[vox-tech] DFE-530TX+, cheap ethernet
vox-tech@lists.lugod.org
vox-tech@lists.lugod.org
Tue, 25 Jun 2002 09:47:16 -0400
This is not an endorsement of either Fry's or D-Link... I am not
recommending anyone patronize or boycott either business.
So a deal going on at Fry's... D-Link DFE-530TX+ 10/100 PCI Adapters
for $5 each. The price tag shows $15, but the register rings up $5,
I've tried arguing with them twice before about things like this and
its not worth the hassle because they don't adjust the price (lower on
the tags or higher on the receipt :).
There are two different boxes available a mostly white box and a
mostly blue box. The pictures are slightly different, and the chip
right next to the ethernet jack is a little different, but everything
else on the boards look the same.
Thinking I was going to pay $30, I only bought two. One of each box
(obviously).
I see three software drivers available for this card:
========
1 - downloadable from the D-Link web site
ftp://ftp.dlink.com/NIC/dfe530tx+/Driver/dfe530TX+_linux_driver_051801.zip
(D-Link's sites are behind a antiquated firewall so you will have
to disable tcp_ecn if you want to connect to them).
2 - from Donald Becker's web site
http://www.scyld.com/network/rtl8139.html
3 - in the 2.4.18 kernel tree
/usr/src/linux/drivers/net/8139too.c
1:
Probably works in a 2.2 kernel, it does not compile on 2.4.18.
2:
Donald's Driver requires a number of auxiliary files
(pci-scan.[ch] kernel_compat.h) which are also available from his
site. His driver is version v1.17 last updated 2002-01-28.
His directions for compiling these drivers do not produce loadable
modules on a 2.4.18 kernel. I can see dozens of other people having
this problem on the bug lists he maintains for this driver ...
but I'm still trying things.
He has a section that talks about 'How this driver differs from the
2.4 "8139too.c"', I'm not sure this document still applies:
- I can plainly see the copyright text has been corrected.
- The mmio read-after-write that he complains about I can't actually
see happening in the driver for regular use there is a debug switch
that will turn that behavior on, but it's off by default.
One opinion I did make is that the kernel 8139too.c driver appears to
be more readable source code, due to formating and comment blocks.
I'm not equipt to comment on the quality of either code base.
3:
The kernel driver compiles and works. Benchmarks to follow.
So if you are at Fry's and looking for some inexpensive cards you
may want to take some to the register and if they ring up at $5 buy a
dozen.
=============
DFE-530TX+ 10/100 Ethernet Adapter
lspci -vvn
00:10.0 Class 0200: 1186:1300 (rev 10)
Subsystem: 1186:1301
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B-
Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 64 (8000ns min, 16000ns max)
Interrupt: pin A routed to IRQ 16
Region 0: I/O ports at e400 [size=256]
Region 1: Memory at ffaffe00 (32-bit, non-prefetchable) [size=256]
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1+,D2+,D3hot+,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
lspci -vv
00:10.0 Ethernet controller: D-Link System Inc RTL8139 Ethernet (rev 10)
Subsystem: D-Link System Inc: Unknown device 1301
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B-
Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 64 (8000ns min, 16000ns max)
Interrupt: pin A routed to IRQ 16
Region 0: I/O ports at e400 [size=256]
Region 1: Memory at ffaffe00 (32-bit, non-prefetchable) [size=256]
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1+,D2+,D3hot+,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-