Javi Roman wrote: > On 1/9/07, Jan Kiszka <[EMAIL PROTECTED]> wrote: >> Javi Roman wrote: >> > Hi all! >> > >> > I'm workig with this set: >> > linux-2.4.18 >> > RTAI 3.0 (rthal-5g-2.4.18.patch for PPC) >> > RTNET 0.7.1 >> >> /* no comment */ >> >> > >> > I'm getting a very strange behavior when I going to set up Rtnet: >> > >> > PPC host with IP: 10.10.1.1: >> > [ with the network wire unplugged from switch] >> > # lsmod >> > loopback-rt 1412 1 >> > mpc8xx_fec-rt 5398 1 >> > rtnet 45856 0 [loopback-rt mpc8xx_fec-rt] >> > rtai_rtdm 11424 0 [rtnet] >> > rtai_sem 18544 0 [mpc8xx_fec-rt rtnet] >> > rtai_up 44400 0 [mpc8xx_fec-rt rtnet rtai_rtdm >> rtai_sem] >> > rtai_malloc 5224 0 [rtai_rtdm rtai_sem rtai_up] >> > rtai_hal 22480 0 [loopback-rt mpc8xx_fec-rt rtnet >> > rtai_rtdm rtai_sem rtai_up rtai_malloc >> > # ./rtifconfig rtlo up 127.0.0.1 >> > # ./rtifconfig rteth0 up 10.10.1.1 >> > # ./rtroute solicit 10.10.1.2 dev rteth0 >> > fec_enet_start_xmit: data len is 42 >> > fec_enet_start_xmit: transmission start triggered >> > fec_enet_interrupt: interrupt handler >> > fec_enet_tx: TX path >> > RTnet: TXI: c669f040 c6531330 0 >> > fec_enet_tx: dev_kfree_rtskb, sk buffer freed! >> > >> > [These are my traces. Obiously the ARP cache is not correct because of >> > we are not connected to switch] >> > >> > # ./rtroute >> > Host Routing Table >> > Hash Destination HW Address Device >> > 00 0.0.0.0 00:00:00:00:00:00 rtlo >> > 01 255.255.255.1 FF:FF:FF:FF:FF:FF rteth0 >> > 01 10.10.1.1 00:00:00:00:00:00 rtlo >> > 01 127.0.0.1 00:00:00:00:00:00 rtlo >> > >> > [At this moment I connect the network wire to switch and try again] >> > >> > # ./rtroute solicit 10.10.1.2 dev rteth0 >> > fec_enet_start_xmit: data len is 42 >> > fec_enet_start_xmit: transmission start triggered >> > >> > [I repeat this command several times with the same behavior until I >> > obtain the output queue full] >> > >> > # ./rtroute solicit 10.10.1.2 dev rteth0 >> > fec_enet_start_xmit: data len is 42 >> > fec_enet_start_xmit: transmission start triggered >> > fec_enet_start_xmit: rtnetif_stop_queue! >> > # ./rtroute solicit 10.10.1.2 dev rteth0 >> > fec_enet_start_xmit: data len is 42 >> > RTnet: rteth0: tx queue full!. >> > hard_start_xmit returned 1 >> > >> > Please, anybody has an idea? >> >> Do packets arrive on the other side? Maybe you are not getting any IRQs >> handled with the FEC (or they are not correctly ack'ed - I recall such >> issues with PPC, dig in the RTnet history). In that case neither the >> sent skbs will be freed nor incoming replies will be noticed. >> >> Jan >> >> >> >> > > The packets doesn't arrive on the other side because the FEC only > raises the interrupt handler (fec_enet_interrupt) when the wire is > unplugged from NIC. The original fec driver from montavista worked > fine, so the problem is with the rtnet driver porting. I suppose so > the mpc8xx_fec-rt.c driver never worked for this rtnet version > (v0.7.1). I have to make to work this driver without change the > versions (kernel, RTAI, rtnet), so any help (clue) would be greatly > appreciated.
If you have a working fec driver, check for differences in the code. Also compare the old RTnet version with the latest one, if there are potentially relevant changes. And if some diff remains unclear, post your findings here. Assuming this is a generic RTnet problem with your board, we are certainly very interested in finding a fix and propagating it into the latest version. Jan
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
