Hi Arun, kernel test robot noticed the following build warnings: [auto build test WARNING on drm-tip/drm-tip] url: https://github.com/intel-lab-lkp/linux/commits/Arun-R-Murthy/drm-i915-display-dp-On-AUX-xfer-timeout-restart-freshly/20230619-163622 base: git://anongit.freedesktop.org/drm/drm-tip drm-tip patch link: https://lore.kernel.org/r/20230619082715.922094-1-arun.r.murthy%40intel.com patch subject: [PATCHv4] drm/i915/display/dp: On AUX xfer timeout restart freshly config: x86_64-rhel-8.3-rust (https://download.01.org/0day-ci/archive/20230619/202306191845.yMTzbDgG-lkp@xxxxxxxxx/config) compiler: clang version 15.0.7 (https://github.com/llvm/llvm-project.git 8dfdcc7b7bf66834a761bd8de445840ef68e4d1a) reproduce: (https://download.01.org/0day-ci/archive/20230619/202306191845.yMTzbDgG-lkp@xxxxxxxxx/reproduce) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp@xxxxxxxxx> | Closes: https://lore.kernel.org/oe-kbuild-all/202306191845.yMTzbDgG-lkp@xxxxxxxxx/ All warnings (new ones prefixed by >>): >> drivers/gpu/drm/i915/display/intel_dp_aux.c:284:12: warning: variable 'aux_clock_divider' is uninitialized when used here [-Wuninitialized] aux_clock_divider); ^~~~~~~~~~~~~~~~~ drivers/gpu/drm/i915/display/intel_dp_aux.c:222:23: note: initialize the variable 'aux_clock_divider' to silence this warning u32 aux_clock_divider; ^ = 0 1 warning generated. vim +/aux_clock_divider +284 drivers/gpu/drm/i915/display/intel_dp_aux.c 209 210 static int 211 intel_dp_aux_xfer(struct intel_dp *intel_dp, 212 const u8 *send, int send_bytes, 213 u8 *recv, int recv_size, 214 u32 aux_send_ctl_flags) 215 { 216 struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); 217 struct drm_i915_private *i915 = 218 to_i915(dig_port->base.base.dev); 219 enum phy phy = intel_port_to_phy(i915, dig_port->base.port); 220 bool is_tc_port = intel_phy_is_tc(i915, phy); 221 i915_reg_t ch_ctl, ch_data[5]; 222 u32 aux_clock_divider; 223 enum intel_display_power_domain aux_domain; 224 intel_wakeref_t aux_wakeref; 225 intel_wakeref_t pps_wakeref; 226 int i, ret, recv_bytes; 227 int try, clock = 0; 228 u32 status; 229 u32 send_ctl; 230 bool vdd; 231 232 ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp); 233 for (i = 0; i < ARRAY_SIZE(ch_data); i++) 234 ch_data[i] = intel_dp->aux_ch_data_reg(intel_dp, i); 235 236 if (is_tc_port) { 237 intel_tc_port_lock(dig_port); 238 /* 239 * Abort transfers on a disconnected port as required by 240 * DP 1.4a link CTS 4.2.1.5, also avoiding the long AUX 241 * timeouts that would otherwise happen. 242 * TODO: abort the transfer on non-TC ports as well. 243 */ 244 if (!intel_tc_port_connected_locked(&dig_port->base)) { 245 ret = -ENXIO; 246 goto out_unlock; 247 } 248 } 249 250 aux_domain = intel_aux_power_domain(dig_port); 251 252 aux_wakeref = intel_display_power_get(i915, aux_domain); 253 pps_wakeref = intel_pps_lock(intel_dp); 254 255 /* 256 * We will be called with VDD already enabled for dpcd/edid/oui reads. 257 * In such cases we want to leave VDD enabled and it's up to upper layers 258 * to turn it off. But for eg. i2c-dev access we need to turn it on/off 259 * ourselves. 260 */ 261 vdd = intel_pps_vdd_on_unlocked(intel_dp); 262 263 /* 264 * dp aux is extremely sensitive to irq latency, hence request the 265 * lowest possible wakeup latency and so prevent the cpu from going into 266 * deep sleep states. 267 */ 268 cpu_latency_qos_update_request(&intel_dp->pm_qos, 0); 269 270 intel_pps_check_power_unlocked(intel_dp); 271 272 /* 273 * FIXME PSR should be disabled here to prevent 274 * it using the same AUX CH simultaneously 275 */ 276 277 /* Only 5 data registers! */ 278 if (drm_WARN_ON(&i915->drm, send_bytes > 20 || recv_size > 20)) { 279 ret = -E2BIG; 280 goto out; 281 } 282 send_ctl = intel_dp->get_aux_send_ctl(intel_dp, 283 send_bytes, > 284 aux_clock_divider); 285 send_ctl |= aux_send_ctl_flags; 286 287 while ((aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, clock++))) { 288 /* Re-visit : Must try at least 3 times according to DP spec */ 289 for (try = 0; try < 5; try++) { 290 /* Try to wait for any previous AUX channel activity */ 291 status = intel_dp_aux_wait_done(intel_dp); 292 /* just trace the final value */ 293 trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true); 294 295 if (status & DP_AUX_CH_CTL_SEND_BUSY) { 296 drm_WARN(&i915->drm, 1, 297 "%s: not started, previous Tx still in process (status 0x%08x)\n", 298 intel_dp->aux.name, status); 299 intel_dp->aux_busy_last_status = status; 300 if (try > 3) { 301 ret = -EBUSY; 302 goto out; 303 } else 304 continue; 305 } 306 307 /* Load the send data into the aux channel data registers */ 308 for (i = 0; i < send_bytes; i += 4) 309 intel_de_write(i915, ch_data[i >> 2], 310 intel_dp_aux_pack(send + i, 311 send_bytes - i)); 312 313 /* Send the command and wait for it to complete */ 314 intel_de_write(i915, ch_ctl, send_ctl); 315 316 /* TODO: if typeC then 4.2ms else 800us. For DG2 add 1.5ms for both cases */ 317 status = intel_dp_aux_wait_done(intel_dp); 318 319 /* Clear done status and any errors */ 320 intel_de_write(i915, ch_ctl, 321 status | DP_AUX_CH_CTL_DONE | 322 DP_AUX_CH_CTL_TIME_OUT_ERROR | 323 DP_AUX_CH_CTL_RECEIVE_ERROR); 324 325 /* 326 * DP CTS 1.2 Core Rev 1.1, 4.2.1.1 & 4.2.1.2 327 * 400us delay required for errors and timeouts 328 * Timeout errors from the HW already meet this 329 * requirement so skip to next iteration 330 */ 331 if (status & (DP_AUX_CH_CTL_TIME_OUT_ERROR | 332 DP_AUX_CH_CTL_SEND_BUSY)) 333 continue; 334 335 if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) { 336 usleep_range(400, 500); 337 continue; 338 } 339 if (status & DP_AUX_CH_CTL_DONE) 340 goto done; 341 } 342 } 343 344 if ((status & DP_AUX_CH_CTL_DONE) == 0) { 345 drm_err(&i915->drm, "%s: not done (status 0x%08x)\n", 346 intel_dp->aux.name, status); 347 ret = -EBUSY; 348 goto out; 349 } 350 351 done: 352 /* 353 * Check for timeout or receive error. Timeouts occur when the sink is 354 * not connected. 355 */ 356 if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) { 357 drm_err(&i915->drm, "%s: receive error (status 0x%08x)\n", 358 intel_dp->aux.name, status); 359 ret = -EIO; 360 goto out; 361 } 362 363 /* 364 * Timeouts occur when the device isn't connected, so they're "normal" 365 * -- don't fill the kernel log with these 366 */ 367 if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR) { 368 drm_dbg_kms(&i915->drm, "%s: timeout (status 0x%08x)\n", 369 intel_dp->aux.name, status); 370 ret = -ETIMEDOUT; 371 goto out; 372 } 373 374 /* Unload any bytes sent back from the other side */ 375 recv_bytes = REG_FIELD_GET(DP_AUX_CH_CTL_MESSAGE_SIZE_MASK, status); 376 377 /* 378 * By BSpec: "Message sizes of 0 or >20 are not allowed." 379 * We have no idea of what happened so we return -EBUSY so 380 * drm layer takes care for the necessary retries. 381 */ 382 if (recv_bytes == 0 || recv_bytes > 20) { 383 drm_dbg_kms(&i915->drm, 384 "%s: Forbidden recv_bytes = %d on aux transaction\n", 385 intel_dp->aux.name, recv_bytes); 386 ret = -EBUSY; 387 goto out; 388 } 389 390 if (recv_bytes > recv_size) 391 recv_bytes = recv_size; 392 393 for (i = 0; i < recv_bytes; i += 4) 394 intel_dp_aux_unpack(intel_de_read(i915, ch_data[i >> 2]), 395 recv + i, recv_bytes - i); 396 397 ret = recv_bytes; 398 out: 399 cpu_latency_qos_update_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE); 400 401 if (vdd) 402 intel_pps_vdd_off_unlocked(intel_dp, false); 403 404 intel_pps_unlock(intel_dp, pps_wakeref); 405 intel_display_power_put_async(i915, aux_domain, aux_wakeref); 406 out_unlock: 407 if (is_tc_port) 408 intel_tc_port_unlock(dig_port); 409 410 return ret; 411 } 412 -- 0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
