_gem_object_get(obj);
}
EXPORT_SYMBOL(drm_gem_vm_open);
/**
* drm_gem_vm_close - vma->ops->close implementation for GEM
* @vma: VM area structure
*
* This function implements the #vm_operations_struct close() callback for GEM
* drivers. This must be used together with drm_gem_vm_open().
*/
void drm_gem_vm_close(struct vm_area_struct *vma)
{
struct drm_gem_object *obj = vma->vm_private_data;
drm_gem_object_put_unlocked(obj);
}
EXPORT_SYMBOL(drm_gem_vm_close);
/**
* drm_gem_mmap_obj - memory map a GEM object
* @obj: the GEM object to map
* @obj_size: the object size to be mapped, in bytes
* @vma: VMA for the area to be mapped
*
* Set up the VMA to prepare mapping of the GEM object using the gem_vm_ops
* provided by the driver. Depending on their requirements, drivers can either
* provide a fault handler in their gem_vm_ops (in which case any accesses to
* the object will be trapped, to perform migration, GTT binding, surface
* register allocation, or performance monitoring), or mmap the buffer memory
* synchronously after calling drm_gem_mmap_obj.
*
* This function is mainly intended to implement the DMABUF mmap operation, when
* the GEM object is not looked up based on its fake offset. To implement the
* DRM mmap operation, drivers should use the drm_gem_mmap() function.
*
* drm_gem_mmap_obj() assumes the user is granted access to the buffer while
* drm_gem_mmap() prevents unprivileged users from mapping random objects. So
* callers must verify access restrictions before calling this helper.
*
* Return 0 or success or -EINVAL if the object size is smaller than the VMA
* size, or if no gem_vm_ops are provided.
*/
int drm_gem_mmap_obj(struct drm_gem_object *obj, unsigned long obj_size,
struct vm_area_struct *vma)
{
struct drm_device *dev = obj->dev;
/* Check for valid size. */
if (obj_size < vma->vm_end - vma->vm_start)
return -EINVAL;
if (!dev->driver->gem_vm_ops)
return -EINVAL;
vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
vma->vm_ops = dev->driver->gem_vm_ops;
vma->vm_private_data = obj;
vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
/* Take a ref for this mapping of the object, so that the fault
* handler can dereference the mmap offset's pointer to the object.
* This reference is cleaned up by the corresponding vm_close
* (which should happen whether the vma was created by this call, or
* by a vm_open due to mremap or partial unmap or whatever).
*/
drm_gem_object_get(obj);
return 0;
}
EXPORT_SYMBOL(drm_gem_mmap_obj);
/**
* drm_gem_mmap - memory map routine for GEM objects
* @filp: DRM file pointer
* @vma: VMA for the area to be mapped
*
* If a driver supports GEM object mapping, mmap calls on the DRM file
* descriptor will end up here.
*
* Look up the GEM object based on the offset passed in (vma->vm_pgoff will
* contain the fake offset we created when the GTT map ioctl was called on
* the object) and map it with a call to drm_gem_mmap_obj().
*
* If the caller is not granted access to the buffer object, the mmap will fail
* with EACCES. Please see the vma manager for more information.
*/
int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
{
struct drm_file *priv = filp->private_data;
struct drm_device *dev = priv->minor->dev;
struct drm_gem_object *obj = NULL;
struct drm_vma_offset_node *node;
int ret;
if (drm_device_is_unplugged(dev))
return -ENODEV;
drm_vma_offset_lock_lookup(dev->vma_offset_manager);
node = drm_vma_offset_exact_lookup_locked(dev->vma_offset_manager,
vma->vm_pgoff,
vma_pages(vma));
if (likely(node)) {
obj = container_of(node, struct drm_gem_object, vma_node);
/*
* When the object is being freed, after it hits 0-refcnt it
* proceeds to tear down the object. In the process it will
* attempt to remove the VMA offset and so acquire this
* mgr->vm_lock. Therefore if we find an object with a 0-refcnt
* that matches our range, we know it is in the process of being
* destroyed and will be freed as soon as we release the lock -
* so we have to check for the 0-refcnted object and treat it as
* invalid.
*/
if (!kref_get_unless_zero(&obj->refcount))
obj = NULL;
}
drm_vma_offset_unlock_lookup(dev->vma_offset_manager);
if (!obj)
return -EINVAL;
if (!drm_vma_node_is_allowed(node, priv)) {
drm_gem_object_put_unlocked(obj);
return -EACCES;
}
ret = drm_gem_mmap_obj(obj, drm_vma_node_size(node) << PAGE_SHIFT,
vma);
drm_gem_object_put_unlocked(obj);
return ret;
}
EXPORT_SYMBOL(drm_gem_mmap);