Hi Matti, I love your patch! Yet something to improve: [auto build test ERROR on 6efb943b8616ec53a5e444193dccf1af9ad627b5] url: https://github.com/0day-ci/linux/commits/Matti-Vaittinen/Extend-regulator... base: 6efb943b8616ec53a5e444193dccf1af9ad627b5 config: x86_64-randconfig-a015-20210510 (attached as .config) compiler: clang version 13.0.0 (https://github.com/llvm/llvm-project 492173d42b32cb91d5d0d72d5ed84fcab80d059a) reproduce (this is a W=1 build): wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross # install x86_64 cross compiling tool for clang build # apt-get install binutils-x86-64-linux-gnu # https://github.com/0day-ci/linux/commit/904edb46fa37ac86bc1e7a1629141e037... git remote add linux-review https://github.com/0day-ci/linux git fetch --no-tags linux-review Matti-Vaittinen/Extend-regulator-notification-support/20210510-203125 git checkout 904edb46fa37ac86bc1e7a1629141e037f45ebed # save the attached .config to linux build tree COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 ARCH=x86_64 If you fix the issue, kindly add following tag as appropriate Reported-by: kernel test robot <lkp(a)intel.com&gt; All errors (new ones prefixed by >>): pr_dbg("Sending regulator notification EVT 0x%lx\r\n", ^ 1 error generated. vim +/pr_dbg +244 drivers/regulator/irq_helpers.c 153 154 static irqreturn_t regulator_notifier_isr(int irq, void *data) 155 { 156 struct regulator_irq *h = data; 157 struct regulator_irq_desc *d; 158 struct regulator_irq_data *rid; 159 unsigned long rdev_map = 0; 160 int num_rdevs; 161 int ret, i, j; 162 163 d = &h->desc; 164 rid = &h->rdata; 165 num_rdevs = rid->num_states; 166 167 if (d->fatal_cnt) 168 h->retry_cnt++; 169 170 /* 171 * we spare a few cycles by not clearing statuses prior to this call. 172 * The IC driver must initialize the status buffers for rdevs 173 * which it indicates having active events via rdev_map. 174 * 175 * Maybe we should just to be on a safer side(?) 176 */ 177 ret = d->map_event(irq, rid, &rdev_map); 178 179 /* 180 * If status reading fails (which is unlikely) we don't ack/disable 181 * IRQ but just increase fail count and retry when IRQ fires again. 182 * If retry_count exceeds the given safety limit we call IC specific die 183 * handler which can try disabling regulator(s). 184 * 185 * If no die handler is given we will just bug() as a last resort. 186 * 187 * We could try disabling all associated rdevs - but we might shoot 188 * ourselves in the head and leave the problematic regulator enabled. So 189 * if IC has no die-handler populated we just assume the regulator 190 * can't be disabled. 191 */ 192 if (unlikely(ret == REGULATOR_FAILED_RETRY)) 193 goto fail_out; 194 195 h->retry_cnt = 0; 196 /* 197 * Let's not disable IRQ if there were no status bits for us. We'd 198 * better leave spurious IRQ handling to genirq 199 */ 200 if (ret || !rdev_map) 201 return IRQ_NONE; 202 203 /* 204 * Some events are bogus if the regulator is disabled. Skip such events 205 * if all relevant regulators are disabled 206 */ 207 if (d->skip_off) { 208 for_each_set_bit(i, &rdev_map, num_rdevs) { 209 struct regulator_dev *rdev; 210 const struct regulator_ops *ops; 211 212 rdev = rid->states[i].rdev; 213 ops = rdev->desc->ops; 214 215 /* 216 * If any of the flagged regulators is enabled we do 217 * handle this 218 */ 219 if (ops->is_enabled(rdev)) 220 break; 221 } 222 if (i == num_rdevs) 223 return IRQ_NONE; 224 } 225 226 /* Disable IRQ if HW keeps line asserted */ 227 if (d->irq_off_ms) 228 disable_irq_nosync(irq); 229 230 /* 231 * IRQ seems to be for us. Let's fire correct notifiers / store error 232 * flags 233 */ 234 for_each_set_bit(i, &rdev_map, num_rdevs) { 235 unsigned long evt; 236 struct regulator_err_state *stat; 237 struct regulator_dev *rdev; 238 239 stat = &rid->states[i]; 240 rdev = stat->rdev; 241 242 for_each_set_bit(j, &stat->notifs, BITS_PER_TYPE(stat->notifs)) 243 evt = BIT(j); 245 stat->notifs, evt); 246 regulator_notifier_call_chain(rdev, evt, NULL); 247 248 rdev_flag_err(rdev, stat->errors); 249 } 250 251 if (d->irq_off_ms) { 252 if (!d->high_prio) 253 schedule_delayed_work(&h->isr_work, 254 msecs_to_jiffies(d->irq_off_ms)); 255 else 256 mod_delayed_work(system_highpri_wq, 257 &h->isr_work, 258 msecs_to_jiffies(d->irq_off_ms)); 259 } 260 261 return IRQ_HANDLED; 262 263 fail_out: 264 if (d->fatal_cnt && h->retry_cnt > d->fatal_cnt) { 265 /* If we have no recovery, just try shut down straight away */ 266 if (!d->die) { 267 hw_protection_shutdown("Regulator failure. Retry count exceeded", 268 REGULATOR_FORCED_SAFETY_SHUTDOWN_WAIT_MS); 269 } else { 270 ret = d->die(rid); 271 /* If die() failed shut down as a last attempt to save the HW */ 272 if (ret) 273 hw_protection_shutdown("Regulator failure. Recovery failed", 274 REGULATOR_FORCED_SAFETY_SHUTDOWN_WAIT_MS); 275 } 276 } 277 278 return IRQ_NONE; 279 } 280 --- 0-DAY CI Kernel Test Service, Intel Corporation https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org