> > Any ideas on the Tx errors?
> > Are they Tx FIFO errors on the NIC?
> >
> > If so which one?
> >
> > I have this problem on 2.2.x kernels, running the Symbios scsi/ethernet
> > combo (PCNet32 driver), when running on 100Base lan, but none on ten. It
> > appears to be largely occuring when connecting to an HP-UX 11.x server
> > (CVS and NFS).
>
> Tx FIFO errors occur when the PCNet32 driver is sending a packet and cannot
> get it across the PCI bus fast enough. That typically occurs only if another
> device has a very high PCI latency. It occurs a lot with the ISA ones on
> ISA/PCI boards for similar reasons
are you sure? I had pretty similar problems with a NI6510SC ISA card (AM79C60)
in an old DX2/66 system whenever there is a bit high netload from this PC
(even using NFS with 1k block size was too fast and the card/driver locked up
pretty frequently).
now, recently I raised HZ for this PC from 100 to 1024 (for completely unrelated
reasons) and to my big surprise this fixed all those network hangs, now sustained
transfer and 8k NFS are no problem at all...
I don't know if this really can be compared and I have no idea why HZ=1024
results in much better networking than HZ=100 and I haven't spent any time
trying to figure out what's going on.
for referece, some ni65 driver output (using 2.2.5):
Apr 12 13:15:19 dach kernel: eth0: ni6510 EtherBlaster found at 0x300, Version 0
x30033000, DMA 3 (autodetected), IRQ 9 (autodetected).
Apr 12 13:16:10 dach kernel: nfs: server turtle not responding, still trying
Apr 12 13:16:13 dach kernel: eth0: xmitter timed out, try to restart!
Apr 12 13:16:13 dach kernel: 83 83 83 03
Apr 12 13:16:13 dach kernel: nfs: server turtle OK
Apr 12 13:16:14 dach kernel: eth0: RX or TX was offline -> restart
Apr 12 13:16:14 dach kernel: eth0: RX or TX was offline -> restart
Apr 12 13:16:26 dach kernel: nfs: server turtle not responding, still trying
Apr 12 13:16:28 dach kernel: eth0: xmitter timed out, try to restart!
Apr 12 13:16:28 dach kernel: 83 83 83 03
Apr 12 13:16:28 dach kernel: nfs: server turtle OK
Apr 12 13:24:12 dach kernel: eth0: RX or TX was offline -> restart
Apr 12 13:24:12 dach last message repeated 29 times
lance driver:
Apr 12 13:25:11 dach kernel: lance.c:v1.14ac 1998/11/20 dplatt@3do.com, becker@c
esdis.gsfc.nasa.gov
Apr 12 13:25:13 dach rwhod[143]: sendto(10.2.170.255): Network is unreachable
Apr 12 13:25:48 dach kernel: eth0: transmit timed out, status 004c, resetting.
Apr 12 13:25:48 dach kernel: Ring data dump: dirty_tx 56 cur_tx 72 (full) cur_r
x 55.
Apr 12 13:25:48 dach kernel: 8338c000 0608 0066 8334b800 0608 0066 8334b000 06
62 0000 83b40d62 0062 0000
Apr 12 13:26:08 dach kernel: eth0: transmit timed out, status 004c, resetting.
Apr 12 13:26:08 dach kernel: Ring data dump: dirty_tx 99 cur_tx 115 (full) cur_
rx 41.
Apr 12 13:26:08 dach kernel: 832db000 0608 0066 83caa000 0608 0066 83531800 06
Harald
--
All SCSI disks will from now on ___ _____
be required to send an email notice 0--,| /OOOOOOO\
24 hours prior to complete hardware failure! <_/ / /OOOOOOOOOOO\
\ \/OOOOOOOOOOOOOOO\
\ OOOOOOOOOOOOOOOOO|//
Harald Koenig, \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik // / \\ \
koenig@tat.physik.uni-tuebingen.de ^^^^^ ^^^^^
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/