drm/i915: Copy the staged connector config to the legacy atomic state
Hello Ander Conselvan de Oliveira,
This is a semi-automatic email about new static checker warnings.
The patch 944b0c765757: "drm/i915: Copy the staged connector config to the legacy atomic state" from Mar 20, 2015, leads to the following Smatch complaint:
drivers/gpu/drm/i915/intel_display.c:9118 intel_release_load_detect_pipe()
error: we previously assumed 'state' could be null (see line 9082)
drivers/gpu/drm/i915/intel_display.c
9081 state = drm_atomic_state_alloc(dev);
9082 if (!state)
9083 goto fail;
^^^^^^^^^Patch changes a return into a goto.
9084
9085 state->acquire_ctx = ctx;
9086
9087 connector_state = drm_atomic_get_connector_state(state, connector);
9088 if (IS_ERR(connector_state))
9089 goto fail;
9090
9091 to_intel_connector(connector)->new_encoder = NULL;
9092 intel_encoder->new_crtc = NULL;
9093 intel_crtc->new_enabled = false;
9094 intel_crtc->new_config = NULL;
9095
9096 connector_state->best_encoder = NULL;
9097 connector_state->crtc = NULL;
9098
9099 intel_set_mode(crtc, NULL, 0, 0, NULL, state);
9100
9101 drm_atomic_state_free(state);
9102
9103 if (old->release_fb) {
9104 drm_framebuffer_unregister_private(old->release_fb);
9105 drm_framebuffer_unreference(old->release_fb);
9106 }
9107
9108 return;
9109 }
9110
9111 /* Switch crtc and encoder back off if necessary */
9112 if (old->dpms_mode != DRM_MODE_DPMS_ON)
9113 connector->funcs->dpms(connector, old->dpms_mode);
9114
9115 return;
9116 fail:
9117 DRM_DEBUG_KMS("Couldn't release load detect pipe.\n");
9118 drm_atomic_state_free(state);
^^^^^Dereferenced inside function call.
9119 }
9120
regards,
dan carpenter
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 52 lines · Dan Carpenter
Hello Ander Conselvan de Oliveira,
This is a semi-automatic email about new static checker warnings.
The patch
944b0c765757: "drm/i915: Copy the staged connector config to the legacy atomic state" from Mar 20, 2015, leads to the following Smatch complaint:
drivers/gpu/drm/i915/intel_display.c:9118intel_release_load_detect_pipe()
error: we previously assumed 'state' could be null (see line 9082)
drivers/gpu/drm/i915/intel_display.c
9081 state =drm_atomic_state_alloc(dev);
9082 if (!state)9083 goto fail; ^^^^^^^^^Patch changes a return into a goto.
9084
9085state->acquire_ctx= ctx;
9086
9087connector_state=drm_atomic_get_connector_state(state, connector);
9088 if (IS_ERR(connector_state))
9089 goto fail;
9090
9091 to_intel_connector(connector)->new_encoder = NULL;
9092intel_encoder->new_crtc= NULL;
9093intel_crtc->new_enabled= false;
9094intel_crtc->new_config= NULL;
9095
9096connector_state->best_encoder= NULL;
9097connector_state->crtc= NULL;
9098
9099intel_set_mode(crtc, NULL, 0, 0, NULL, state);
9100
9101drm_atomic_state_free(state);
9102
9103 if (old->release_fb) {
9104drm_framebuffer_unregister_private(old->release_fb);
9105drm_framebuffer_unreference(old->release_fb);
9106 }
9107
9108 return;
9109 }
9110
9111 /* Switch crtc and encoder back off if necessary */
9112 if (old->dpms_mode!=DRM_MODE_DPMS_ON)
9113connector->funcs->dpms(connector, old->dpms_mode);
9114
9115 return;
9116 fail:
9117DRM_DEBUG_KMS("Couldn't release load detect pipe.\n");
Quoted 3 lines
9118 drm_atomic_state_free(state); ^^^^^Dereferenced inside function call.
That's odd, this is checking the "fail" label inside another function. The change above was in intel_get_load_detect_pipe(), and this is in intel_release_load_detect_pipe().
Anyway, the fail path in intel_get_load_detect_pipe() does check for state being NULL before using it.
Cheers,
Ander
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 52 lines · Dan Carpenter
Hello Ander Conselvan de Oliveira,
This is a semi-automatic email about new static checker warnings.
The patch
944b0c765757: "drm/i915: Copy the staged connector config to the legacy atomic state" from Mar 20, 2015, leads to the following Smatch complaint:
drivers/gpu/drm/i915/intel_display.c:9118intel_release_load_detect_pipe()
error: we previously assumed 'state' could be null (see line 9082)
drivers/gpu/drm/i915/intel_display.c
9081 state =drm_atomic_state_alloc(dev);
9082 if (!state)9083 goto fail; ^^^^^^^^^Patch changes a return into a goto.
9084
9085state->acquire_ctx= ctx;
9086
9087connector_state=drm_atomic_get_connector_state(state, connector);
9088 if (IS_ERR(connector_state))
9089 goto fail;
9090
9091 to_intel_connector(connector)->new_encoder = NULL;
9092intel_encoder->new_crtc= NULL;
9093intel_crtc->new_enabled= false;
9094intel_crtc->new_config= NULL;
9095
9096connector_state->best_encoder= NULL;
9097connector_state->crtc= NULL;
9098
9099intel_set_mode(crtc, NULL, 0, 0, NULL, state);
9100
9101drm_atomic_state_free(state);
9102
9103 if (old->release_fb) {
9104drm_framebuffer_unregister_private(old->release_fb);
9105drm_framebuffer_unreference(old->release_fb);
9106 }
9107
9108 return;
9109 }
9110
9111 /* Switch crtc and encoder back off if necessary */
9112 if (old->dpms_mode!=DRM_MODE_DPMS_ON)
9113connector->funcs->dpms(connector, old->dpms_mode);
9114
9115 return;
9116 fail:
9117DRM_DEBUG_KMS("Couldn't release load detect pipe.\n");
Quoted 6 lines
9118 drm_atomic_state_free(state); ^^^^^Dereferenced inside function call.
9119 }
9120
Wouldn't it make more sense to make drm_atomic_state_free() follow the same semantics as *free() functions typically do (no operation performed)?
Kind regards, David Weinehall
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 57 lines · David Weinehall
Quoted 56 lines · Dan Carpenter
Hello Ander Conselvan de Oliveira,
This is a semi-automatic email about new static checker warnings.
The patch
944b0c765757: "drm/i915: Copy the staged connector config to the legacy atomic state" from Mar 20, 2015, leads to the following Smatch complaint:
drivers/gpu/drm/i915/intel_display.c:9118intel_release_load_detect_pipe()
error: we previously assumed 'state' could be null (see line 9082)
drivers/gpu/drm/i915/intel_display.c
9081 state =drm_atomic_state_alloc(dev);
9082 if (!state)9083 goto fail; ^^^^^^^^^Patch changes a return into a goto.
9084
9085state->acquire_ctx= ctx;
9086
9087connector_state=drm_atomic_get_connector_state(state, connector);
9088 if (IS_ERR(connector_state))
9089 goto fail;
9090
9091 to_intel_connector(connector)->new_encoder = NULL;
9092intel_encoder->new_crtc= NULL;
9093intel_crtc->new_enabled= false;
9094intel_crtc->new_config= NULL;
9095
9096connector_state->best_encoder= NULL;
9097connector_state->crtc= NULL;
9098
9099intel_set_mode(crtc, NULL, 0, 0, NULL, state);
9100
9101drm_atomic_state_free(state);
9102
9103 if (old->release_fb) {
9104drm_framebuffer_unregister_private(old->release_fb);
9105drm_framebuffer_unreference(old->release_fb);
9106 }
9107
9108 return;
9109 }
9110
9111 /* Switch crtc and encoder back off if necessary */
9112 if (old->dpms_mode!=DRM_MODE_DPMS_ON)
9113connector->funcs->dpms(connector, old->dpms_mode);
9114
9115 return;
9116 fail:
9117DRM_DEBUG_KMS("Couldn't release load detect pipe.\n");9118 drm_atomic_state_free(state); ^^^^^Dereferenced inside function call.
Quoted 6 lines
Quoted 2 lines
9119 }
9120Wouldn't it make more sense to make
drm_atomic_state_free()follow the same semantics as*free()functions typically do (no operation performed)?
It would. I already signed up to send that patch, even. Just didn't get around to yet.
Ander
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
This is a typical "one err bug" which is a kind of bug where there is just one err: label which does all the error handling. One err bugs are one of the most common class of bugs in the kernel. "Oh, we didn't initialize that struct member" or "That pointer is NULL."
In networking they have resisted patches to add sanity checks to free_netdev() so that people can't do one err style error handling. You always have to know if the dev is allocated or not.
Having one error label makes sense if the label is "unlock:" but labels like fail:, err:, out:, or bail are not very good label names.
The disadvantage of having multiple returns is that it means you can't have a debug statement, but debug statements are mostly a waste of RAM and we can use ftrace for that.
regards,
dan carpenter
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx