1.\" Copyright (c) 1997, 1998 2.\" Bill Paul <wpaul@ctr.columbia.edu>. All rights reserved. 3.\" 4.\" Redistribution and use in source and binary forms, with or without 5.\" modification, are permitted provided that the following conditions 6.\" are met: 7.\" 1. Redistributions of source code must retain the above copyright 8.\" notice, this list of conditions and the following disclaimer. 9.\" 2. Redistributions in binary form must reproduce the above copyright 10.\" notice, this list of conditions and the following disclaimer in the 11.\" documentation and/or other materials provided with the distribution. 12.\" 3. All advertising materials mentioning features or use of this software 13.\" must display the following acknowledgement: 14.\" This product includes software developed by Bill Paul. 15.\" 4. Neither the name of the author nor the names of any co-contributors 16.\" may be used to endorse or promote products derived from this software 17.\" without specific prior written permission. 18.\" 19.\" THIS SOFTWARE IS PROVIDED BY Bill Paul AND CONTRIBUTORS ``AS IS'' AND 20.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22.\" ARE DISCLAIMED. IN NO EVENT SHALL Bill Paul OR THE VOICES IN HIS HEAD 23.\" BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 24.\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 25.\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 26.\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 27.\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 28.\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF 29.\" THE POSSIBILITY OF SUCH DAMAGE. 30.\" 31.\" $FreeBSD: src/share/man/man4/rl.4,v 1.15.2.8 2003/02/17 21:20:39 trhodes Exp $ 32.\" $DragonFly: src/share/man/man4/rl.4,v 1.8 2008/07/11 10:38:50 thomas Exp $ 33.\" 34.Dd November 4, 1998 35.Dt RL 4 36.Os 37.Sh NAME 38.Nm rl 39.Nd RealTek 8129/8139 fast ethernet device driver 40.Sh SYNOPSIS 41.Cd "device miibus" 42.Cd "device rl" 43.Sh DESCRIPTION 44The 45.Nm 46driver provides support for PCI ethernet adapters and embedded 47controllers based on the RealTek 8129 and 8139 fast ethernet controller 48chips. 49This includes the Allied Telesyn AT2550, Farallon NetLINE 10/100 PCI, 50Genius GF100TXR, 51NDC Communications NE100TX-E, OvisLink LEF-8129TX, OvisLink LEF-8139TX, 52Netronix Inc. EA-1210 NetEther 10/100, KTX-9130TX 10/100 Fast Ethernet, 53Encore ENL832-TX 10/100 M PCI, Longshine LCS-8038TX-R, the 54SMC EZ Card 10/100 PCI 1211-TX, and various other cheap adapters. 55It also supports the Accton EN1207D which has a 56chip labeled MPX5030 (or MPX5038) which appears to be a RealTek workalike. 57.Pp 58The RealTek controllers use bus master DMA but do not use a 59descriptor-based data transfer mechanism. 60The receiver uses a 61single fixed size ring buffer from which packets must be copied 62into mbufs. 63For transmission, there are only four outbound packet 64address registers which require all outgoing packets to be stored 65as contiguous buffers. 66Furthermore, outbound packet buffers must 67be longword aligned or else transmission will fail. 68.Pp 69The 8129 differs from the 8139 in that the 8139 has an internal 70PHY which is controlled through special direct access registers 71whereas the 8129 uses an external PHY via an MII bus. 72The 8139 73supports both 10 and 100Mbps speeds in either full or half duplex. 74The 8129 can support the same speeds and modes given an appropriate 75PHY chip. 76.Pp 77The 78.Nm 79driver supports the following media types: 80.Bl -tag -width xxxxxxxxxxxxxxxxxxxx 81.It autoselect 82Enable autoselection of the media type and options. 83This is only 84supported if the PHY chip attached to the RealTek controller 85supports NWAY autonegotiation. 86The user can manually override 87the autoselected mode by adding media options to the 88.Pa /etc/rc.conf 89file. 90.It 10baseT/UTP 91Set 10Mbps operation. 92The 93.Ar mediaopt 94option can also be used to select either 95.Ar full-duplex 96or 97.Ar half-duplex 98modes. 99.It 100baseTX 100Set 100Mbps (fast ethernet) operation. 101The 102.Ar mediaopt 103option can also be used to select either 104.Ar full-duplex 105or 106.Ar half-duplex 107modes. 108.El 109.Pp 110The 111.Nm 112driver supports the following media options: 113.Bl -tag -width xxxxxxxxxxxxxxxxxxxx 114.It full-duplex 115Force full duplex operation 116.It half-duplex 117Force half duplex operation. 118.El 119.Pp 120Note that the 100baseTX media type is only available if supported 121by the adapter. 122.Pp 123For more information on configuring this device, see 124.Xr ifconfig 8 . 125The 126.Nm 127driver supports 128.Xr polling 4 . 129.Sh DIAGNOSTICS 130.Bl -diag 131.It "rl%d: couldn't map memory" 132A fatal initialization error has occurred. 133.It "rl%d: couldn't map interrupt" 134A fatal initialization error has occurred. 135.It "rl%d: watchdog timeout" 136The device has stopped responding to the network, or there is a problem with 137the network connection (cable). 138.It "rl%d: no memory for rx list" 139The driver failed to allocate an mbuf for the receiver ring. 140.It "rl%d: no memory for tx list" 141The driver failed to allocate an mbuf for the transmitter ring when 142allocating a pad buffer or collapsing an mbuf chain into a cluster. 143.It "rl%d: chip is in D3 power state -- setting to D0" 144This message applies only to adapters which support power 145management. 146Some operating systems place the controller in low power 147mode when shutting down, and some PCI BIOSes fail to bring the chip 148out of this state before configuring it. 149The controller loses all of 150its PCI configuration in the D3 state, so if the BIOS does not set 151it back to full power mode in time, it won't be able to configure it 152correctly. 153The driver tries to detect this condition and bring 154the adapter back to the D0 (full power) state, but this may not be 155enough to return the driver to a fully operational condition. 156If 157you see this message at boot time and the driver fails to attach 158the device as a network interface, you will have to perform second 159warm boot to have the device properly configured. 160.Pp 161Note that this condition only occurs when warm booting from another 162operating system. 163If you power down your system prior to booting 164.Dx , 165the card should be configured correctly. 166.El 167.Sh SEE ALSO 168.Xr arp 4 , 169.Xr ifmedia 4 , 170.Xr miibus 4 , 171.Xr netintro 4 , 172.Xr ng_ether 4 , 173.Xr polling 4 , 174.Xr ifconfig 8 175.Rs 176.%B The RealTek 8129 and 8139 datasheets 177.%O ftp.realtek.com.tw:/lancard/data sheet 178.Re 179.Sh HISTORY 180The 181.Nm 182device driver first appeared in 183.Fx 3.0 . 184.Sh AUTHORS 185The 186.Nm 187driver was written by 188.An Bill Paul Aq wpaul@ctr.columbia.edu . 189.Sh BUGS 190Since outbound packets must be longword aligned, the transmit 191routine has to copy an unaligned packet into an mbuf cluster buffer 192before transmission. 193The driver abuses the fact that the cluster buffer 194pool is allocated at system startup time in a contiguous region starting 195at a page boundary. 196Since cluster buffers are 2048 bytes, they are 197longword aligned by definition. 198The driver probably should not be 199depending on this characteristic. 200.Pp 201The RealTek data sheets are of especially poor quality, 202and there is a lot of information missing 203particularly concerning the receiver operation. 204One particularly 205important fact that the data sheets fail to mention relates to the 206way in which the chip fills in the receive buffer. 207When an interrupt 208is posted to signal that a frame has been received, it is possible that 209another frame might be in the process of being copied into the receive 210buffer while the driver is busy handling the first one. 211If the driver 212manages to finish processing the first frame before the chip is done 213DMAing the rest of the next frame, the driver may attempt to process 214the next frame in the buffer before the chip has had a chance to finish 215DMAing all of it. 216.Pp 217The driver can check for an incomplete frame by inspecting the frame 218length in the header preceding the actual packet data: an incomplete 219frame will have the magic length of 0xFFF0. 220When the driver encounters 221this value, it knows that it has finished processing all currently 222available packets. 223Neither this magic value nor its significance are 224documented anywhere in the RealTek data sheets. 225