Thanks a lot, I am looking forward to test the binary --zoly
-----Original Message----- From: Garrett D'Amore [mailto:gdam...@opensolaris.org] Sent: Tuesday, January 13, 2009 5:43 PM To: Zoltan Farkas Cc: solaris driver Subject: Re: [driver-discuss] mxfe driver issues? You've given me plenty of information. I may have a test binary for you to try later this week. (I'm balancing a lot of different priorities right now... sorry...) If you need "immediate" relief, go pick yourself up a different NIC from your local electronics store. Devices with realtek gige on them (rtl8169) are relatively cheap. The Linksys LNE100TX (afe driver) should work fine too. -- Garrett Zoltan Farkas wrote: > It is a Intel Core I7 at 2.6 Ghz. > I am not using vlans > My system has 2 onboard GBE NICs that have no driver installed. > (BCM5784M) > > Anything else I can look up? > > prtdiag > System Configuration: DELL Inc. XPS 730X BIOS Configuration: DELL INC. > 1.0.0 11/05/2008 > > ==== Processor Sockets ==================================== > > Version Location Tag > -------------------------------- -------------------------- > Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz CPU 1 > > ==== Memory Device Sockets ================================ > > Type Status Set Device Locator Bank Locator > ----------- ------ --- ------------------- ---------------- > other in use 0 DIMM0 BANK0 > other in use 0 DIMM1 BANK1 > other in use 0 DIMM2 BANK2 > > ==== On-Board Devices ===================================== > > ==== Upgradeable Slots ==================================== > > ID Status Type Description > --- --------- ---------------- ---------------------------- > > > -----Original Message----- > From: Garrett D'Amore [mailto:gdam...@opensolaris.org] > Sent: Tuesday, January 13, 2009 2:03 PM > To: Zoltan Farkas > Cc: solaris driver > Subject: Re: [driver-discuss] mxfe driver issues? > > This is helpful. It suggests that the NIC is melting down due to FIFO > overflow errors. The RX fifo shouldn't fill up, but perhaps there is > a lot of bus contention on the system? > > Its pretty bad if the host can't keep up with a 100Mbps connection. > What type of system are you running this on? > > Its possible that I need to take more severe corrective action when I > notice this problem in the driver (such as resetting the chip) -- it > wouldn't surprise me if there was a silicon bug that left the chip in > some weird state when the overflow condition is noticed. Maybe the > chip is engaging some sort of flow control (pause frame) as well... > > What I don't understand is why snv105 is worse, unless snv105 is > somehow making the bus contention worse. (It would make sense if we > were dropping frames due to inability to retrieve them from the > descriptors quickly enough. But *that* shouldn't happen, and my > understanding is that the error status bit that is getting counted > only refers to a hardware FIFO overrun, not rx buffer exhaustion.) > > You're not using this with VLANs by any chance are you? > > -- Garrett > > Zoltan Farkas wrote: > >> here is the output: >> >> kstat mxfe >> module: mxfe instance: 0 >> name: fm class: misc >> acc_err 0 >> crtime 59.938337218 >> dma_err 0 >> erpt_dropped 0 >> fm_cache_full 0 >> fm_cache_grew 0 >> snaptime 7313.355485801 >> >> module: mxfe instance: 0 >> name: intr class: controller >> crtime 59.923994787 >> hard 57348 >> multiple_service 0 >> snaptime 7313.355683824 >> soft 0 >> spurious 68298 >> watchdog 0 >> >> module: mxfe instance: 0 >> name: mac class: net >> adv_cap_1000fdx 0 >> adv_cap_1000hdx 0 >> adv_cap_100fdx 1 >> adv_cap_100hdx 1 >> adv_cap_100T4 0 >> adv_cap_10fdx 1 >> adv_cap_10hdx 1 >> adv_cap_asmpause 0 >> adv_cap_autoneg 1 >> adv_cap_pause 0 >> adv_rem_fault 0 >> align_errors 0 >> brdcstrcv 13935 >> brdcstxmt 165 >> cap_1000fdx 0 >> cap_1000hdx 0 >> cap_100fdx 1 >> cap_100hdx 1 >> cap_100T4 0 >> cap_10fdx 1 >> cap_10hdx 1 >> cap_asmpause 0 >> cap_autoneg 1 >> cap_pause 0 >> cap_rem_fault 0 >> carrier_errors 1 >> collisions 0 >> crtime 59.938315688 >> defer_xmts 0 >> ex_collisions 0 >> fcs_errors 0 >> first_collisions 0 >> ierrors 8670 >> ifspeed 100000000 >> ipackets 86527 >> ipackets64 86527 >> jabber_errors 0 >> link_asmpause 0 >> link_autoneg 1 >> link_duplex 2 >> link_pause 0 >> link_state 1 >> link_up 1 >> lp_cap_1000fdx 0 >> lp_cap_1000hdx 0 >> lp_cap_100fdx 1 >> lp_cap_100hdx 1 >> lp_cap_100T4 0 >> lp_cap_10fdx 1 >> lp_cap_10hdx 1 >> lp_cap_asmpause 0 >> lp_cap_autoneg 1 >> lp_cap_pause 0 >> lp_rem_fault 0 >> macrcv_errors 0 >> macxmt_errors 0 >> multi_collisions 0 >> multircv 21752 >> multixmt 317 >> norcvbuf 0 >> noxmtbuf 0 >> obytes 16459926 >> obytes64 16459926 >> oerrors 1 >> oflo 8669 >> opackets 40681 >> opackets64 40681 >> promisc 1 >> rbytes 37938197 >> rbytes64 37938197 >> runt_errors 1 >> snaptime 7313.355823978 >> sqe_errors 0 >> toolong_errors 0 >> tx_late_collisions 0 >> uflo 0 >> unknowns 0 >> xcvr_addr 0 >> xcvr_id 0 >> xcvr_inuse 4 >> >> module: mxfe instance: 0 >> name: mxfe0 class: net >> brdcstrcv 13935 >> brdcstxmt 165 >> collisions 0 >> crtime 60.481214787 >> ierrors 8670 >> ifspeed 100000000 >> ipackets 86527 >> ipackets64 86527 >> link_duplex 2 >> link_state 1 >> multircv 21752 >> multixmt 317 >> norcvbuf 0 >> noxmtbuf 0 >> obytes 16459926 >> obytes64 16459926 >> oerrors 1 >> opackets 40681 >> opackets64 40681 >> rbytes 37938197 >> rbytes64 37938197 >> snaptime 7313.357017426 >> unknowns 16340 >> >> >> -----Original Message----- >> From: Garrett D'Amore [mailto:gdam...@opensolaris.org] >> Sent: Tuesday, January 13, 2009 1:19 PM >> To: Zoltan Farkas >> Cc: solaris driver >> Subject: Re: [driver-discuss] mxfe driver issues? >> >> Hmm... I guess I also want the statistics for the device. How about >> "kstat mxfe" output? >> >> -- Garrett >> >> Zoltan Farkas wrote: >> >> >>> there you go: >>> >>> dladm show-linkprop >>> LINK PROPERTY VALUE DEFAULT POSSIBLE >>> mxfe0 speed 100 100 -- >>> mxfe0 autopush -- -- -- >>> mxfe0 zone -- -- -- >>> mxfe0 duplex full full half,full >>> mxfe0 state up up up,down >>> mxfe0 adv_autoneg_cap 1 1 1,0 >>> mxfe0 mtu 1500 1500 -- >>> mxfe0 adv_100fdx_cap 1 1 1,0 >>> mxfe0 en_100fdx_cap 1 1 1,0 >>> mxfe0 adv_100hdx_cap 1 1 1,0 >>> mxfe0 en_100hdx_cap 1 1 1,0 >>> mxfe0 adv_10fdx_cap 1 1 1,0 >>> mxfe0 en_10fdx_cap 1 1 1,0 >>> mxfe0 adv_10hdx_cap 1 1 1,0 >>> mxfe0 en_10hdx_cap 1 1 1,0 >>> >>> >>> --zoly >>> >>> -----Original Message----- >>> From: Garrett D'Amore [mailto:gdam...@opensolaris.org] >>> Sent: Tuesday, January 13, 2009 12:21 PM >>> To: Zoltan Farkas >>> Cc: solaris driver >>> Subject: Re: [driver-discuss] mxfe driver issues? >>> >>> Hmm... that's strange. I'd have to take a look at it. I do wonder >>> why it happens more frequently on b105. b105 has Crossbow though, >>> and >>> >>> >> >> >>> there could be changes as a result of that. >>> >>> The fact that it happens on b101b suggests that the root problem is >>> *not* crossbow, though. >>> >>> This particular NIC has some ugliness surrounding its use of legacy >>> nway instead of more modern MII. >>> >>> Can you send me the output of dladm show-linkprop for this device? >>> >>> -- Garrett >>> >>> Zoltan Farkas wrote: >>> >>> >>> >>>> >>>> >>>> -----Original Message----- >>>> From: driver-discuss-boun...@opensolaris.org >>>> [mailto:driver-discuss-boun...@opensolaris.org] On Behalf Of Zoltan >>>> Farkas >>>> Sent: Tuesday, January 13, 2009 11:48 AM >>>> To: solaris driver >>>> Subject: Re: [driver-discuss] Is HP 2180 Deskjet printer supported >>>> onbuild105 >>>> >>>> >>>> I have a problem with my Network Card not receiving data after a >>>> >>>> >>>> >>> while. >>> >>> >>> >>>> It happens both in SNV_101b and SNV_105, in SNV_105 it happens a >>>> lot >>>> > > >>>> more often. >>>> >>>> functionality resumes after a unplumb/plumb >>>> >>>> What can I do to find out what is going on? >>>> >>>> >>>> Here are the details of my card: >>>> >>>> node name: pci15e8,8110 >>>> Vendor: Macronix, Inc. [MXIC] >>>> Device: MX987x5 >>>> binding name: pci10d9,531 >>>> devfs path: >>>> >>>> >> /p...@0,0/pci8086,2...@1e/pci15e8,8...@0 >> >> >>>> compatible name: >>>> (pci10d9,531.15e8.8110.25)(pci10d9,531.15e8.8110)(pci15e8,8110)(pci >>>> 1 >>>> 0 >>>> d >>>> 9, >>>> 531.25)(pci10d9,531)(pciclass,020000)(pciclass,0200) >>>> driver name: mxfe >>>> instance: 0 >>>> driver state: Attached >>>> fm-errcb-capable: TRUE >>>> model: Macronix MX98715AEC >>>> assigned-addresses: 81010010 >>>> reg: 10000 >>>> compatible: pci10d9,531.15e8.8110.25 >>>> model: Ethernet controller >>>> power-consumption: 1 >>>> fast-back-to-back: TRUE >>>> devsel-speed: 1 >>>> interrupts: 1 >>>> max-latency: 38 >>>> min-grant: 8 >>>> subsystem-vendor-id: 15e8 >>>> subsystem-id: 8110 >>>> unit-address: 0 >>>> class-code: 20000 >>>> revision-id: 25 >>>> vendor-id: 10d9 >>>> device-id: 531 >>>> >>>> >>>> >>>> _______________________________________________ >>>> driver-discuss mailing list >>>> driver-discuss@opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/driver-discuss >>>> >>>> >>>> _______________________________________________ >>>> driver-discuss mailing list >>>> driver-discuss@opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/driver-discuss >>>> >>>> >>>> >>>> >>> _______________________________________________ >>> driver-discuss mailing list >>> driver-discuss@opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/driver-discuss >>> >>> >>> >> >> >> > > > > _______________________________________________ driver-discuss mailing list driver-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/driver-discuss
