Hello Tero Kristo, This is a semi-automatic email about new static checker warnings. The patch ad18cc9d0f91: "crypto: omap-aes - Add support for GCM mode" from May 24, 2017, leads to the following Smatch complaint: drivers/crypto/omap-aes.c:1262 omap_aes_probe() error: we previously assumed 'dd->pdata->aead_algs_info' could be null (see line 1237) drivers/crypto/omap-aes.c 1236 1237 if (dd->pdata->aead_algs_info && ^^^^^^^^^^^^^^^^^^^^^^^^^ Check for NULL. 1238 !dd->pdata->aead_algs_info->registered) { 1239 for (i = 0; i < dd->pdata->aead_algs_info->size; i++) { 1240 aalg = &dd->pdata->aead_algs_info->algs_list[i]; 1241 algp = &aalg->base; 1242 1243 pr_debug("reg alg: %s\n", algp->cra_name); 1244 INIT_LIST_HEAD(&algp->cra_list); 1245 1246 err = crypto_register_aead(aalg); 1247 if (err) 1248 goto err_aead_algs; 1249 1250 dd->pdata->aead_algs_info->registered++; 1251 } 1252 } 1253 1254 err = sysfs_create_group(&dev->kobj, &omap_aes_attr_group); 1255 if (err) { 1256 dev_err(dev, "could not create sysfs device attrs\n"); 1257 goto err_aead_algs; ^^^^^^^^^^^^^^^^^^ Goto. It was actually this goto added in Feb which introduced the warning, but my scripts got confused. 1258 } 1259 1260 return 0; 1261 err_aead_algs: 1262 for (i = dd->pdata->aead_algs_info->registered - 1; i >= 0; i--) { ^^^^^^^^^^^^^^^^^^^^^^^^^ Unchecked dereference. 1263 aalg = &dd->pdata->aead_algs_info->algs_list[i]; 1264 crypto_unregister_aead(aalg); regards, dan carpenter