Hello Linus Walleij,

This is a semi-automatic email about new static checker warnings.

The patch 5fc537bfd000: "drm/mcde: Add new driver for ST-Ericsson MCDE" from May 24, 2019, leads to the following Smatch complaint:

drivers/gpu/drm/mcde/mcde_dsi.c:908 mcde_dsi_bind() error: we previously assumed 'd->mdsi' could be null (see line 879)

drivers/gpu/drm/mcde/mcde_dsi.c 878 /* If the display attached before binding, set this up */ 879 if (d->mdsi) ^^^^^^^ Check for NULL

880 d->mcde->mdsi = d->mdsi; 881 882 /* Obtain the clocks */ 883 d->hs_clk = devm_clk_get(dev, "hs"); 884 if (IS_ERR(d->hs_clk)) { 885 dev_err(dev, "unable to get HS clock\n"); 886 return PTR_ERR(d->hs_clk); 887 } 888 889 d->lp_clk = devm_clk_get(dev, "lp"); 890 if (IS_ERR(d->lp_clk)) { 891 dev_err(dev, "unable to get LP clock\n"); 892 return PTR_ERR(d->lp_clk); 893 } 894 895 /* Assert RESET through the PRCMU, active low */ 896 /* FIXME: which DSI block? */ 897 regmap_update_bits(d->prcmu, PRCM_DSI_SW_RESET, 898 PRCM_DSI_SW_RESET_DSI0_SW_RESETN, 0); 899 900 usleep_range(100, 200); 901 902 /* De-assert RESET again */ 903 regmap_update_bits(d->prcmu, PRCM_DSI_SW_RESET, 904 PRCM_DSI_SW_RESET_DSI0_SW_RESETN, 905 PRCM_DSI_SW_RESET_DSI0_SW_RESETN); 906 907 /* Start up the hardware */ 908 mcde_dsi_start(d); ^ d->mdsi is dereferenced without checking for NULL inside here.

909 910 /* Look for a panel as a child to this node */

regards, dan carpenter