(In reply to Takashi Iwai from comment #5) > Thanks. The dmesg shows > > [ 0.164138] AMD-Vi: [Firmware Bug]: : IOAPIC[4] not in IVRS table > [ 0.164141] AMD-Vi: [Firmware Bug]: : IOAPIC[5] not in IVRS table > [ 0.164143] AMD-Vi: [Firmware Bug]: : No southbridge IOAPIC found > [ 0.164144] AMD-Vi: Disabling interrupt remapping Yes, these messages report a BIOS bug in the IVRS table setup, but I doubt that this is relevant for the issue discussed here. The IVRS bug only prevents interrupt remapping from getting enabled, normal IRQ routing should still work. > that indicates a BIOS bug, and that's the possible cause that screws up the > interrupts setup after hibernation resume. > > Could you try to boot with amd_iommu_dump=1 kernel option and give the > kernel boot messages? > > Adding Joerg to Cc, as this seems relevant with AMD IOMMU. I don't think it has to do with the AMD IOMMU, as the parts that would use the buggy BIOS settings are not active.