我最近注意到我的一个 CPU 闲置率约为 85-90%,并且根据top
使用情况来自中断,因此就像这个问题一样我结合使用dmesg
和 定期cat
ing/proc/interrupts
并发现了这一点:
CPU0 CPU1 CPU2 CPU3
0: 17 0 0 0 IR-IO-APIC 2-edge timer
1: 11548 0 2429 0 IR-IO-APIC 1-edge i8042
8: 0 0 0 1 IR-IO-APIC 8-edge rtc0
9: 7 16 0 0 IR-IO-APIC 9-fasteoi acpi
12: 14530 108887 0 0 IR-IO-APIC 12-edge i8042
16: 78464100 0 0 11702812 IR-IO-APIC 16-fasteoi idma64.0, i2c_designware.0, i801_smbus
120: 0 0 0 0 DMAR-MSI 0-edge dmar0
121: 0 0 0 0 DMAR-MSI 1-edge dmar1
正如你所看到的,IRQ #16 疯狂地发送中断(每次 CPU 从 S3 唤醒时,它似乎开始向不同的 CPU 发送垃圾邮件),我还发现我的触摸板使用相同的 IRQ,并且如果启用了 I2C 模式(或者先进的模式,根据我的 BIOS),它随机停止工作并显示以下消息(来自dmesg
):
[ 167.851139] irq 16: nobody cared (try booting with the "irqpoll" option)
[ 167.851158] CPU: 2 PID: 3874 Comm: firefox Not tainted 4.15.3-300.fc27.x86_64 #1
[ 167.851160] Hardware name: Acer Aspire E5-575/Ironman_SK , BIOS V1.04 04/26/2016
[ 167.851162] Call Trace:
[ 167.851171] <IRQ>
[ 167.851185] dump_stack+0x5c/0x85
[ 167.851193] __report_bad_irq+0x30/0xc0
[ 167.851196] note_interrupt+0x235/0x280
[ 167.851198] handle_irq_event_percpu+0x51/0x70
[ 167.851201] handle_irq_event+0x27/0x50
[ 167.851204] handle_fasteoi_irq+0x6b/0x120
[ 167.851209] handle_irq+0xaf/0x120
[ 167.851214] do_IRQ+0x41/0xc0
[ 167.851219] common_interrupt+0xa2/0xa2
[ 167.851222] </IRQ>
[ 167.851224] RIP: 0010:_raw_spin_lock+0x10/0x20
[ 167.851226] RSP: 0000:ffffa85a857dfdd0 EFLAGS: 00000246 ORIG_RAX: ffffffffffffffdb
[ 167.851230] RAX: 0000000000000000 RBX: ffff8d0a268930c8 RCX: 00003ffffffff000
[ 167.851231] RDX: 0000000000000001 RSI: 8000000000000025 RDI: ffffd21648d7ca70
[ 167.851232] RBP: ffffd2164892e100 R08: 0000000000000000 R09: 0000000000171800
[ 167.851233] R10: 0000000000271800 R11: 0000000000001000 R12: 0000000000000000
[ 167.851234] R13: 8000000224b84867 R14: ffffd21648d7ca70 R15: ffff8d0a35f29810
[ 167.851244] __handle_mm_fault+0xa4c/0x1290
[ 167.851249] handle_mm_fault+0xaa/0x1f0
[ 167.851255] __do_page_fault+0x25d/0x4e0
[ 167.851262] ? SyS_mmap_pgoff+0xfb/0x250
[ 167.851264] do_page_fault+0x32/0x110
[ 167.851267] ? page_fault+0x36/0x60
[ 167.851269] page_fault+0x4c/0x60
[ 167.851272] RIP: 0033:0x7ff86dc0b205
[ 167.851273] RSP: 002b:00007ffe6493e888 EFLAGS: 00010206
[ 167.851276] handlers:
[ 167.851291] [<0000000061f0cd43>] idma64_irq [idma64]
[ 167.851296] [<000000006909bb35>] i2c_dw_isr
[ 167.851302] [<00000000ee17d26d>] i801_isr [i2c_i801]
[ 167.851304] Disabling IRQ #16
这是硬件问题吗?我能做些什么?
最后我有机会深入研究这个问题,通过运行lspci -nnkv
我发现有 2 个设备正在使用 IRQ 16:
00:15.0 Signal processing controller [1180]: Intel Corporation Sunrise Point-LP Serial IO I2C Controller #0 [8086:9d60] (rev 21)
Subsystem: Acer Incorporated [ALI] Device [1025:1094]
Flags: fast devsel, IRQ 16
Memory at a132b000 (64-bit, non-prefetchable) [size=4K]
Capabilities: [80] Power Management version 3
Capabilities: [90] Vendor Specific Information: Len=14 <?>
Kernel driver in use: intel-lpss
Kernel modules: intel_lpss_pci
和:
00:1f.4 SMBus [0c05]: Intel Corporation Sunrise Point-LP SMBus [8086:9d23] (rev 21)
Subsystem: Acer Incorporated [ALI] Device [1025:1094]
Flags: medium devsel, IRQ 16
Memory at a132e000 (64-bit, non-prefetchable) [size=256]
I/O ports at 4040 [size=32]
Kernel driver in use: i801_smbus
Kernel modules: i2c_i801
intel_lpss_pci
如果我卸载模块,问题似乎就会消失rmmod intel_lpss_pci
,但触摸板当然会停止工作。但我想这比 CPU 始终处于 100% 更好。
答案1
可能需要将其添加到 PM 挂起列表中,以便在挂起/解冻周期后重新加载。
SUSPEND_MODULES="brcmfmac intel_lpss_pci intel_lpss_acpi"