PCIe Master Abort

取消
显示结果 
显示  仅  | 搜索替代 
您的意思是: 

PCIe Master Abort

3,893 次查看
emilviding
Contributor I

Hi,

I have a P4080 configured as RC connected to a P1022 through a switch. During a test the P4080 read data from P1022 and P1022 was reseted.

When this happened the P4080 crashed. I attached a Lauterbach debugging probe and saw that the Received_master_abort bit in Status_Register was set. I wasn't able to do any configuration accesses to any bus but the internal (bus 0). Only a hot reset resolved this issue.

After consulting the datasheet I thought that the master abort was caused by a unsupported request completion. I disabled the unsupported request error reporting in the switch but that didn't solve anything.

Am I on the right track? Is the error caused by some error forwarded to the P4080 because it's RC? I have disabled every error bit I can find without any progress.

Regards,

Emil

标签 (1)
标记 (1)
0 项奖励
回复
1 回复

3,035 次查看
emilviding
Contributor I

The problem seems to be PCIe Ack time-out (PEX_ERR_DR[PAT]). When that bit is set it's not possible to access any external PCIe bus. In some cases it's possible to regain access by clearing the bit but not always. Disabling the bit does nothing to solve the problem, it just masks the error.

I want to be able to access other endpoints even when the P1022 is in reset. It that possible?

/Emil

0 项奖励
回复