drm/msm/dpu: Correct dpu destroy and disable order
authorJayant Shekhar <jshekhar@codeaurora.org>
Fri, 2 Nov 2018 12:49:17 +0000 (18:19 +0530)
committerRob Clark <robdclark@gmail.com>
Tue, 11 Dec 2018 18:07:08 +0000 (13:07 -0500)
In case of msm drm bind failure, dpu_mdss_destroy is triggered.
In this function, resources are freed and pm runtime disable is
called, which triggers dpu_mdss_disable. Now in dpu_mdss_disable,
driver tries to access a memory which is already freed. This
results in kernel panic. Fix this by ensuring proper sequence
of dpu destroy and disable calls.

Changes in v2:
   - Removed double spacings [Jeykumar]

Tested-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Jayant Shekhar <jshekhar@codeaurora.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
drivers/gpu/drm/msm/disp/dpu1/dpu_mdss.c

index 2235ef8129f4b02e85e0779ef52429fbf3f29836..38576f8b90b66d220514304e2d814cceac2d7f55 100644 (file)
@@ -156,18 +156,15 @@ static void dpu_mdss_destroy(struct drm_device *dev)
        struct dpu_mdss *dpu_mdss = to_dpu_mdss(priv->mdss);
        struct dss_module_power *mp = &dpu_mdss->mp;
 
+       pm_runtime_disable(dev->dev);
        _dpu_mdss_irq_domain_fini(dpu_mdss);
-
        free_irq(platform_get_irq(pdev, 0), dpu_mdss);
-
        msm_dss_put_clk(mp->clk_config, mp->num_clk);
        devm_kfree(&pdev->dev, mp->clk_config);
 
        if (dpu_mdss->mmio)
                devm_iounmap(&pdev->dev, dpu_mdss->mmio);
        dpu_mdss->mmio = NULL;
-
-       pm_runtime_disable(dev->dev);
        priv->mdss = NULL;
 }