Skip to contentSkip to conversation

drm/vkms: Fix some memory related bug

PATCH In review Yuan Can 2 participants 5 messages Nov 1, 2022 – Nov 29, 2022

Subsystems (3)
lore
Share & tools
lore

Yuan Can

author
#0 lore raw

cover drm/vkms: Fix some memory related bug

This series contains two memory related bugfixes about vkms driver.

Yuan Can (2):
drm/vkms: Fix memory leak in vkms_init()
drm/vkms: Fix null-ptr-deref in vkms_release()

 drivers/gpu/drm/vkms/vkms_drv.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH 1/2 drm/vkms: Fix memory leak in vkms_init()

A memory leak was reported after the vkms module install failed.

unreferenced object 0xffff88810bc28520 (size 16):
comm "modprobe", pid 9662, jiffies 4298009455 (age 42.590s) hex dump (first 16 bytes):

    01 01 00 64 81 88 ff ff 00 00 dc 0a 81 88 ff ff  ...d............
  backtrace:
    [<00000000e7561ff8>] kmalloc_trace+0x27/0x60
    [<000000000b1954a0>] 0xffffffffc45200a9
    [<00000000abbf1da0>] do_one_initcall+0xd0/0x4f0
    [<000000001505ee87>] do_init_module+0x1a4/0x680
    [<00000000958079ad>] load_module+0x6249/0x7110
    [<00000000117e4696>] __do_sys_finit_module+0x140/0x200
    [<00000000f74b12d2>] do_syscall_64+0x35/0x80
    [<000000008fc6fcde>] entry_SYSCALL_64_after_hwframe+0x46/0xb0

The reason is that the vkms_init() returns without checking the return value of vkms_create(), and if the vkms_create() failed, the config allocated at the beginning of vkms_init() is leaked.

vkms_init()
config = kmalloc(...) # config allocated
...
return vkms_create() # vkms_create failed and config is leaked

Fix this problem by checking return value of vkms_create() and free the config if error happened.

Fixes 2df7af93fdad ("drm/vkms: Add vkms_config type")
---
 drivers/gpu/drm/vkms/vkms_drv.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/vkms/vkms_drv.c +6 -1
index 0ffe5f0e33f7..dfe983eaa07f 100644
@@ -218,6 +218,7 @@ static int vkms_create(struct vkms_config *config)
 
 static int __init vkms_init(void)
 {
+	int ret;
 	struct vkms_config *config;
 
 	config = kmalloc(sizeof(*config), GFP_KERNEL);
@@ -230,7 +231,11 @@ static int __init vkms_init(void)
 	config->writeback = enable_writeback;
 	config->overlay = enable_overlay;
 
-	return vkms_create(config);
+	ret = vkms_create(config);
+	if (ret)
+		kfree(config);
+
+	return ret;
 }
 
 static void vkms_destroy(struct vkms_config *config)
Signature · 1 line
2.17.1
Reviewed-by on v1 1/2
#2 lore raw

Re: [PATCH 1/2] drm/vkms: Fix memory leak in vkms_init()

Quoted 47 lines · Yuan Can

A memory leak was reported after the vkms module install failed.

unreferenced object 0xffff88810bc28520 (size 16):
comm "modprobe", pid 9662, jiffies 4298009455 (age 42.590s) hex dump (first 16 bytes):

    01 01 00 64 81 88 ff ff 00 00 dc 0a 81 88 ff ff  ...d............
  backtrace:
    [<00000000e7561ff8>] kmalloc_trace+0x27/0x60
    [<000000000b1954a0>] 0xffffffffc45200a9
    [<00000000abbf1da0>] do_one_initcall+0xd0/0x4f0
    [<000000001505ee87>] do_init_module+0x1a4/0x680
    [<00000000958079ad>] load_module+0x6249/0x7110
    [<00000000117e4696>] __do_sys_finit_module+0x140/0x200
    [<00000000f74b12d2>] do_syscall_64+0x35/0x80
    [<000000008fc6fcde>] entry_SYSCALL_64_after_hwframe+0x46/0xb0

The reason is that the vkms_init() returns without checking the return value of vkms_create(), and if the vkms_create() failed, the config allocated at the beginning of vkms_init() is leaked.

vkms_init()
config = kmalloc(...) # config allocated
...
return vkms_create() # vkms_create failed and config is leaked

Fix this problem by checking return value of vkms_create() and free the config if error happened.

Fixes 2df7af93fdad ("drm/vkms: Add vkms_config type")
---
 drivers/gpu/drm/vkms/vkms_drv.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/vkms/vkms_drv.c +1 -0
index 0ffe5f0e33f7..dfe983eaa07f 100644
@@ -218,6 +218,7 @@ static int vkms_create(struct vkms_config *config)
 
 static int __init vkms_init(void)
 {
+	int ret;
 	struct vkms_config *config;
 
 	config = kmalloc(sizeof(*config), GFP_KERNEL);

drivers/gpu/drm/vkms/vkms_drv.c static int __init vkms_init(void) · line 238

Quoted 9 lines
drivers/gpu/drm/vkms/vkms_drv.c +5 -1
@@ -230,7 +231,11 @@ static int __init vkms_init(void)
 	config->writeback = enable_writeback;
 	config->overlay = enable_overlay;
 
-	return vkms_create(config);
+	ret = vkms_create(config);
+	if (ret)
+		kfree(config);
+
+	return ret;

Good catch. Thanks,

Quoted 6 lines
drivers/gpu/drm/vkms/vkms_drv.c +0 -0
 }
 
 static void vkms_destroy(struct vkms_config *config)
Signature · 1 line
2.17.1
author
#3 lore raw

PATCH 2/2 drm/vkms: Fix null-ptr-deref in vkms_release()

In reply to Yuan Can

A null-ptr-deref is triggered when it tries to destroy the workqueue in vkms->output.composer_workq in vkms_release().

 KASAN: null-ptr-deref in range [0x0000000000000118-0x000000000000011f]
 CPU: 5 PID: 17193 Comm: modprobe Not tainted 6.0.0-11331-gd465bff130bf #24
 RIP: 0010:destroy_workqueue+0x2f/0x710
 ...
 Call Trace:
  <TASK>
  ? vkms_config_debugfs_init+0x50/0x50 [vkms]
  __devm_drm_dev_alloc+0x15a/0x1c0 [drm]
  vkms_init+0x245/0x1000 [vkms]
  do_one_initcall+0xd0/0x4f0
  do_init_module+0x1a4/0x680
  load_module+0x6249/0x7110
  __do_sys_finit_module+0x140/0x200
  do_syscall_64+0x35/0x80
  entry_SYSCALL_64_after_hwframe+0x46/0xb0

The reason is that an OOM happened which triggers the destroy of the workqueue, however, the workqueue is alloced in the later process, thus a null-ptr-deref happened. A simple call graph is shown as below:

vkms_init()
vkms_create()

    devm_drm_dev_alloc()
      __devm_drm_dev_alloc()
        devm_drm_dev_init()
          devm_add_action_or_reset()
            devm_add_action() # an error happened
            devm_drm_dev_init_release()
              drm_dev_put()
                kref_put()
                  drm_dev_release()
                    vkms_release()
                      destroy_workqueue() # null-ptr-deref happened
    vkms_modeset_init()
      vkms_output_init()
        vkms_crtc_init() # where the workqueue get allocated

Fix this by checking if composer_workq is NULL before passing it to the destroy_workqueue() in vkms_release().

Fixes 6c234fe37c57 ("drm/vkms: Implement CRC debugfs API")
---
 drivers/gpu/drm/vkms/vkms_drv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/vkms/vkms_drv.c +2 -1
index dfe983eaa07f..f716c5796f5f 100644
@@ -57,7 +57,8 @@ static void vkms_release(struct drm_device *dev)
 {
 	struct vkms_device *vkms = drm_device_to_vkms_device(dev);
 
-	destroy_workqueue(vkms->output.composer_workq);
+	if (vkms->output.composer_workq)
+		destroy_workqueue(vkms->output.composer_workq);
 }
 
 static void vkms_atomic_commit_tail(struct drm_atomic_state *old_state)
Signature · 1 line
2.17.1
Reviewed-by on v1 2/2
#4 lore raw

Re: [PATCH 2/2] drm/vkms: Fix null-ptr-deref in vkms_release()

Quoted 54 lines · Yuan Can

A null-ptr-deref is triggered when it tries to destroy the workqueue in vkms->output.composer_workq in vkms_release().

 KASAN: null-ptr-deref in range [0x0000000000000118-0x000000000000011f]
 CPU: 5 PID: 17193 Comm: modprobe Not tainted 6.0.0-11331-gd465bff130bf #24
 RIP: 0010:destroy_workqueue+0x2f/0x710
 ...
 Call Trace:
  <TASK>
  ? vkms_config_debugfs_init+0x50/0x50 [vkms]
  __devm_drm_dev_alloc+0x15a/0x1c0 [drm]
  vkms_init+0x245/0x1000 [vkms]
  do_one_initcall+0xd0/0x4f0
  do_init_module+0x1a4/0x680
  load_module+0x6249/0x7110
  __do_sys_finit_module+0x140/0x200
  do_syscall_64+0x35/0x80
  entry_SYSCALL_64_after_hwframe+0x46/0xb0

The reason is that an OOM happened which triggers the destroy of the workqueue, however, the workqueue is alloced in the later process, thus a null-ptr-deref happened. A simple call graph is shown as below:

vkms_init()
vkms_create()

    devm_drm_dev_alloc()
      __devm_drm_dev_alloc()
        devm_drm_dev_init()
          devm_add_action_or_reset()
            devm_add_action() # an error happened
            devm_drm_dev_init_release()
              drm_dev_put()
                kref_put()
                  drm_dev_release()
                    vkms_release()
                      destroy_workqueue() # null-ptr-deref happened
    vkms_modeset_init()
      vkms_output_init()
        vkms_crtc_init() # where the workqueue get allocated

Fix this by checking if composer_workq is NULL before passing it to the destroy_workqueue() in vkms_release().

Fixes 6c234fe37c57 ("drm/vkms: Implement CRC debugfs API")
---
 drivers/gpu/drm/vkms/vkms_drv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/vkms/vkms_drv.c +0 -0
index dfe983eaa07f..f716c5796f5f 100644

drivers/gpu/drm/vkms/vkms_drv.c static void vkms_release(struct drm_device *dev) · line 61

Quoted 6 lines
drivers/gpu/drm/vkms/vkms_drv.c +2 -1
@@ -57,7 +57,8 @@ static void vkms_release(struct drm_device *dev)
 {
 	struct vkms_device *vkms = drm_device_to_vkms_device(dev);
 
-	destroy_workqueue(vkms->output.composer_workq);
+	if (vkms->output.composer_workq)
+		destroy_workqueue(vkms->output.composer_workq);
Quoted 6 lines
drivers/gpu/drm/vkms/vkms_drv.c +0 -0
 }
 
 static void vkms_atomic_commit_tail(struct drm_atomic_state *old_state)
Signature · 1 line
2.17.1