Two ways Intel CPU masks PCIe Errors

two ways to HasWell CPU PCIE Error Masking

in doing HASWELLCPU offers several ways to hot-plug a PCIe card device , or to forcibly disconnect the power from the PCIe card without causing the system to restart .

one way is to use PCIe and AER capability related bits, shielding these bits can prevent the above operation caused the system to restart, after clearing these bits and turn on the shield, you can continue to transfer to the CPU various errors in PCI. The following sections describe how the screen and turn on the error enable bit and then perform the corresponding PCIE card slot operation.

1. Disable AER:

I. CE/NFE/FE in the Devctl of the Disable PCIE capability Reporting

II. Set the Ue/ce mask for AER capability.

III. power-down and pull-out operation for PCIE cards

2. Enable AER:

I. Clear AER capability 's ue/ce status;

                                II.            &NBSP, Enable AER capability ' ue/ce report;

III. Enable PCIE capability in the Devctl of ce/nfe/fe reporting.

Iv.. Power-ON, insert operation for PCIE cards

the second way is to use SMI exception Enable bit. Mask The SMI exception when the error and reboot caused by the PCIE card are required to be shielded, and enable the SMI again after the operation is complete . The steps are as follows:

1. Disable SMI:

refer to 507849_haswell_eds_vol2 Page 1589 set offset 19c/1a0/1a4/1c8, i.e. gnerrmask/gferrmask/ Gcerrmask/gerrctl take gnerrmask as an example:

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/71/7E/wKiom1XR7sigeiovAAPAUHCbqKI001.jpg "title=" Smi_ Ctl.png "alt=" Wkiom1xr7sigeiovaapauhcbqki001.jpg "/>

2. Enable SMI:

Clear out the corresponding mask bit above .

This article is from the "Store Chef" blog, so be sure to keep this source http://xiamachao.blog.51cto.com/10580956/1686194

Two ways Intel CPU masks PCIe Errors