Question on i40e PCIe relaxed ordering (RO)
Subsystems (2)
Hi,
Running iperf tcp test on 2 sparc systems with i40e connected back to back, I see huge number of 'port.rx_dropped' (on iperf server). Based on past experience with ixgbe, this could very well because of PCIe RO (relaxed ordering) not enabled.
I am trying to confirm RO is enabled. i40e datasheet mentioned RO settings in 3 different sections:
1. section 10.2.2.2.38 PCIe Global Config 2 - GLPCI_CNF2 register contains global status fields of PCIe configuration. The bit 0 of the register is "RO_DIS". If this bit is set to 1 RO is disabled.
RO_DIS in my setup is 0 imply RO is not disabled.
2. section 12.3.5.5 Device Control Register (0xA8; RW) has bit 4 that enable/disable RO. This is pcie cap register.
In my i40e pcie config space value at offset 0xA8 is "2110". i.e 4th bit set to 1 imply RO is enabled.
3. section 3.1.2.7.2 mentions some relaxed ordering rules e.g. "The GLLAN_RCTL.RXDESCRDROEN bit (loaded from NVM) enables relaxed ordering for Rx descriptor reads"
However, GLLAN_RCTL register definition has not bit like RXDESCRDROEN. Same goes for GLLAN_TCTL.TXDESCRDROEN.
Am I missing anything? please advise.
Thanks.
-Tushar
Quoted 22 lines · tndave
Hi,
Running iperf tcp test on 2 sparc systems with i40e connected back to back, I see huge number of 'port.rx_dropped' (on iperf server). Based on past experience with ixgbe, this could very well because of PCIe RO (relaxed ordering) not enabled.
I am trying to confirm RO is enabled. i40e datasheet mentioned RO settings in 3 different sections:
1. section 10.2.2.2.38 PCIe Global Config 2 -
GLPCI_CNF2register contains global status fields of PCIe configuration. The bit 0 of the register is "RO_DIS". If this bit is set to 1 RO is disabled.
RO_DISin my setup is 0 imply RO is not disabled.2. section 12.3.5.5 Device Control Register (
0xA8; RW) has bit 4 that enable/disable RO. This is pcie cap register.In my i40e pcie config space value at offset
0xA8is "2110". i.e 4th bit set to 1 imply RO is enabled.
Quoted 8 lines
3. section 3.1.2.7.2 mentions some relaxed ordering rules e.g. "The GLLAN_RCTL.RXDESCRDROEN bit (loaded from NVM) enables relaxed ordering for Rx descriptor reads"
However,
GLLAN_RCTLregister definition has not bit like RXDESCRDROEN. Same goes for GLLAN_TCTL.TXDESCRDROEN.Am I missing anything? please advise.
I would try posting this question to the e1000 developer list over at Source Forge. The Intel customer support folks used to monitor that list closely when I was there, hopefully they still are.
Regards,
- Greg
Quoted 9 lines
Thanks.
-Tushar
Quoted 34 lines · Greg
Quoted 4 lines · tndave
Hi,
?
Running iperf tcp test on 2 sparc systems with i40e connected back to back, I see huge number of 'port.rx_dropped' (on iperf server). Basedon
Quoted 8 lines
past experience with ixgbe, this could very well because of PCIe RO (relaxed ordering) not enabled.
?
I am trying to confirm RO is enabled. i40e datasheet mentioned RO settings in 3 different sections:
?
1. section 10.2.2.2.38 PCIe Global Config 2 -GLPCI_CNF2register contains global status fields of PCIe configuration. The bit 0 of theQuoted 8 lines
register is "
RO_DIS". If this bit is set to 1 RO is disabled. ?RO_DISin my setup is 0 imply RO is not disabled.
?
2. section 12.3.5.5 Device Control Register (0xA8; RW) has bit 4 that enable/disable RO. This is pcie cap register.
?
In my i40e pcie config space value at offset0xA8is "2110". i.e 4thbit
Quoted 10 lines
set to 1 imply RO is enabled.
?
3. section 3.1.2.7.2 mentions some relaxed ordering rules e.g. "The GLLAN_RCTL.RXDESCRDROEN bit (loaded from NVM) enables relaxed ordering for Rx descriptor reads"
?
However,GLLAN_RCTLregister definition has not bit like RXDESCRDROEN. Same goes for GLLAN_TCTL.TXDESCRDROEN.
?
Am I missing anything? please advise.
Quoted 3 lines
I would try posting this question to the e1000 developer list over at Source Forge.? The Intel customer support folks used to monitor that list closely when I was there, hopefully they still are.
We are supposed to be monitoring both lists, but just in case I have added e1000-devel list...
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://lists.osuosl.org/pipermail/intel-wired-lan/attachments/20160629/401d5a31/attachment.asc>