FPGA Intellectual Property
PCI Express*, Networking and Connectivity, Memory Interfaces, DSP IP, and Video IP
Announcements
Intel Support hours are Monday-Fridays, 8am-5pm PST, except Holidays. Thanks to our community members who provide support during our down time or before we get to your questions. We appreciate you!

Need Forum Guidance? Click here
Search our FPGA Knowledge Articles here.

PCIE link -up status

Altera_Forum
Honored Contributor II
1,501 Views

Hi , I am migrating an old design i had using a soft PCIE core to cyclone 4's hard IP. I had a 'link-up' status signal which I used to drive an LED whenever the link became active. Is there an equivalent in the PCIE hard IP implementation in QSYS. In the reference designs it seems the test_out signals played a similar role but not sure exactly how it works. thanks

0 Kudos
5 Replies
Altera_Forum
Honored Contributor II
391 Views

The link is consider up when the 10G MAC successfully transmit or receive idle to/from the link partner. Use the 10g MAC "link_fault_status_xgmii_rx" output.  

 

1) link_idle = 2b'00 <- after powerup, this show link idle because the system is still in reset stage. 

 

2) link_training = 2b'01 <- after transceiver reset sequence is complete, the link is still not stable and trying to transmit the idle here. 

 

3) link_up = 2'b00 <- the link is stable and successfully transmit/receive idle, the mac will show link up.
Altera_Forum
Honored Contributor II
391 Views

pcie_hard_ip_0_fixedclk_clk : in std_logic := '0'; -- pcie_hard_ip_0_fixedclk.clk 

pcie_hard_ip_0_pipe_ext_pipe_mode : in std_logic := '0'; -- pcie_hard_ip_0_pipe_ext.pipe_mode 

pcie_hard_ip_0_pipe_ext_phystatus_ext : in std_logic := '0'; -- .phystatus_ext 

pcie_hard_ip_0_pipe_ext_pll_powerdown : in std_logic := '0'; -- .pll_powerdown 

pcie_hard_ip_0_pipe_ext_gxb_powerdown : in std_logic := '0'; -- .gxb_powerdown 

pcie_hard_ip_0_pipe_ext_rate_ext : out std_logic; -- .rate_ext 

pcie_hard_ip_0_pipe_ext_powerdown_ext : out std_logic_vector(1 downto 0); -- .powerdown_ext 

pcie_hard_ip_0_pipe_ext_txdetectrx_ext : out std_logic; -- .txdetectrx_ext 

pcie_hard_ip_0_pipe_ext_rxelecidle0_ext : in std_logic := '0'; -- .rxelecidle0_ext 

pcie_hard_ip_0_pipe_ext_rxdata0_ext : in std_logic_vector(7 downto 0) := (others => '0'); -- .rxdata0_ext 

pcie_hard_ip_0_pipe_ext_rxstatus0_ext : in std_logic_vector(2 downto 0) := (others => '0'); -- .rxstatus0_ext 

pcie_hard_ip_0_pipe_ext_rxvalid0_ext : in std_logic := '0'; -- .rxvalid0_ext 

pcie_hard_ip_0_pipe_ext_rxdatak0_ext : in std_logic := '0'; -- .rxdatak0_ext 

pcie_hard_ip_0_pipe_ext_txdata0_ext : out std_logic_vector(7 downto 0); -- .txdata0_ext 

pcie_hard_ip_0_pipe_ext_txdatak0_ext : out std_logic; -- .txdatak0_ext 

pcie_hard_ip_0_pipe_ext_rxpolarity0_ext : out std_logic; -- .rxpolarity0_ext 

pcie_hard_ip_0_pipe_ext_txcompl0_ext : out std_logic; -- .txcompl0_ext 

pcie_hard_ip_0_pipe_ext_txelecidle0_ext : out std_logic; -- .txelecidle0_ext 

pcie_hard_ip_0_cal_blk_clk_clk : in std_logic := '0'; -- pcie_hard_ip_0_cal_blk_clk.clk 

pcie_hard_ip_0_reconfig_togxb_data : in std_logic_vector(3 downto 0) := (others => '0'); -- pcie_hard_ip_0_reconfig_togxb.data 

pcie_hard_ip_0_reconfig_busy_busy_altgxb_reconfig : in std_logic := '0'; -- pcie_hard_ip_0_reconfig_busy.busy_altgxb_reconfig 

pcie_hard_ip_0_pcie_rstn_export : in std_logic := '0'; -- pcie_hard_ip_0_pcie_rstn.export 

pcie_hard_ip_0_test_out_test_out : out std_logic_vector(63 downto 0); -- pcie_hard_ip_0_test_out.test_out 

pcie_hard_ip_0_rx_in_rx_datain_0 : in std_logic := '0'; -- pcie_hard_ip_0_rx_in.rx_datain_0 

pcie_hard_ip_0_reconfig_fromgxb_0_data : out std_logic_vector(4 downto 0); -- pcie_hard_ip_0_reconfig_fromgxb_0.data 

pcie_hard_ip_0_reconfig_gxbclk_clk : in std_logic := '0'; -- pcie_hard_ip_0_reconfig_gxbclk.clk 

pcie_hard_ip_0_tx_out_tx_dataout_0 : out std_logic; -- pcie_hard_ip_0_tx_out.tx_dataout_0 

pcie_hard_ip_0_clocks_sim_clk250_export : out std_logic; -- pcie_hard_ip_0_clocks_sim.clk250_export 

pcie_hard_ip_0_clocks_sim_clk500_export : out std_logic; -- .clk500_export 

pcie_hard_ip_0_clocks_sim_clk125_export : out std_logic; -- .clk125_export 

pcie_hard_ip_0_test_in_test_in : in std_logic_vector(39 downto 0) := (others => '0'); -- pcie_hard_ip_0_test_in.test_in 

pcie_hard_ip_0_refclk_export : in std_logic := '0' --  

 

 

this is the top level for the system i generated in QSYS (+plus some other signals from my custom components)  

 

none of these is the "k_fault_status_xgmii_rx" do i need to change some settings in order for this signal to come up? My system consist of a 1 lane PCIE hard IP, DMA and a custom memory-like module.
Altera_Forum
Honored Contributor II
391 Views

Sorry about the 10g link up answer - wrong post.  

I do have an PCIe Qsys block in our Stratix IV device. From Qsys, I instantiated an "IP_Compiler for PCI Express" (altera_pcie_hard_ip) component. This component provides a "pcie_rstn" output. We use this output as a link up status. We started with Beta Qsys software and now are using 11.0sp1. I mention this because your PICe ports are slightly different than what I have. I am not sure if it is because we started from Beta software or because you using a different Qsys component
Altera_Forum
Honored Contributor II
391 Views

i'll try this. Thanks for the response.

Altera_Forum
Honored Contributor II
391 Views

"pcie_rstn" is the PCIe edge connector reset line, so you can detect when you are hot-plugged into a host or the bus is reset. Not much point in showing this. 

 

For link state, read the datasheet, and you will find a couple of bits in pcie_hard_ip_0_test_out_test_out show what you are looking for. LTSSM states are useful to decode if you have enough LEDs - PCIe in L0 state usually is good enough for single bit display.
Reply