fxp(4) device timeouts

Matthew Dillon dillon at apollo.backplane.com
Wed Dec 31 12:10:03 PST 2003

:I installed dragonfly a couple days ago using the dec 29 snapshot, and been
:having issues with fxp0 and fxp1 on my machine timing out. This problem only
:happens when I'm using two intel network cards (i've tried a combination of
:82550's, 82555's, 82558's and 82559's all yeild the same results, even when
:intermixed). If I remove the second nic, all is fine (except my NAT is screwed
:up). I'm running a GENERIC kernel with SMP enabled, below is the output of
:my dmesg
     Hmm.  I would suspect that it's an interrupt sharing issue.  The
     dmesg output from a boot with both cards installed would be useful,
     as would the output showing the timeout errors.

     But I don't know if I can fix it without getting two fxp cards and
     reproducing the problem locally.

					Matthew Dillon 
					<dillon at xxxxxxxxxxxxx>

:aic7880: Ultra Single Channel A, SCSI Id=2, 16/253 SCBs
:pci0: <S3 ViRGE DX/GX graphics accelerator> at 11.0 irq 10
:fxp0: <Intel 82550 Pro/100 Ethernet> port 0xf400-0xf43f mem 0xffbc0000-0xffbdfff
:f,0xffbec000-0xffbecfff irq 11 at device 14.0 on pci0
:installed MI handler for int 11
:fxp0: Ethernet address 00:02:b3:2b:18:cd
:inphy0: <i82555 10/100 media interface> on miibus0
:inphy0:  10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto

More information about the Bugs mailing list