有人知道下面这个错误（dmesg 输出）表示什么吗？我在 Linux 下定期写入 Intel NVME 驱动器（连接到 PCI 卡）时遇到此错误。不确定“无需进一步操作”是否意味着我应该忽略它，或者 PCI 卡只是垃圾。
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: Hardware error from APEI Generic Hardware Error Source: 0
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: It has been corrected by h/w and requires no further action
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: event severity: corrected
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: Error 0, type: corrected
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: section_type: PCIe error
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: port_type: 0, PCIe end point
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: version: 3.0
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: command: 0x0506, status: 0x0010
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: device_id: 0000:17:00.0
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: slot: 0
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: secondary_bus: 0x00
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: vendor_id: 0x8086, device_id: 0xf1a6
[Mon Oct 1 13:46:53 2018] {24}[Hardware Error]: class_code: 020801
[Mon Oct 1 13:46:53 2018] nvme 0000:17:00.0: aer_status: 0x000010c0, aer_mask: 0x00002000
[Mon Oct 1 13:46:53 2018] Bad TLP, Bad DLLP, Replay Timer Timeout
[Mon Oct 1 13:46:53 2018] nvme 0000:17:00.0: aer_layer=Data Link Layer, aer_agent=Transmitter ID
[Mon Oct 1 14:21:56 2018] perf: interrupt took too long (3147 > 3135), lowering kernel.perf_event_max_sample_rate to 63500
答案1
这是 RAS 功能，告诉您有一个错误但已更正。无需针对此特定故障采取进一步措施。有时，较高的更正错误率是故障的早期指标。
合理的反应是介于忽略和丢弃磁盘之间。准备好备用磁盘，验证备份，并检查它是否作为阵列的一部分具有冗余。