Hi, since gentoo kernels >4.9.34 I get the following two lines in dmesg (booting 4.14.8): [drm:intel_set_cpu_fifo_underrun_reporting] *ERROR* uncleared fifo underrun on pipe A [drm:intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun Scrolling up I found a message about the intel microcode being outdated so I followed https://wiki.gentoo.org/wiki/Intel_microcode#New_method_without_initram-fs.2Fdisk but that did not fix it (the microcode loads correctly). I see similar report when googling but was not able to derive a solution for myself from that research. fwiw I compared the kernel configs of 4.9.34 (last "working" version) and 4.14.8 (most recent version showing that bug) uname -a Linux pax 4.14.8-gentoo-r1 #2 SMP Fri Dec 22 09:21:40 CET 2017 x86_64 Intel(R) Core(TM) i5-4210U CPU @ 1.70GHz GenuineIntel GNU/Linux diff <(sort kernel-config-4.9.34-gentoo) <(sort linux/.config) | grep -E "[<|>] C" > CONFIG_ARCH_HAS_ADD_PAGES=y > CONFIG_ARCH_HAS_FORTIFY_SOURCE=y < CONFIG_ARCH_HAS_GIGANTIC_PAGE=y < CONFIG_ARCH_HAS_MMIO_FLUSH=y > CONFIG_ARCH_HAS_MEM_ENCRYPT=y > CONFIG_ARCH_HAS_REFCOUNT=y > CONFIG_ARCH_HAS_UACCESS_FLUSHCACHE=y > CONFIG_ARCH_HAS_ZONE_DEVICE=y < CONFIG_ARCH_WANT_FRAME_POINTERS=y > CONFIG_ARCH_WANTS_THP_SWAP=y < CONFIG_CIFS_SMB2=y > CONFIG_DRM_PANEL_BRIDGE=y < CONFIG_EXTRA_FIRMWARE="" > CONFIG_EXTRA_FIRMWARE_DIR="/lib/firmware" > CONFIG_EXTRA_FIRMWARE="intel-ucode/06-45-01" > CONFIG_FRAME_POINTER_UNWINDER=y > CONFIG_FUTEX_PI=y > CONFIG_GENERIC_IRQ_EFFECTIVE_AFF_MASK=y > CONFIG_GENERIC_IRQ_MIGRATION=y > CONFIG_HAVE_GENERIC_GUP=y > CONFIG_HAVE_HARDLOCKUP_DETECTOR_PERF=y > CONFIG_HAVE_MOD_ARCH_SPECIFIC=y > CONFIG_HAVE_RCU_TABLE_FREE=y > CONFIG_HID_ITE=y > CONFIG_IR_JVC_DECODER=y > CONFIG_IR_MCE_KBD_DECODER=y > CONFIG_IR_NEC_DECODER=y > CONFIG_IR_RC5_DECODER=y > CONFIG_IR_RC6_DECODER=y > CONFIG_IR_SANYO_DECODER=y > CONFIG_IR_SHARP_DECODER=y > CONFIG_IR_SONY_DECODER=y > CONFIG_IR_XMP_DECODER=y > CONFIG_MDIO_BUS=y < CONFIG_MMC_BLOCK_BOUNCE=y > CONFIG_NET_VENDOR_HUAWEI=y > CONFIG_NVMEM=y > CONFIG_PCI_LOCKLESS_CONFIG=y > CONFIG_RC_CORE=y > CONFIG_RC_DECODERS=y > CONFIG_RC_MAP=y > CONFIG_RTC_NVMEM=y > CONFIG_SLAB_MERGE_DEFAULT=y > CONFIG_SND_SEQ_DEVICE=y > CONFIG_SND_SEQ_MIDI_EVENT=y > CONFIG_THIN_ARCHIVES=y > CONFIG_VIDEO_IR_I2C=y > CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED=y > CONFIG_WLAN_VENDOR_QUANTENNA=y Kind regards
Following https://bbs.archlinux.org/viewtopic.php?id=196370 https://bbs.archlinux.org/viewtopic.php?id=198157 which recommend to edit /etc/X11/xorg.conf.d/20-intel.conf : Section "Device" Identifier "Intel Graphics" Driver "intel" Option "AccelMethod" "uxa" #Option "AccelMethod" "sna" EndSection Resulted in X11 not starting and the FIFO underrun messages remaining in dmesg. Following https://01.org/linuxgraphics/forum/graphics-update-tool-discussions/drmintelcpufifounderrunirqhandler-i915-error-cpu-pipe-fifo which recommends to append `i915.enable_rc6=0` to the kernel command line prevents X11 from starting but at least there are no more FIFO underrun messages in dmesg.
Good old fifo underrun. My laptop had them for years before it died. Upstream is the place to get in touch with, there are existing bugs for this on https://bugs.freedesktop.org/
I can confirm the same behavior, plus screen going to black, on 4.14.11-r2. Le sigh. This bug seemingly just won't die.
still problem with 4.14.13 ?
I am on 4.9.95 The messages are gone.