Hello Dmitry Baryshkov,

The patch f25f656608e3: "drm/msm/dpu: merge struct dpu_irq into struct dpu_hw_intr" from Jun 18, 2021, leads to the following Smatch static checker warnings:

drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c:569 dpu_core_irq_preinstall() error: potential null dereference 'dpu_kms->hw_intr->irq_cb_tbl'. (kcalloc returns null)

drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c:570 dpu_core_irq_preinstall() error: potential null dereference 'dpu_kms->hw_intr->irq_counts'. (kcalloc returns null)

drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c 554 void dpu_core_irq_preinstall(struct dpu_kms *dpu_kms) 555 { 556 int i; 557 558 pm_runtime_get_sync(&dpu_kms->pdev->dev); 559 dpu_clear_irqs(dpu_kms); 560 dpu_disable_all_irqs(dpu_kms); 561 pm_runtime_put_sync(&dpu_kms->pdev->dev); 562 563 /* Create irq callbacks for all possible irq_idx */ 564 dpu_kms->hw_intr->irq_cb_tbl = kcalloc(dpu_kms->hw_intr->total_irqs, 565 sizeof(struct list_head), GFP_KERNEL); 566 dpu_kms->hw_intr->irq_counts = kcalloc(dpu_kms->hw_intr->total_irqs, 567 sizeof(atomic_t), GFP_KERNEL);

No checks for these kcallocs... It's not actually new code but shuffling them around makes them show up as new in my tests.

568 for (i = 0; i < dpu_kms->hw_intr->total_irqs; i++) { --> 569 INIT_LIST_HEAD(&dpu_kms->hw_intr->irq_cb_tbl[i]); 570 atomic_set(&dpu_kms->hw_intr->irq_counts[i], 0); 571 } 572 }

regards, dan carpenter