- ixp1200 - lists.cs.princeton.edu

RE: [ixp1200] what is the difference between IXP1200 Revision C0 and the olders
by Ye, Irvin 25 Nov '02

25 Nov '02

Hi, all, Is there a 2.4.18 for ixp1200? Where can I find it? What I, and a lot of others, I think, are still using 2.3.99. Irvin -----Original Message----- From: Li Weihua [mailto:liweihua@cosix.com.cn] Sent: 25 November 2002 01:40 To: ixp1200(a)CS.Princeton.EDU Subject: [ixp1200] what is the difference between IXP1200 Revision C0 and the olders Hi, The README of "inux 2.4.18 for ixp1200" says that it can not run on silicon older then Rev C0 IXP1200. But, what is the difference between the Rev C0 and the olders? And where can I find the detailed documents about this? Thanks a lot! 　　　　　　　　Li Weihua 　　　　　　　　liweihua(a)cosix.com.cn 　　　　　　　　　　2002-11-25

2 1

what is the difference between IXP1200 Revision C0 and the olders
by Li Weihua 25 Nov '02

25 Nov '02

Hi, The README of "inux 2.4.18 for ixp1200" says that it can not run on silicon older then Rev C0 IXP1200. But, what is the difference between the Rev C0 and the olders? And where can I find the detailed documents about this? Thanks a lot! 　　　　　　　　Li Weihua 　　　　　　　　liweihua(a)cosix.com.cn 　　　　　　　　　　2002-11-25

2 1

Re: FW: [ixp1200] Regarding to IXP1200 Ethernet Port configuration
by Ning Weng 22 Nov '02

22 Nov '02

Joel, Thanks a lot. It works perfect for me, thank you so much for you so detail instructions. After run the L3forearder example, I could ping each ports of the ENP2505 card from my PC. Attached file(ixsys_count_8_1.config) is a configuration file for the simple counting application from the SDK CDs. Would you like to help me to have a look at it? Compared ixsys_count_8_1.config with ixsys.config-l3fwdr, looks like that they are pretty much same in terms of format. Why I can ping ports after running ./ixstart ixsys.config-l3fwdr; but I cannot ping after running ./ixstart ixsys_8_1.config? what I mean, the total number of packet doesn't increase even though I pinged the right port. Thanks so much for your time and your help. Really appreciate it. P.S. Should I have to reboot my IXP1200 after running an application and before load another one? -ning "Schuetze, Joel D" wrote: > Greetings, > > Attached is a document and two patch files. The document describes the > process for applying the patch, setting up the L3forwarder example, and > testing it. Let me know if you have any questions regarding it. > > Joel Schuetze > > -----Original Message----- > From: Brandewie, Dirk J > Sent: Tuesday, November 19, 2002 10:26 AM > To: Schuetze, Joel D > Subject: FW: [ixp1200] Regarding to IXP1200 Ethernet Port configuration > > -----Original Message----- > From: Ning Weng [mailto:nweng@ecs.umass.edu] > Sent: Tuesday, November 19, 2002 6:37 AM > To: ixp1200(a)CS.Princeton.EDU > Subject: [ixp1200] Regarding to IXP1200 Ethernet Port configuration > > Hi, > > I have one question here regarding to IXP1200 ENP2505. > > My case is I reboot my target(IXP1200) from my host machine through a > serial cable, and I can ping each other via virtual interface( PCI) as > well. Now I want the two system(IXP1200 and my host Linux machine) > communicate by Ethernet port via a cross over cable. My question is : > what is the detail steps that I should do, such as network > configurations or other? That will be a great help for me if you would > like share your experiences or point me the the documentation that you > followed. > > Thanks a lot. > > Ning > > ------------------------------------------------------------------------ > Name: Setup and Running of Layer 3 Forwarder Example.ZIP > Setup and Running of Layer 3 Forwarder Example.ZIP Type: Zip Compressed Data (application/x-zip-compressed) > Encoding: base64 > > Name: config_1.ZIP > config_1.ZIP Type: Zip Compressed Data (application/x-zip-compressed) > Encoding: base64 > > Name: board.ZIP > board.ZIP Type: Zip Compressed Data (application/x-zip-compressed) > Encoding: base64 # ********************************************************************** # # Sample file for the ixconfig application # # It is run at boot time and specifies # # - the interfaces that are to be started at system boot time # - the micro aces that are to be started at system boot time # - the regular aces that are to be started at system boot time # - bind configuration # - Shell commands to be run # # For each microace or ace it also lets us specify another # configuration file that can be used to configure that ace # # # ********************************************************************** # # First we specify the interfaces we want # # For the SI board # -- 0-15 are fast ethernet ports. 16 and 17 are gigabit ports # # <port number> <ip addr> <broadcast> <netmask> <mac address> <flags> # # Values for flags are # # 0x0 Unicast # 0x1 Promiscous Mode # 0x2 All Multicast packets are allowed # 0x3 Multicast packets in set only # interface 0 10.1.0.1 10.1.0.255 255.255.255.0 00:01:02:03:04:05 1 interface 1 10.2.0.1 10.2.0.255 255.255.255.0 00:01:02:03:04:06 1 interface 2 10.3.0.1 10.3.0.255 255.255.255.0 00:01:02:03:04:07 1 interface 3 10.4.0.1 10.4.0.255 255.255.255.0 00:01:02:03:04:08 1 interface 4 10.5.0.1 10.5.0.255 255.255.255.0 00:01:02:03:04:09 1 interface 5 10.6.0.1 10.6.0.255 255.255.255.0 00:01:02:03:04:10 1 interface 6 10.7.0.1 10.7.0.255 255.255.255.0 00:01:02:03:04:11 1 interface 7 10.8.0.1 10.8.0.255 255.255.255.0 00:01:02:03:04:12 1 interface 8 10.9.0.1 10.9.0.255 255.255.255.0 00:01:02:03:04:13 1 interface 9 10.10.0.1 10.10.0.255 255.255.255.0 00:01:02:03:04:14 1 interface 10 10.11.0.1 10.11.0.255 255.255.255.0 00:01:02:03:04:15 1 interface 11 10.12.0.1 10.12.0.255 255.255.255.0 00:01:02:03:04:16 1 interface 12 10.13.0.1 10.13.0.255 255.255.255.0 00:01:02:03:04:17 1 interface 13 10.14.0.1 10.14.0.255 255.255.255.0 00:01:02:03:04:18 1 interface 14 10.15.0.1 10.15.0.255 255.255.255.0 00:01:02:03:04:19 1 interface 15 10.16.0.1 10.16.0.255 255.255.255.0 00:01:02:03:04:20 1 interface 16 10.17.0.1 10.17.0.255 255.255.255.0 00:01:02:03:04:21 1 interface 17 10.18.0.1 10.18.0.255 255.255.255.0 00:01:02:03:04:22 1 # ********************************************************************* # Specify if we will be debugging with the workbench and downloading # code via it # # mode <mode> # # Values for mode are # # 0x0 No workbench. # 0x1 Download and debug via workbench #mode 0 mode 1 # ********************************************************************* # Next specify the Microcode files (UOF Files) # # <fileType> <fileName> # # Values for fileType are # # 0x0 Slow Ingress File # 0x1 Slow Egress File # 0x2 Fast Ingress File for Port 1 # 0x3 Fast Ingress File for Port 2 # 0x4 Fast Egress File # file 0 ./SlowIngressCount.uof file 1 ./SlowEgressRR.uof file 2 ./FastIngressCount-seq1.uof file 3 ./FastIngressCount-seq2.uof file 4 ./FastEgressFifo.uof # ********************************************************************** # # Now specify the microaces # # <name of ace> <name of executable> <config file name> <runsOn> \ # <type> <additional parameters> # # runsOn can be -- 0 (RUNS_ON_INGRESS_SIDE) # -- 1 (RUNS_ON_EGRESS_SIDE) # # type can be -- 0 (unknown type) # 1 (Ingress) # 2 (egress) # 3 (L3) # 4 (L2) # 5 (input nat) # 6 (output nat) # # # The first parameter to every microace is the mask of microengines it # runs on. This is passed automatically by the application. The <additional # parameters> specify the rest of the command line parameters # # no config file for ingress and egress needed and no parameters microace ifaceInput ./ingressAce none 0 1 microace ifaceOutput ./egressAce none 1 2 # This microace will substitute for L3 for now # # options for CountMicroAce are # # -v : verbose # -target : send packets over the default target # -loop : loop back packets to the microblock # -blockname <name> : name of the microblock # -port <portNumber>: output port to send to (default is dont change what's in # the ofnum of the packet) # microace CountMicroAce ./CountMicroAce 0 0 none -target -blockname COUNT # ********************************************************************** # # Now specify the regular aces like stack ace, nat control ace, spanning # tree ace etc # # <name of ace> <name of executable> <config file name> <type> <parameters> # # # type can be -- 0 (unknown type) # 1 (Spanning Tree) # 2 (Nat control) # 3 (Stack ace) # # <parameters> are any command line (argc/argv) parameters to be passed to # the ace # # ********************************************************************** # ********************************************************************** # # Now specify the all the bind operations # # <static/regular> <target name> <ace name> # # Specify static for microace to microace static targets # bind static ifaceInput/default CountMicroAce bind static CountMicroAce/default ifaceOutput # ********************************************************************** # Now specify any shell commands. These will always be executed last # no matter where they are specified. Note that these are to be used # typically for things like adding arp entries to the linux stack. # # Below we add arp entries to the linux stack. This is needed because # the traffic generator we are using does not respond to ARP requests. # Otherwise this is typically not needed as long as the ARP target of # of the interface ace is bound to the stack ace # # Below we add a couple of arp entries # # Syntax is # # sh [command string to be executed] # # sh arp -s 10.1.0.2 01:02:03:04:05:06 # sh arp -s 10.2.0.2 02:02:03:04:05:06 # sh arp -s 10.3.0.2 03:02:03:04:05:06 # sh arp -s 10.4.0.2 04:02:03:04:05:06 # sh arp -s 10.5.0.2 05:02:03:04:05:06 # sh arp -s 10.6.0.2 06:02:03:04:05:06 # sh arp -s 10.7.0.2 07:02:03:04:05:06 # sh arp -s 10.8.0.2 08:02:03:04:05:06 # sh arp -s 10.9.0.2 09:02:03:04:05:06 # **********************************************************************

1 0

RE: [ixp1200] How to communicate IXP1200 ENP2505 card with PC usi ng Ethernet ports
by Ye, Irvin 22 Nov '02

22 Nov '02

Hi, Nimit, I used ixsys.config-sys-l3fwdr (along with route.config-sys-l3fwdr). Is that OK. I really saw some other simliar files. Regards. Irvin -----Original Message----- From: Nimit S Sawhney [mailto:nimits@cmu.edu] Sent: 21 November 2002 18:11 To: ixp1200(a)CS.Princeton.EDU Subject: Re: [ixp1200] How to communicate IXP1200 ENP2505 card with PC using Ethernet ports Irvin, What config file(s) did you use when you called "ixstart" for the l3fwdr? -Nimit ----- Original Message ----- From: "Ye, Irvin" <i.ye(a)lancaster.ac.uk> To: <ixp1200(a)CS.Princeton.EDU> Sent: Thursday, November 21, 2002 5:25 AM Subject: RE: [ixp1200] How to communicate IXP1200 ENP2505 card with PC using Ethernet ports > Hi, Nimit, > > I've tried Intel's samples, both l2bridge and l3forward, but failed to let them > work. There're always errors in the egress side. The ueng always stops to wait > for the ports to be ready, while the ports keep unready. Would you give me some > hints? > > Irvin > > -----Original Message----- > From: Nimit S Sawhney [mailto:nimits@cmu.edu] > Sent: 20 November 2002 21:29 > To: ixp1200(a)CS.Princeton.EDU > Subject: Re: [ixp1200] How to communicate IXP1200 ENP2505 card with PC using > Ethernet ports > > > Ning... > Try running the hardware version of the l3fwdr project on the IXP by > configuring it for the number of interfaces you want to use. > > -Nimit > > ----- Original Message ----- > From: "Ning Weng" <nweng(a)ecs.umass.edu> > To: <ixp1200(a)CS.Princeton.EDU> > Cc: "Ye, Irvin" <i.ye(a)lancaster.ac.uk> > Sent: Wednesday, November 20, 2002 4:01 PM > Subject: [ixp1200] How to communicate IXP1200 ENP2505 card with PC using > Ethernet ports > > > > > > Please forgive me to let me clarify my question: > > > > How to communicate my IXP1200 ENP2505 Card with my PC using Ethernet ports > > presented on the ENP2505 card? > > > > Yes, I can boot the IXP1200 using serial cable, and also can ping my PC > > through PCI bus. > > > > Thanks. > > > > > > > > On Wed, 20 Nov 2002, Ye, Irvin wrote: > > > > > No need to do any configuration, ixconfig will do for you. > > > > > > As for packet generator, Intel suggest use IXIA or Shomiti, but I even > don't > > > know what are they? Who know? do tell me. If they're some kinds of > special > > > devices, I can't afford, I'm affraid. So I just use Linux host to > generate to > > > packets. > > > > > > Irvin > > > > > > -----Original Message----- > > > From: Ning Weng [mailto:nweng@sisko.ecs.umass.edu] > > > Sent: 18 November 2002 22:45 > > > To: ixp1200(a)CS.Princeton.EDU; Ye, Irvin > > > Subject: Re: [ixp1200] SOS, a question about connection and output! > > > > > > > > > > > > Irvin, > > > > > > Sorry for not trying to answer your question, but ask you question.... > > > > > > > > > Q1: My case is I reboot my target(IXP1200) from my host machine through > a > > > serial cable, and I can ping each other via virtual interface( PCI) as > well. > > > Now I want the two system(IXP1200 and my host Linux machine) communicate > by > > > Ethernet port via a cross over cable. My question is : what is the > detail > > > steps that I should do, such as network configurations or other? That > will be a > > > great help for me if you would like share your experiences or point me > the the > > > documentation that you followed. > > > Q2: What is the packet generator that you used? > > > > > > > > > Thanks a lot. > > > > > > > > > Ning > > > > > > > > > > > > > > > "Ye, Irvin" > > > > > > Q2: > > > > > > > > > > > > "Ye, Irvin" wrote: > > > > > > > > > Hi,I'm puzzled on the connection for ixp1200.I had tried to connect PCs > and > > > ixp1200 ports in standard way: via a hub / switch, but it failed to > work. When > > > I tried direct connection through crossover twisted pairs, it seemed OK, > > > packets could be successfully sent to ixp1200 ports. But then it found > that > > > packets could not be sent from the ports. When I tried Intel's simple > bridge > > > example, I found there were problem in the egress side: packets could > not be > > > sent out, so finally there was a RmSendPacket error. Upon Erik's > suggestion, > > > I've checked some status of the board and found that when there were > packets to > > > send, the value of the XMIT_PTR regisger would stick to a fix number, > and the > > > active transmit thread would cycle around these codes: > > > Egress_WaitForPortReady[abs_port] ctx_arb[sdram]I've tried long, but > failed > > > to resolve the problem. It's making me crazy. Who can give me a > > > help?Regards.Irvin > > > > > > > > > > > > > >

1 0

RE: [ixp1200] How to communicate IXP1200 ENP2505 card with PC usi ng Ethernet ports
by Ye, Irvin 21 Nov '02

21 Nov '02

Hi, Nimit, I've tried Intel's samples, both l2bridge and l3forward, but failed to let them work. There're always errors in the egress side. The ueng always stops to wait for the ports to be ready, while the ports keep unready. Would you give me some hints? Irvin -----Original Message----- From: Nimit S Sawhney [mailto:nimits@cmu.edu] Sent: 20 November 2002 21:29 To: ixp1200(a)CS.Princeton.EDU Subject: Re: [ixp1200] How to communicate IXP1200 ENP2505 card with PC using Ethernet ports Ning... Try running the hardware version of the l3fwdr project on the IXP by configuring it for the number of interfaces you want to use. -Nimit ----- Original Message ----- From: "Ning Weng" <nweng(a)ecs.umass.edu> To: <ixp1200(a)CS.Princeton.EDU> Cc: "Ye, Irvin" <i.ye(a)lancaster.ac.uk> Sent: Wednesday, November 20, 2002 4:01 PM Subject: [ixp1200] How to communicate IXP1200 ENP2505 card with PC using Ethernet ports > > Please forgive me to let me clarify my question: > > How to communicate my IXP1200 ENP2505 Card with my PC using Ethernet ports > presented on the ENP2505 card? > > Yes, I can boot the IXP1200 using serial cable, and also can ping my PC > through PCI bus. > > Thanks. > > > > On Wed, 20 Nov 2002, Ye, Irvin wrote: > > > No need to do any configuration, ixconfig will do for you. > > > > As for packet generator, Intel suggest use IXIA or Shomiti, but I even don't > > know what are they? Who know? do tell me. If they're some kinds of special > > devices, I can't afford, I'm affraid. So I just use Linux host to generate to > > packets. > > > > Irvin > > > > -----Original Message----- > > From: Ning Weng [mailto:nweng@sisko.ecs.umass.edu] > > Sent: 18 November 2002 22:45 > > To: ixp1200(a)CS.Princeton.EDU; Ye, Irvin > > Subject: Re: [ixp1200] SOS, a question about connection and output! > > > > > > > > Irvin, > > > > Sorry for not trying to answer your question, but ask you question.... > > > > > > Q1: My case is I reboot my target(IXP1200) from my host machine through a > > serial cable, and I can ping each other via virtual interface( PCI) as well. > > Now I want the two system(IXP1200 and my host Linux machine) communicate by > > Ethernet port via a cross over cable. My question is : what is the detail > > steps that I should do, such as network configurations or other? That will be a > > great help for me if you would like share your experiences or point me the the > > documentation that you followed. > > Q2: What is the packet generator that you used? > > > > > > Thanks a lot. > > > > > > Ning > > > > > > > > > > "Ye, Irvin" > > > > Q2: > > > > > > > > "Ye, Irvin" wrote: > > > > > > Hi,I'm puzzled on the connection for ixp1200.I had tried to connect PCs and > > ixp1200 ports in standard way: via a hub / switch, but it failed to work. When > > I tried direct connection through crossover twisted pairs, it seemed OK, > > packets could be successfully sent to ixp1200 ports. But then it found that > > packets could not be sent from the ports. When I tried Intel's simple bridge > > example, I found there were problem in the egress side: packets could not be > > sent out, so finally there was a RmSendPacket error. Upon Erik's suggestion, > > I've checked some status of the board and found that when there were packets to > > send, the value of the XMIT_PTR regisger would stick to a fix number, and the > > active transmit thread would cycle around these codes: > > Egress_WaitForPortReady[abs_port] ctx_arb[sdram]I've tried long, but failed > > to resolve the problem. It's making me crazy. Who can give me a > > help?Regards.Irvin > > > > > > >

2 1

RE: [ixp1200] WorkBench in HARDWARE Mode?
by Ye, Irvin 21 Nov '02

21 Nov '02

Thanks! -----Original Message----- From: Abhijeet Joglekar [mailto:abhijeet@cs.utah.edu] Sent: 21 November 2002 13:33 To: 'ixp1200(a)cs.princeton.edu' Subject: Re: [ixp1200] WorkBench in HARDWARE Mode? Irvin, There was some discussion about this in August 2002 on this mailing list. Try looking in the archives to see if that is useful to you. -- Abhijeet On Thu, 21 Nov 2002, Ye, Irvin wrote: > Hi, > > Now, I meet problems in using WorkBench in HARDWARE mode. The microcodes have > been successfully loaded and ran, but failed to get any packets from the ports. > The same examples from Intel, if run out of WorkBench, it can receive packets > from the ports; but if run in WorkBench, it fails. I've checked the > correspondent registers, eg, RCV_RDY_LO. The value sounded abnormal in the > later case. So, does it mean that ports or FBI and like stuffs are not > properlly configured in WorkBench. Those honorable veterans, who can give me > some hints? > > Thank in advance! > > Irvin > > >

1 0

WorkBench in HARDWARE Mode?
by Ye, Irvin 21 Nov '02

21 Nov '02

Hi, Now, I meet problems in using WorkBench in HARDWARE mode. The microcodes have been successfully loaded and ran, but failed to get any packets from the ports. The same examples from Intel, if run out of WorkBench, it can receive packets from the ports; but if run in WorkBench, it fails. I've checked the correspondent registers, eg, RCV_RDY_LO. The value sounded abnormal in the later case. So, does it mean that ports or FBI and like stuffs are not properlly configured in WorkBench. Those honorable veterans, who can give me some hints? Thank in advance! Irvin

2 1

RE: [ixp1200] SOS, a question about connection and output!
by Ye, Irvin 20 Nov '02

20 Nov '02

No need to do any configuration, ixconfig will do for you. As for packet generator, Intel suggest use IXIA or Shomiti, but I even don't know what are they? Who know? do tell me. If they're some kinds of special devices, I can't afford, I'm affraid. So I just use Linux host to generate to packets. Irvin -----Original Message----- From: Ning Weng [mailto:nweng@sisko.ecs.umass.edu] Sent: 18 November 2002 22:45 To: ixp1200(a)CS.Princeton.EDU; Ye, Irvin Subject: Re: [ixp1200] SOS, a question about connection and output! Irvin, Sorry for not trying to answer your question, but ask you question.... Q1: My case is I reboot my target(IXP1200) from my host machine through a serial cable, and I can ping each other via virtual interface( PCI) as well. Now I want the two system(IXP1200 and my host Linux machine) communicate by Ethernet port via a cross over cable. My question is : what is the detail steps that I should do, such as network configurations or other? That will be a great help for me if you would like share your experiences or point me the the documentation that you followed. Q2: What is the packet generator that you used? Thanks a lot. Ning "Ye, Irvin" Q2: "Ye, Irvin" wrote: Hi,I'm puzzled on the connection for ixp1200.I had tried to connect PCs and ixp1200 ports in standard way: via a hub / switch, but it failed to work. When I tried direct connection through crossover twisted pairs, it seemed OK, packets could be successfully sent to ixp1200 ports. But then it found that packets could not be sent from the ports. When I tried Intel's simple bridge example, I found there were problem in the egress side: packets could not be sent out, so finally there was a RmSendPacket error. Upon Erik's suggestion, I've checked some status of the board and found that when there were packets to send, the value of the XMIT_PTR regisger would stick to a fix number, and the active transmit thread would cycle around these codes: Egress_WaitForPortReady[abs_port] ctx_arb[sdram]I've tried long, but failed to resolve the problem. It's making me crazy. Who can give me a help?Regards.Irvin

4 4

RE: [ixp1200] SOS, a question about connection and output!
by Ye, Irvin 20 Nov '02

20 Nov '02

Hi, Austen, Your information is really helpful, but I think the MAC of the port is set by ixconfig, so it shouldn't cause problem. Cheers! Irvin -----Original Message----- From: Austen McDonald [mailto:austen@cc.gatech.edu] Sent: 18 November 2002 23:01 To: 'ixp1200(a)cs.princeton.edu' Subject: Re: [ixp1200] SOS, a question about connection and output! > I had tried to connect PCs and ixp1200 ports in standard way: via a hub / > switch, but it failed to work. When I tried direct connection through crossover So, hub != switch. Let me explain: I dont know anything about the Intel app but it could be that youre trying to send packets to the wrong eth addr---if you dont set one up on the IXPs (I have an app that will do that, bvmac, check out www.cercs.gatech.edu/projects/asan/software.html) AND send a packet (or by some means let the switch know about your eth addr) then the switch will probably just drop the packet. However, if youre using a hub, it probably wont drop it and youve probably got some other problem... Austen McDonald austen(a)cc.gatech.edu

1 0

Regarding to IXP1200 Ethernet Port configuration
by Ning Weng 19 Nov '02

19 Nov '02

Hi, I have one question here regarding to IXP1200 ENP2505. My case is I reboot my target(IXP1200) from my host machine through a serial cable, and I can ping each other via virtual interface( PCI) as well. Now I want the two system(IXP1200 and my host Linux machine) communicate by Ethernet port via a cross over cable. My question is : what is the detail steps that I should do, such as network configurations or other? That will be a great help for me if you would like share your experiences or point me the the documentation that you followed. Thanks a lot. Ning

1 0

Re: [ixp1200] waiting for memory writes
by Shyamal Pandya 18 Nov '02

18 Nov '02

Use the sram/sdram instruction with a sig_done optional token. This won't swap out the context. A wait loop can then be written using the br_!signal instruction. Refer the microcode programmer's reference for details. It even has an example. All this if you are using microcode. >From: Gregory Prier <greg(a)Lever.cs.ucla.edu> >Reply-To: ixp1200(a)CS.Princeton.EDU >To: ixp mailing list <ixp1200(a)CS.Princeton.EDU> >Subject: [ixp1200] waiting for memory writes >Date: Mon, 18 Nov 2002 15:31:26 -0800 (PST) > >Is there any way to get the uengine to wait for a memory write to complete >without swapping context? I can swap context and wait for a signal but >then >other threads have a chance to run, which I don't want to happen. I just >want >the processor to spin and waste cycles while the write completes. Yes, I >know >this is inefficient, but it would be a big help right now for debugging. > >Thanks for the help. >GKP > >---------------------------------------------------------------------- >Greg Prier Previously: A Mudder >3564 Boelter Hall Currently: A Bruin working >UCLA CA 90095 toward a PhD >(310) 825-8899 After that: No idea > >"Hell, there are no rules here. We're trying to accomplish something." > Thomas Edison _________________________________________________________________ Add photos to your messages with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail

1 0

waiting for memory writes
by Gregory Prier 18 Nov '02

18 Nov '02

Is there any way to get the uengine to wait for a memory write to complete without swapping context? I can swap context and wait for a signal but then other threads have a chance to run, which I don't want to happen. I just want the processor to spin and waste cycles while the write completes. Yes, I know this is inefficient, but it would be a big help right now for debugging. Thanks for the help. GKP ---------------------------------------------------------------------- Greg Prier Previously: A Mudder 3564 Boelter Hall Currently: A Bruin working UCLA CA 90095 toward a PhD (310) 825-8899 After that: No idea "Hell, there are no rules here. We're trying to accomplish something." Thomas Edison

1 0

SOS, a question about connection and output!
by Ye, Irvin 18 Nov '02

18 Nov '02

Hi, I'm puzzled on the connection for ixp1200. I had tried to connect PCs and ixp1200 ports in standard way: via a hub / switch, but it failed to work. When I tried direct connection through crossover twisted pairs, it seemed OK, packets could be successfully sent to ixp1200 ports. But then it found that packets could not be sent from the ports. When I tried Intel's simple bridge example, I found there were problem in the egress side: packets could not be sent out, so finally there was a RmSendPacket error. Upon Erik's suggestion, I've checked some status of the board and found that when there were packets to send, the value of the XMIT_PTR regisger would stick to a fix number, and the active transmit thread would cycle around these codes: Egress_WaitForPortReady[abs_port] ctx_arb[sdram] I've tried long, but failed to resolve the problem. It's making me crazy. Who can give me a help? Regards. Irvin

2 1

Prefetch and Non-Prefetch SDRAM
by Shyamal Pandya 18 Nov '02

18 Nov '02

Hi, Could someone tell me the distinction between prefetch and non-prefetch SDRAM on the IXP1200? And why does the resource manager provided with the IXA Ace framework use only the uncached SDRAM? While at it could someone tell me what is the cache flush area meant to be? Thanks! Shyamal Pandya. Arizona State University. _________________________________________________________________ Help STOP SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail

2 1

recycling NICs from IXP12EB's, gig NICs & multiple IXP12EB's in a PC (running VERA)
by Andrew Moore 18 Nov '02

18 Nov '02

Q1. I have vera working happily on an ixp12eb (rev C.), can the ethernet board that comes with the ix12eb kit be recycled into a linux box? I'm tempted to try, and did in fact do this, but the machine wasn't happy booting so I figured it could be either the ancient machine I was trying or something more fundamental in the ethernet nic that comes with the ixp12eb. Q2. I'm interested in talking to people that have used vera and the 1 Gig interfaces... anyone out there? I'm interested in using them for traffic generation, but would like to talk to anyone who has done more than the basic "turn it on" and "yep the physical layer is working" test. Q3 now this one might be a bit odd, however.... What problems are there with sticking several ixp12eb's into a machine? I assume in its current incarnation the vera drivers are not aware of several boards? (no I've not tried it) any ideas what the load is for an ixp12eb? is it going to 'cause the mobo pci supplies to be draining too much? This would be for a "lots of ports" traffic generator.... although a "lots of ports" might be good too. cheers, Andrew. Andrew Moore, Computer Laboratory, University of Cambridge, Ph: +44-1223-763446 Fax: +44-1223-334678 andrew.moore(a)cl.cam.ac.uk www.cl.cam.ac.uk/users/awm22/

2 1

princeton roms on a rev Bixp1200 Eth. Eval. Board.
by Andrew Moore 18 Nov '02

18 Nov '02

Hi, I've been attempting to get VERA up on a board with rev B arm core (at least its what I had always thought was a REV B - this may be the problem) Possibly there are many more versions of the IXP1200 Ethernet Evaluation Board than I realised and I may be stuck with one that will never fly... I've been following the "update the flash" instructions (http://www.cs.princeton.edu/nsg/vera/README.flash.txt) However everything is fine until futil, Instead of getting C:\>FUtil -c 1 -n -a 0 -o 0 -e 200000 -f FILE.bin Flash Utility, Version 1.2.76 FlashUtil, Version 1.2.76 CPU Revision: 6901C123 Flash Width: 32 Flash Bank Size: 2048 KBytes Flash Type Bank[0]: 28F800B3-B Flash Type Bank[1]: 28F160B3-T Flash Type Bank[2]: 28F160B3-T Flash Type Bank[3]: 28F160B3-T Sending data to remote system... Percent complete: 100 Updating flash... Percent Complete: 100 done I get C:\> FUtil -c 1 -n -a 0 -o 0 -e 200000 -f flashImage.bin Flash Utility, Version 2.01.112 FlashUtil, Version 1.1.109 CPU Revision: 6901C121 Flash Width: 32 Flash Bank Size: 2048 KBytes Flash Type Bank[0]: 28F800B3-B Flash Type Bank[1]: 28F160B3-T Flash Type Bank[2]: Unknown Flash Type Bank[3]: Unknown Sending data to remote system... Percent complete: 100 Error in doing inflate: 'Error FFFFFFFD C:\> Clearly this board doesn't have 2Mb of flash, or its not in the configuration expected. I've tried the same process on other IXP ether Eval boards and it works, but not this particular type. Details: the boards I have that work call themselves Rev C the board has the identifier A13033-001 A02 L1 (this is from the sth/wst corner of the board) (near serial port) the board that gives the error above identifies itself as a Rev B and has the identifier A13033-002 Is this board a waste of time??? Thanks for your thoughts, Andrew. Andrew Moore, Computer Laboratory, University of Cambridge, Ph: +44-1223-763446 Fax: +44-1223-334678 andrew.moore(a)cl.cam.ac.uk www.cl.cam.ac.uk/users/awm22/

6 8

IXP1200 - RHL
by Rayashish＠aol.com 15 Nov '02

15 Nov '02

Hi all, Can anyone help me to load Linux2.4 kernel on IXP12eb sitting on host pc running RHL2.4. --I have VERA drivers loaded on Linux Host. --Can connect to ixp-board through Hyper teminal. --Cygmon is running in flash. --Also Vera loader is running in flash which waits for start address of srec-image. How do i load arm-kernel and ramdisk image through host pci usinng vera tools? Any help or pointer- TIX --Ashish

2 1

RE: [ixp1200] IXP1200-2505 PCI card booting
by Brandewie, Dirk J 15 Nov '02

15 Nov '02

The problem here is that the card saw a space bar that told it to drop out of loop where it was waiting for the kernel to be downloaded and go to cygmon. The ixstart script waits to see the login prompt from the kernel booting on the board. The script times out if it does not see the login prompt. If you reboot the system and connect to the serial port that the board is connected to you should see the message: "Press space bar to abort host download mode...." repeated over and over. If this is the case you should be able to exit minicom and use the bootixp script to boot the board. --Dirk > -----Original Message----- > From: Ning Weng [mailto:nweng@ecs.umass.edu] > Sent: Wednesday, November 13, 2002 3:45 PM > To: ixp1200(a)CS.Princeton.EDU > Subject: [ixp1200] IXP1200-2505 PCI card booting > > > > Hello, > > Sorry for bother but I really need some kind inputs... > > Did I any one have this kind of experience? Plese help and give me > some hints. > > For IXP1200 ENP-2505 PCI card, following the IXA Education > Workstation > Setup Guide, I did successfully boot the card with the > message" Finished > ENP-2505 Driver Install..." . But after a few days, when I > reboot, this is > what I got : > > ".... > > Welcome to minicom 1.83.1 > > OPTIONS: History Buffer, F-key Macros, Search History Buffer, I18n > Compiled on Aug 28 2001, 15:09:33. > > Press CTRL-A Z for help on special keys > > > cygmon> script "./.ixstart": global timeout > > ...." > > I cann't remember what I exactly did after my first > successful booting. > But I try to reinstall the software and hardware, but I got the same > message. > > Any help is greatly appreciated!!!! > > Thanks a lot. > > -Ning >

2 2

microengine loader (w/correct URL)
by Kenneth M. Mackenzie 13 Nov '02

13 Nov '02

[doh -- always check the URLs] Hi all, Here's an updated package of our local tools, just updated to include a microcode loader, etc. that we use to operate the microengines from the StrongARM under Linux without ACEs or the workbench. http://www.cercs.gatech.edu/projects/asan/software.html Note that the StrongARM tools (in the "utils/" directory) don't depend on our alternative host-side driver scheme, I'm just packaging them in one tarball. ---------------- Various folks have asked about alternatives to ACEs. We use three schemes to program the microengines: 1. Homebrew tools including loader and etc., that run stand-alone under Linux on the StrongARM and communicate with the uEngines via Intel's "ue.o" kernel module on the StrongARM (same as used by rs_udebug with the workbench). We also use command-line versions of the assembler and linker in this configuration under "wine" so it's an all-linux solution. 2. Intel's Workbench. You can use this to load code onto the hardware instead of the simulator and it doesn't require ACEs (in fact it doesn't work with ACEs although it's not obvious that they should be incompatible). 3. ACEs Note: I would not characterize any of these three toolchains as exactly "easy". Each just offers its own kind of pain. In other words, don't switch toolchains just because you're frustrated with ACEs. -- Ken

2 1

VERA LibIXPRouter Minor Software Update
by Yitzchak M. Gottlieb 13 Nov '02

13 Nov '02

We are very happy to announce an update of the Vera IXPRouter library. This is a very minor update to correct small problems in compiling the IXPRouter library. Some relevant changes include: 1) Removed references to atoi in Libixprouter/sa/src/control.c (I applied the patch previously posted to the mailing list.) 2) Removed reference to p2tosa.h in Libixprouter/common/src/Makefile The software is available at http://www.cs.princeton.edu/nsg/vera. Please feel free to post any questions. Zuki -- Yitzchak Gottlieb zuki(a)CS.Princeton.EDU

1 0

RE: [ixp1200] Workbench, debug with "hardware" option?
by Botham, Mick 12 Nov '02

12 Nov '02

Irwin, Check the portmapper is running on the host, this error message results if the portmapper is not started. Rgds Mick -----Original Message----- From: Ye, Irvin [mailto:i.ye@lancaster.ac.uk] Sent: 08 November 2002 14:09 To: 'ixp1200(a)CS.Princeton.EDU' Subject: RE: [ixp1200] Workbench, debug with "hardware" option? I'm using Linux. -----Original Message----- From: Botham, Mick [mailto:mick.botham@intel.com] Sent: 08 November 2002 13:43 To: ixp1200(a)CS.Princeton.EDU Subject: RE: [ixp1200] Workbench, debug with "hardware" option? Irvin, Are you using VxWorks or Linux? The workbench will function the same on either EB or SI card. I prefer a direct network connection from host to target board, it saves a lot of confusion. Rgds Michael Botham EMEA IXA TME -----Original Message----- From: Ye, Irvin [mailto:i.ye@lancaster.ac.uk] Sent: 08 November 2002 11:46 To: 'ixp1200(a)cs.princeton.edu' Subject: [ixp1200] Workbench, debug with "hardware" option? Hi, I wonder, if I can use Workbench to do "hardware" debug? It's said that it's OK on EB boards, but I only have a SI board from Radisys. So I'm not sure whether it works. I've tried. Sometimes, it seemed successifully connected with the board, while in the other times, it just reported "Attach to debug library operation failed". It let me puzzled. Who can give some hints? BTW, to my opinion, play with ixp1200 is a game far from fun. There are enormous documents, while it's rather difficult to find wanted information. Cheers! Irvin

1 0

microengine loader
by Kenneth M. Mackenzie 11 Nov '02

11 Nov '02

Hi all, Here's an updated package of our local tools, just updated to include a microcode loader, etc. that we use to operate the microengines from the StrongARM under Linux without ACEs or the workbench. http://www.cercs.gatech.edu/projects/asan/software/ Note that the StrongARM tools (in the "utils/" directory) don't depend on our alternative host-side driver scheme, I'm just packaging them in one tarball. ---------------- Various folks have asked about alternatives to ACEs. We use three schemes to program the microengines: 1. Homebrew tools including loader and etc., that run stand-alone under Linux on the StrongARM and communicate with the uEngines via Intel's "ue.o" kernel module on the StrongARM (same as used by rs_udebug with the workbench). We also use command-line versions of the assembler and linker in this configuration under "wine" so it's an all-linux solution. 2. Intel's Workbench. You can use this to load code onto the hardware instead of the simulator and it doesn't require ACEs (in fact it doesn't work with ACEs although it's not obvious that they should be incompatible). 3. ACEs Note: I would not characterize any of these three toolchains as exactly "easy". Each just offers its own kind of pain. In other words, don't switch toolchains just because you're frustrated with ACEs. -- Ken

1 0

RmSendPacket error "Ring to microengine is full"
by Ye, Irvin 11 Nov '02

11 Nov '02

Hi, I use Intel's MicroACE. But there's a problem on egress side. The problem is between the core and micro block. When the core tried to send packets to micro block, there may be a RmSendPacket error, and the message is "Ring to microengine is full". Who can give me some help? Regards. Irvin

1 0

uengine interrupt
by Arthur Berkowitz 11 Nov '02

11 Nov '02

Excuse me in advance if this is too long. Let me explain what I'm trying to do: I'm using the uengine interrupt to StrongARM to ask for an ARP service performed by VxWorks. I'm not using the OS ARP Cache directly from uengines, rather I maintain a private hash table (SRAM and DRAM) for IP to MAC entries. This table is written by the core and read by uengines. All locking, hash unit and duplicates issues are just fine. When a lookup from uengine thread fails to find an entry for IP (the MAC is unknown!), the uengine thread will write the request in a mailbox and raise an interrupt to core. ACK from core is waited (inter thread signal). when ACK received, uthread will drop the packet, hopefully next packet required MAC will be resolved since ARP Service on core will add the information to IP2MAC table. On core there is an interrupt handler, that is a callback function registered with the Global uengine ISR (yeah, the same one breakpoint is using!). This handler will read the uthread request from mailbox and send it by message queue to ARP resolver task. The interrupt handler send an inter thread signal as an ACK back to uengines. The ARP resolver task perform the ARP and adds the resolved MAC to IP2MAC hash table. some code snippets: #macro arp_resolve_request(_output_intf, _ip) .local mailbox xbuf_alloc($request, 2) move($request0, _output_intf) move($request1, _ip) #if (UENGINE_ID == 0) move(mailbox, LWOFFSET0) #else // UENGINE_ID == 1 move(mailbox, LWOFFSET2) #endif sram_write($request0, UTHREAD_TO_CORE_MAILBOXES, mailbox, LWCOUNT2, sig_done) block_on_sram#: br_!signal[sram, block_on_sram#] ; block until mailbox written to continue fast_wr[1, ireg] ; raise interrupt block_on_inter_thread#: br_!signal[inter_thread, block_on_inter_thread#] ; wait for core ACK xbuf_free($request) alu[--,--,b, 0] ; this will cause drop pkt .endlocal // mailbox #endm note that I use only the first 2 (0, 1) uengines to Receive & Forward so just two mailboxes. now at the core... I connect the interrupt handler as: #include "ueng.h" ......... /* statics initialization */ unsigned int & ArpServ::_ireg(*(unsigned int *)0xB00401E0); ................. uEng_intConnect(UENG_MSG_INT, 0xFFFFFF, interrupt_handler, 0); ................ void ArpServ::interrupt_handler(int){ REQUEST_t request; unsigned int ifname, ipaddr, request_size; unsigned int iregShadow; iregShadow= _ireg; _ireg = iregShadow & 0xF8FFFFFF; iregShadow &= 0xFFFFFF; for(int bit=0; bit<FWD_THR_NUM; bit++){ if (iregShadow & (1 << bit)){ iregShadow = iregShadow & ~( 1<< bit); request_size = (bit >> 2) ? REQ_SIZE : 0; READ_MAILBOX(request_size, ifname, ipaddr); request.ifname = ifname; request.ipaddr = ipaddr; logMsg("request.ifname = %d request.ipaddr = 0x%08x\n", request.ifname, request.ipaddr, 3, 4, 5, 6); msgQSend(msgQId, (char *)&request, REQ_SIZE, 0, 0); /* ack uengine thread by inter thread signal */ *(unsigned int *)0xB00401C0 = bit; } } } At this point let me thank you for reading so far! Now the problem: 3 boards with same configuration work inconsistently. Board A always performs as expected. Board B and Board C not always. The interrupt isn't received on core. BUT, when I set breakpoint in uthread and stop as close before the fast_wr instruction (on sram_write) then just continue running interrupt is received on core! Then even if I unset the breakpoint the board function as expected The software looks OK and (surprisingly :-) it sometimes (always on one of the Boards) work as expected. Is it some timing issue with SRAM reference? Maybe I mask something wrong in ireg and not register proper to ISR? At the time of writing this an idea came to me: I should link out debug from vxworksnetapp so Breakpoint interrupt handler is not getting connected. Could this be a H/W problem?! Any ideas and all are welcomed ! Arthur Berkowitz COMGATES Ltd. 15 Hagalim Avenue. Herzliya, 46725 Israel Tel: +972.9.950.0404 Ext: 288 Fax: +972.9.950.0385 Mobile: +972.55.535.349 Email: arthur(a)comgates.co.il Visit Us: www.comgates.com

1 0

Boards without debug port.
by Ye, Irvin 11 Nov '02

11 Nov '02

Hi, It seems that most people work on ixp1200 boards with a debug ethernet port, which can be used to boot and then debug, with tools such as WorkBench, the board. While, ENP2505 from Radisys without such a port. It can be booted by the host machine via PCI. And when booted, a pair of psedo ethernet interfaces are setup also via PCI, for the communication between the board and the host. It's convenient and efficient. Yet, just wonder, is it still possible for it to be used in an unhosted environment? Is there any idea from those who also uses ENP2505? Regards. Irvin

1 0