freebsd-dev/sys/pci/if_rlreg.h

1153 lines
35 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1997, 1998-2003
* Bill Paul <wpaul@ctr.columbia.edu>. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by Bill Paul.
* 4. Neither the name of the author nor the names of any co-contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY Bill Paul AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL Bill Paul OR THE VOICES IN HIS HEAD
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
* THE POSSIBILITY OF SUCH DAMAGE.
*
1999-08-28 01:08:13 +00:00
* $FreeBSD$
*/
/*
* RealTek 8129/8139 register offsets
*/
#define RL_IDR0 0x0000 /* ID register 0 (station addr) */
#define RL_IDR1 0x0001 /* Must use 32-bit accesses (?) */
#define RL_IDR2 0x0002
#define RL_IDR3 0x0003
#define RL_IDR4 0x0004
#define RL_IDR5 0x0005
/* 0006-0007 reserved */
#define RL_MAR0 0x0008 /* Multicast hash table */
#define RL_MAR1 0x0009
#define RL_MAR2 0x000A
#define RL_MAR3 0x000B
#define RL_MAR4 0x000C
#define RL_MAR5 0x000D
#define RL_MAR6 0x000E
#define RL_MAR7 0x000F
#define RL_TXSTAT0 0x0010 /* status of TX descriptor 0 */
#define RL_TXSTAT1 0x0014 /* status of TX descriptor 1 */
#define RL_TXSTAT2 0x0018 /* status of TX descriptor 2 */
#define RL_TXSTAT3 0x001C /* status of TX descriptor 3 */
#define RL_TXADDR0 0x0020 /* address of TX descriptor 0 */
#define RL_TXADDR1 0x0024 /* address of TX descriptor 1 */
#define RL_TXADDR2 0x0028 /* address of TX descriptor 2 */
#define RL_TXADDR3 0x002C /* address of TX descriptor 3 */
#define RL_RXADDR 0x0030 /* RX ring start address */
#define RL_RX_EARLY_BYTES 0x0034 /* RX early byte count */
#define RL_RX_EARLY_STAT 0x0036 /* RX early status */
#define RL_COMMAND 0x0037 /* command register */
#define RL_CURRXADDR 0x0038 /* current address of packet read */
#define RL_CURRXBUF 0x003A /* current RX buffer address */
#define RL_IMR 0x003C /* interrupt mask register */
#define RL_ISR 0x003E /* interrupt status register */
#define RL_TXCFG 0x0040 /* transmit config */
#define RL_RXCFG 0x0044 /* receive config */
#define RL_TIMERCNT 0x0048 /* timer count register */
#define RL_MISSEDPKT 0x004C /* missed packet counter */
#define RL_EECMD 0x0050 /* EEPROM command register */
/* RTL8139/RTL8139C+ only */
#define RL_8139_CFG0 0x0051 /* config register #0 */
#define RL_8139_CFG1 0x0052 /* config register #1 */
#define RL_8139_CFG3 0x0059 /* config register #3 */
#define RL_8139_CFG4 0x005A /* config register #4 */
#define RL_8139_CFG5 0x00D8 /* config register #5 */
#define RL_CFG0 0x0051 /* config register #0 */
#define RL_CFG1 0x0052 /* config register #1 */
#define RL_CFG2 0x0053 /* config register #2 */
#define RL_CFG3 0x0054 /* config register #3 */
#define RL_CFG4 0x0055 /* config register #4 */
#define RL_CFG5 0x0056 /* config register #5 */
/* 0057 reserved */
#define RL_MEDIASTAT 0x0058 /* media status register (8139) */
/* 0059-005A reserved */
#define RL_MII 0x005A /* 8129 chip only */
#define RL_HALTCLK 0x005B
#define RL_MULTIINTR 0x005C /* multiple interrupt */
#define RL_PCIREV 0x005E /* PCI revision value */
/* 005F reserved */
#define RL_TXSTAT_ALL 0x0060 /* TX status of all descriptors */
/* Direct PHY access registers only available on 8139 */
#define RL_BMCR 0x0062 /* PHY basic mode control */
#define RL_BMSR 0x0064 /* PHY basic mode status */
#define RL_ANAR 0x0066 /* PHY autoneg advert */
#define RL_LPAR 0x0068 /* PHY link partner ability */
#define RL_ANER 0x006A /* PHY autoneg expansion */
#define RL_DISCCNT 0x006C /* disconnect counter */
#define RL_FALSECAR 0x006E /* false carrier counter */
#define RL_NWAYTST 0x0070 /* NWAY test register */
#define RL_RX_ER 0x0072 /* RX_ER counter */
#define RL_CSCFG 0x0074 /* CS configuration register */
/*
* When operating in special C+ mode, some of the registers in an
* 8139C+ chip have different definitions. These are also used for
* the 8169 gigE chip.
*/
#define RL_DUMPSTATS_LO 0x0010 /* counter dump command register */
#define RL_DUMPSTATS_HI 0x0014 /* counter dump command register */
#define RL_TXLIST_ADDR_LO 0x0020 /* 64 bits, 256 byte alignment */
#define RL_TXLIST_ADDR_HI 0x0024 /* 64 bits, 256 byte alignment */
#define RL_TXLIST_ADDR_HPRIO_LO 0x0028 /* 64 bits, 256 byte alignment */
#define RL_TXLIST_ADDR_HPRIO_HI 0x002C /* 64 bits, 256 byte alignment */
#define RL_CFG2 0x0053
#define RL_TIMERINT 0x0054 /* interrupt on timer expire */
#define RL_TXSTART 0x00D9 /* 8 bits */
#define RL_CPLUS_CMD 0x00E0 /* 16 bits */
#define RL_RXLIST_ADDR_LO 0x00E4 /* 64 bits, 256 byte alignment */
#define RL_RXLIST_ADDR_HI 0x00E8 /* 64 bits, 256 byte alignment */
#define RL_EARLY_TX_THRESH 0x00EC /* 8 bits */
/*
* Registers specific to the 8169 gigE chip
*/
#define RL_GTXSTART 0x0038 /* 8 bits */
#define RL_TIMERINT_8169 0x0058 /* different offset than 8139 */
#define RL_PHYAR 0x0060
#define RL_TBICSR 0x0064
#define RL_TBI_ANAR 0x0068
#define RL_TBI_LPAR 0x006A
#define RL_GMEDIASTAT 0x006C /* 8 bits */
#define RL_MACDBG 0x006D /* 8 bits, 8168C SPIN2 only */
#define RL_GPIO 0x006E /* 8 bits, 8168C SPIN2 only */
#define RL_PMCH 0x006F /* 8 bits */
#define RL_MAXRXPKTLEN 0x00DA /* 16 bits, chip multiplies by 8 */
#define RL_INTRMOD 0x00E2 /* 16 bits */
/*
* TX config register bits
*/
#define RL_TXCFG_CLRABRT 0x00000001 /* retransmit aborted pkt */
#define RL_TXCFG_MAXDMA 0x00000700 /* max DMA burst size */
#define RL_TXCFG_QUEUE_EMPTY 0x00000800 /* 8168E-VL or higher */
#define RL_TXCFG_CRCAPPEND 0x00010000 /* CRC append (0 = yes) */
#define RL_TXCFG_LOOPBKTST 0x00060000 /* loopback test */
#define RL_TXCFG_IFG2 0x00080000 /* 8169 only */
#define RL_TXCFG_IFG 0x03000000 /* interframe gap */
#define RL_TXCFG_HWREV 0x7CC00000
#define RL_LOOPTEST_OFF 0x00000000
#define RL_LOOPTEST_ON 0x00020000
#define RL_LOOPTEST_ON_CPLUS 0x00060000
Take the support for the 8139C+/8169/8169S/8110S chips out of the rl(4) driver and put it in a new re(4) driver. The re(4) driver shares the if_rlreg.h file with rl(4) but is a separate module. (Ultimately I may change this. For now, it's convenient.) rl(4) has been modified so that it will never attach to an 8139C+ chip, leaving it to re(4) instead. Only re(4) has the PCI IDs to match the 8169/8169S/8110S gigE chips. if_re.c contains the same basic code that was originally bolted onto if_rl.c, with the following updates: - Added support for jumbo frames. Currently, there seems to be a limit of approximately 6200 bytes for jumbo frames on transmit. (This was determined via experimentation.) The 8169S/8110S chips apparently are limited to 7.5K frames on transmit. This may require some more work, though the framework to handle jumbo frames on RX is in place: the re_rxeof() routine will gather up frames than span multiple 2K clusters into a single mbuf list. - Fixed bug in re_txeof(): if we reap some of the TX buffers, but there are still some pending, re-arm the timer before exiting re_txeof() so that another timeout interrupt will be generated, just in case re_start() doesn't do it for us. - Handle the 'link state changed' interrupt - Fix a detach bug. If re(4) is loaded as a module, and you do tcpdump -i re0, then you do 'kldunload if_re,' the system will panic after a few seconds. This happens because ether_ifdetach() ends up calling the BPF detach code, which notices the interface is in promiscuous mode and tries to switch promisc mode off while detaching the BPF listner. This ultimately results in a call to re_ioctl() (due to SIOCSIFFLAGS), which in turn calls re_init() to handle the IFF_PROMISC flag change. Unfortunately, calling re_init() here turns the chip back on and restarts the 1-second timeout loop that drives re_tick(). By the time the timeout fires, if_re.ko has been unloaded, which results in a call to invalid code and blows up the system. To fix this, I cleared the IFF_UP flag before calling ether_ifdetach(), which stops the ioctl routine from trying to reset the chip. - Modified comments in re_rxeof() relating to the difference in RX descriptor status bit layout between the 8139C+ and the gigE chips. The layout is different because the frame length field was expanded from 12 bits to 13, and they got rid of one of the status bits to make room. - Add diagnostic code (re_diag()) to test for the case where a user has installed a broken 32-bit 8169 PCI NIC in a 64-bit slot. Some NICs have the REQ64# and ACK64# lines connected even though the board is 32-bit only (in this case, they should be pulled high). This fools the chip into doing 64-bit DMA transfers even though there is no 64-bit data path. To detect this, re_diag() puts the chip into digital loopback mode and sets the receiver to promiscuous mode, then initiates a single 64-byte packet transmission. The frame is echoed back to the host, and if the frame contents are intact, we know DMA is working correctly, otherwise we complain loudly on the console and abort the device attach. (At the moment, I don't know of any way to work around the problem other than physically modifying the board, so until/unless I can think of a software workaround, this will have do to.) - Created re(4) man page - Modified rlphy.c to allow re(4) to attach as well as rl(4). Note that this code works for the sample 8169/Marvell 88E1000 NIC that I have, but probably won't work for the 8169S/8110S chips. RealTek has sent me some sample NICs, but they haven't arrived yet. I will probably need to add an rlgphy driver to handle the on-board PHY in the 8169S/8110S (it needs special DSP initialization).
2003-09-08 02:11:25 +00:00
/* Known revision codes. */
#define RL_HWREV_8169 0x00000000
#define RL_HWREV_8169S 0x00800000
#define RL_HWREV_8110S 0x04000000
#define RL_HWREV_8169_8110SB 0x10000000
#define RL_HWREV_8169_8110SC 0x18000000
#define RL_HWREV_8401E 0x24000000
#define RL_HWREV_8102EL 0x24800000
#define RL_HWREV_8102EL_SPIN1 0x24C00000
#define RL_HWREV_8168D 0x28000000
#define RL_HWREV_8168DP 0x28800000
#define RL_HWREV_8168E 0x2C000000
#define RL_HWREV_8168E_VL 0x2C800000
#define RL_HWREV_8168B_SPIN1 0x30000000
#define RL_HWREV_8100E 0x30800000
#define RL_HWREV_8101E 0x34000000
#define RL_HWREV_8102E 0x34800000
#define RL_HWREV_8103E 0x34C00000
#define RL_HWREV_8168B_SPIN2 0x38000000
#define RL_HWREV_8168B_SPIN3 0x38400000
#define RL_HWREV_8168C 0x3C000000
#define RL_HWREV_8168C_SPIN2 0x3C400000
#define RL_HWREV_8168CP 0x3C800000
#define RL_HWREV_8105E 0x40800000
#define RL_HWREV_8105E_SPIN1 0x40C00000
#define RL_HWREV_8402 0x44000000
#define RL_HWREV_8168F 0x48000000
#define RL_HWREV_8411 0x48800000
#define RL_HWREV_8139 0x60000000
#define RL_HWREV_8139A 0x70000000
#define RL_HWREV_8139AG 0x70800000
#define RL_HWREV_8139B 0x78000000
#define RL_HWREV_8130 0x7C000000
#define RL_HWREV_8139C 0x74000000
#define RL_HWREV_8139D 0x74400000
#define RL_HWREV_8139CPLUS 0x74800000
#define RL_HWREV_8101 0x74C00000
#define RL_HWREV_8100 0x78800000
#define RL_HWREV_8169_8110SBL 0x7CC00000
#define RL_HWREV_8169_8110SCE 0x98000000
#define RL_TXDMA_16BYTES 0x00000000
#define RL_TXDMA_32BYTES 0x00000100
#define RL_TXDMA_64BYTES 0x00000200
#define RL_TXDMA_128BYTES 0x00000300
#define RL_TXDMA_256BYTES 0x00000400
#define RL_TXDMA_512BYTES 0x00000500
#define RL_TXDMA_1024BYTES 0x00000600
#define RL_TXDMA_2048BYTES 0x00000700
/*
* Transmit descriptor status register bits.
*/
#define RL_TXSTAT_LENMASK 0x00001FFF
#define RL_TXSTAT_OWN 0x00002000
#define RL_TXSTAT_TX_UNDERRUN 0x00004000
#define RL_TXSTAT_TX_OK 0x00008000
#define RL_TXSTAT_EARLY_THRESH 0x003F0000
#define RL_TXSTAT_COLLCNT 0x0F000000
#define RL_TXSTAT_CARR_HBEAT 0x10000000
#define RL_TXSTAT_OUTOFWIN 0x20000000
#define RL_TXSTAT_TXABRT 0x40000000
#define RL_TXSTAT_CARRLOSS 0x80000000
/*
* Interrupt status register bits.
*/
#define RL_ISR_RX_OK 0x0001
#define RL_ISR_RX_ERR 0x0002
#define RL_ISR_TX_OK 0x0004
#define RL_ISR_TX_ERR 0x0008
#define RL_ISR_RX_OVERRUN 0x0010
#define RL_ISR_PKT_UNDERRUN 0x0020
#define RL_ISR_LINKCHG 0x0020 /* 8169 only */
#define RL_ISR_FIFO_OFLOW 0x0040 /* 8139 only */
#define RL_ISR_TX_DESC_UNAVAIL 0x0080 /* C+ only */
#define RL_ISR_SWI 0x0100 /* C+ only */
#define RL_ISR_CABLE_LEN_CHGD 0x2000
#define RL_ISR_PCS_TIMEOUT 0x4000 /* 8129 only */
#define RL_ISR_TIMEOUT_EXPIRED 0x4000
#define RL_ISR_SYSTEM_ERR 0x8000
#define RL_INTRS \
(RL_ISR_TX_OK|RL_ISR_RX_OK|RL_ISR_RX_ERR|RL_ISR_TX_ERR| \
RL_ISR_RX_OVERRUN|RL_ISR_PKT_UNDERRUN|RL_ISR_FIFO_OFLOW| \
RL_ISR_PCS_TIMEOUT|RL_ISR_SYSTEM_ERR)
#ifdef RE_TX_MODERATION
#define RL_INTRS_CPLUS \
Take the support for the 8139C+/8169/8169S/8110S chips out of the rl(4) driver and put it in a new re(4) driver. The re(4) driver shares the if_rlreg.h file with rl(4) but is a separate module. (Ultimately I may change this. For now, it's convenient.) rl(4) has been modified so that it will never attach to an 8139C+ chip, leaving it to re(4) instead. Only re(4) has the PCI IDs to match the 8169/8169S/8110S gigE chips. if_re.c contains the same basic code that was originally bolted onto if_rl.c, with the following updates: - Added support for jumbo frames. Currently, there seems to be a limit of approximately 6200 bytes for jumbo frames on transmit. (This was determined via experimentation.) The 8169S/8110S chips apparently are limited to 7.5K frames on transmit. This may require some more work, though the framework to handle jumbo frames on RX is in place: the re_rxeof() routine will gather up frames than span multiple 2K clusters into a single mbuf list. - Fixed bug in re_txeof(): if we reap some of the TX buffers, but there are still some pending, re-arm the timer before exiting re_txeof() so that another timeout interrupt will be generated, just in case re_start() doesn't do it for us. - Handle the 'link state changed' interrupt - Fix a detach bug. If re(4) is loaded as a module, and you do tcpdump -i re0, then you do 'kldunload if_re,' the system will panic after a few seconds. This happens because ether_ifdetach() ends up calling the BPF detach code, which notices the interface is in promiscuous mode and tries to switch promisc mode off while detaching the BPF listner. This ultimately results in a call to re_ioctl() (due to SIOCSIFFLAGS), which in turn calls re_init() to handle the IFF_PROMISC flag change. Unfortunately, calling re_init() here turns the chip back on and restarts the 1-second timeout loop that drives re_tick(). By the time the timeout fires, if_re.ko has been unloaded, which results in a call to invalid code and blows up the system. To fix this, I cleared the IFF_UP flag before calling ether_ifdetach(), which stops the ioctl routine from trying to reset the chip. - Modified comments in re_rxeof() relating to the difference in RX descriptor status bit layout between the 8139C+ and the gigE chips. The layout is different because the frame length field was expanded from 12 bits to 13, and they got rid of one of the status bits to make room. - Add diagnostic code (re_diag()) to test for the case where a user has installed a broken 32-bit 8169 PCI NIC in a 64-bit slot. Some NICs have the REQ64# and ACK64# lines connected even though the board is 32-bit only (in this case, they should be pulled high). This fools the chip into doing 64-bit DMA transfers even though there is no 64-bit data path. To detect this, re_diag() puts the chip into digital loopback mode and sets the receiver to promiscuous mode, then initiates a single 64-byte packet transmission. The frame is echoed back to the host, and if the frame contents are intact, we know DMA is working correctly, otherwise we complain loudly on the console and abort the device attach. (At the moment, I don't know of any way to work around the problem other than physically modifying the board, so until/unless I can think of a software workaround, this will have do to.) - Created re(4) man page - Modified rlphy.c to allow re(4) to attach as well as rl(4). Note that this code works for the sample 8169/Marvell 88E1000 NIC that I have, but probably won't work for the 8169S/8110S chips. RealTek has sent me some sample NICs, but they haven't arrived yet. I will probably need to add an rlgphy driver to handle the on-board PHY in the 8169S/8110S (it needs special DSP initialization).
2003-09-08 02:11:25 +00:00
(RL_ISR_RX_OK|RL_ISR_RX_ERR|RL_ISR_TX_ERR| \
RL_ISR_RX_OVERRUN|RL_ISR_PKT_UNDERRUN|RL_ISR_FIFO_OFLOW| \
RL_ISR_PCS_TIMEOUT|RL_ISR_SYSTEM_ERR|RL_ISR_TIMEOUT_EXPIRED)
#else
#define RL_INTRS_CPLUS \
(RL_ISR_RX_OK|RL_ISR_RX_ERR|RL_ISR_TX_ERR|RL_ISR_TX_OK| \
RL_ISR_RX_OVERRUN|RL_ISR_PKT_UNDERRUN|RL_ISR_FIFO_OFLOW| \
RL_ISR_PCS_TIMEOUT|RL_ISR_SYSTEM_ERR|RL_ISR_TIMEOUT_EXPIRED)
#endif
/*
* Media status register. (8139 only)
*/
#define RL_MEDIASTAT_RXPAUSE 0x01
#define RL_MEDIASTAT_TXPAUSE 0x02
#define RL_MEDIASTAT_LINK 0x04
#define RL_MEDIASTAT_SPEED10 0x08
#define RL_MEDIASTAT_RXFLOWCTL 0x40 /* duplex mode */
#define RL_MEDIASTAT_TXFLOWCTL 0x80 /* duplex mode */
/*
* Receive config register.
*/
#define RL_RXCFG_RX_ALLPHYS 0x00000001 /* accept all nodes */
#define RL_RXCFG_RX_INDIV 0x00000002 /* match filter */
#define RL_RXCFG_RX_MULTI 0x00000004 /* accept all multicast */
#define RL_RXCFG_RX_BROAD 0x00000008 /* accept all broadcast */
#define RL_RXCFG_RX_RUNT 0x00000010
#define RL_RXCFG_RX_ERRPKT 0x00000020
#define RL_RXCFG_WRAP 0x00000080
#define RL_RXCFG_MAXDMA 0x00000700
#define RL_RXCFG_BUFSZ 0x00001800
#define RL_RXCFG_FIFOTHRESH 0x0000E000
#define RL_RXCFG_EARLYTHRESH 0x07000000
#define RL_RXDMA_16BYTES 0x00000000
#define RL_RXDMA_32BYTES 0x00000100
#define RL_RXDMA_64BYTES 0x00000200
#define RL_RXDMA_128BYTES 0x00000300
#define RL_RXDMA_256BYTES 0x00000400
#define RL_RXDMA_512BYTES 0x00000500
#define RL_RXDMA_1024BYTES 0x00000600
#define RL_RXDMA_UNLIMITED 0x00000700
#define RL_RXBUF_8 0x00000000
#define RL_RXBUF_16 0x00000800
#define RL_RXBUF_32 0x00001000
#define RL_RXBUF_64 0x00001800
#define RL_RXFIFO_16BYTES 0x00000000
#define RL_RXFIFO_32BYTES 0x00002000
#define RL_RXFIFO_64BYTES 0x00004000
#define RL_RXFIFO_128BYTES 0x00006000
#define RL_RXFIFO_256BYTES 0x00008000
#define RL_RXFIFO_512BYTES 0x0000A000
#define RL_RXFIFO_1024BYTES 0x0000C000
#define RL_RXFIFO_NOTHRESH 0x0000E000
/*
* Bits in RX status header (included with RX'ed packet
* in ring buffer).
*/
#define RL_RXSTAT_RXOK 0x00000001
#define RL_RXSTAT_ALIGNERR 0x00000002
#define RL_RXSTAT_CRCERR 0x00000004
#define RL_RXSTAT_GIANT 0x00000008
#define RL_RXSTAT_RUNT 0x00000010
#define RL_RXSTAT_BADSYM 0x00000020
#define RL_RXSTAT_BROAD 0x00002000
#define RL_RXSTAT_INDIV 0x00004000
#define RL_RXSTAT_MULTI 0x00008000
#define RL_RXSTAT_LENMASK 0xFFFF0000
#define RL_RXSTAT_UNFINISHED 0xFFF0 /* DMA still in progress */
/*
* Command register.
*/
#define RL_CMD_EMPTY_RXBUF 0x0001
#define RL_CMD_TX_ENB 0x0004
#define RL_CMD_RX_ENB 0x0008
#define RL_CMD_RESET 0x0010
#define RL_CMD_STOPREQ 0x0080
/*
* Twister register values. These are completely undocumented and derived
* from public sources.
*/
#define RL_CSCFG_LINK_OK 0x0400
#define RL_CSCFG_CHANGE 0x0800
#define RL_CSCFG_STATUS 0xf000
#define RL_CSCFG_ROW3 0x7000
#define RL_CSCFG_ROW2 0x3000
#define RL_CSCFG_ROW1 0x1000
#define RL_CSCFG_LINK_DOWN_OFF_CMD 0x03c0
#define RL_CSCFG_LINK_DOWN_CMD 0xf3c0
#define RL_NWAYTST_RESET 0
#define RL_NWAYTST_CBL_TEST 0x20
#define RL_PARA78 0x78
#define RL_PARA78_DEF 0x78fa8388
#define RL_PARA7C 0x7C
#define RL_PARA7C_DEF 0xcb38de43
#define RL_PARA7C_RETUNE 0xfb38de03
/*
* EEPROM control register
*/
#define RL_EE_DATAOUT 0x01 /* Data out */
#define RL_EE_DATAIN 0x02 /* Data in */
#define RL_EE_CLK 0x04 /* clock */
#define RL_EE_SEL 0x08 /* chip select */
#define RL_EE_MODE (0x40|0x80)
#define RL_EEMODE_OFF 0x00
#define RL_EEMODE_AUTOLOAD 0x40
#define RL_EEMODE_PROGRAM 0x80
#define RL_EEMODE_WRITECFG (0x80|0x40)
/* 9346 EEPROM commands */
#define RL_9346_ADDR_LEN 6 /* 93C46 1K: 128x16 */
#define RL_9356_ADDR_LEN 8 /* 93C56 2K: 256x16 */
#define RL_9346_WRITE 0x5
#define RL_9346_READ 0x6
#define RL_9346_ERASE 0x7
#define RL_9346_EWEN 0x4
#define RL_9346_EWEN_ADDR 0x30
#define RL_9456_EWDS 0x4
#define RL_9346_EWDS_ADDR 0x00
#define RL_EECMD_WRITE 0x140
#define RL_EECMD_READ_6BIT 0x180
#define RL_EECMD_READ_8BIT 0x600
#define RL_EECMD_ERASE 0x1c0
#define RL_EE_ID 0x00
#define RL_EE_PCI_VID 0x01
#define RL_EE_PCI_DID 0x02
/* Location of station address inside EEPROM */
#define RL_EE_EADDR 0x07
/*
* MII register (8129 only)
*/
#define RL_MII_CLK 0x01
#define RL_MII_DATAIN 0x02
#define RL_MII_DATAOUT 0x04
#define RL_MII_DIR 0x80 /* 0 == input, 1 == output */
/*
* Config 0 register
*/
#define RL_CFG0_ROM0 0x01
#define RL_CFG0_ROM1 0x02
#define RL_CFG0_ROM2 0x04
#define RL_CFG0_PL0 0x08
#define RL_CFG0_PL1 0x10
#define RL_CFG0_10MBPS 0x20 /* 10 Mbps internal mode */
#define RL_CFG0_PCS 0x40
#define RL_CFG0_SCR 0x80
/*
* Config 1 register
*/
#define RL_CFG1_PWRDWN 0x01
2010-11-08 22:05:11 +00:00
#define RL_CFG1_PME 0x01
#define RL_CFG1_SLEEP 0x02
#define RL_CFG1_VPDEN 0x02
#define RL_CFG1_IOMAP 0x04
#define RL_CFG1_MEMMAP 0x08
#define RL_CFG1_RSVD 0x10
#define RL_CFG1_LWACT 0x10
#define RL_CFG1_DRVLOAD 0x20
#define RL_CFG1_LED0 0x40
#define RL_CFG1_FULLDUPLEX 0x40 /* 8129 only */
#define RL_CFG1_LED1 0x80
/*
* Config 2 register
*/
#define RL_CFG2_PCI33MHZ 0x00
#define RL_CFG2_PCI66MHZ 0x01
#define RL_CFG2_PCI64BIT 0x08
#define RL_CFG2_AUXPWR 0x10
#define RL_CFG2_MSI 0x20
/*
* Config 3 register
*/
#define RL_CFG3_GRANTSEL 0x80
#define RL_CFG3_WOL_MAGIC 0x20
#define RL_CFG3_WOL_LINK 0x10
#define RL_CFG3_JUMBO_EN0 0x04 /* RTL8168C or later. */
#define RL_CFG3_FAST_B2B 0x01
/*
* Config 4 register
*/
#define RL_CFG4_LWPTN 0x04
#define RL_CFG4_LWPME 0x10
#define RL_CFG4_JUMBO_EN1 0x02 /* RTL8168C or later. */
/*
* Config 5 register
*/
#define RL_CFG5_WOL_BCAST 0x40
#define RL_CFG5_WOL_MCAST 0x20
#define RL_CFG5_WOL_UCAST 0x10
#define RL_CFG5_WOL_LANWAKE 0x02
#define RL_CFG5_PME_STS 0x01
/*
* 8139C+ register definitions
*/
/* RL_DUMPSTATS_LO register */
#define RL_DUMPSTATS_START 0x00000008
/* Transmit start register */
#define RL_TXSTART_SWI 0x01 /* generate TX interrupt */
#define RL_TXSTART_START 0x40 /* start normal queue transmit */
#define RL_TXSTART_HPRIO_START 0x80 /* start hi prio queue transmit */
/*
* Config 2 register, 8139C+/8169/8169S/8110S only
*/
#define RL_CFG2_BUSFREQ 0x07
#define RL_CFG2_BUSWIDTH 0x08
#define RL_CFG2_AUXPWRSTS 0x10
#define RL_BUSFREQ_33MHZ 0x00
#define RL_BUSFREQ_66MHZ 0x01
2010-11-08 22:05:11 +00:00
#define RL_BUSWIDTH_32BITS 0x00
#define RL_BUSWIDTH_64BITS 0x08
/* C+ mode command register */
#define RL_CPLUSCMD_TXENB 0x0001 /* enable C+ transmit mode */
#define RL_CPLUSCMD_RXENB 0x0002 /* enable C+ receive mode */
#define RL_CPLUSCMD_PCI_MRW 0x0008 /* enable PCI multi-read/write */
#define RL_CPLUSCMD_PCI_DAC 0x0010 /* PCI dual-address cycle only */
#define RL_CPLUSCMD_RXCSUM_ENB 0x0020 /* enable RX checksum offload */
#define RL_CPLUSCMD_VLANSTRIP 0x0040 /* enable VLAN tag stripping */
#define RL_CPLUSCMD_MACSTAT_DIS 0x0080 /* 8168B/C/CP */
#define RL_CPLUSCMD_ASF 0x0100 /* 8168C/CP */
#define RL_CPLUSCMD_DBG_SEL 0x0200 /* 8168C/CP */
#define RL_CPLUSCMD_FORCE_TXFC 0x0400 /* 8168C/CP */
#define RL_CPLUSCMD_FORCE_RXFC 0x0800 /* 8168C/CP */
#define RL_CPLUSCMD_FORCE_HDPX 0x1000 /* 8168C/CP */
#define RL_CPLUSCMD_NORMAL_MODE 0x2000 /* 8168C/CP */
#define RL_CPLUSCMD_DBG_ENB 0x4000 /* 8168C/CP */
#define RL_CPLUSCMD_BIST_ENB 0x8000 /* 8168C/CP */
/* C+ early transmit threshold */
2010-11-08 22:05:11 +00:00
#define RL_EARLYTXTHRESH_CNT 0x003F /* byte count times 8 */
Do not use interrupt taskqueue on controllers with MSI/MSI-X capability. One of reason using interrupt taskqueue in re(4) was to reduce number of TX/RX interrupts under load because re(4) controllers have no good TX/RX interrupt moderation mechanism. Basic TX interrupt moderation is done by hardware for most controllers but RX interrupt moderation through undocumented register showed poor RX performance so it was disabled in r215025. Using taskqueue to handle RX interrupt greatly reduced number of interrupts but re(4) consumed all available CPU cycles to run the taskqueue under high TX/RX network load. This can happen even with RTL810x fast ethernet controller and I believe this is not acceptable for most systems. To mitigate the issue, use one-shot timer register to moderate RX interrupts. The timer register provides programmable one-shot timer and can be used to suppress interrupt generation. The timer runs at 125MHZ on PCIe controllers so the minimum time allowed for the timer is 8ns. Data sheet says the register is 32 bits but experimentation shows only lower 13 bits are valid so maximum time that can be programmed is 65.528us. This yields theoretical maximum number of RX interrupts that could be generated per second is about 15260. Combined with TX completion interrupts re(4) shall generate less than 20k interrupts. This number is still slightly high compared to other intelligent ethernet controllers but system is very responsive even under high network load. Introduce sysctl variable dev.re.%d.int_rx_mod that controls amount of time to delay RX interrupt processing in units of us. Value 0 completely disables RX interrupt moderation. To provide old behavior for controllers that have MSI/MSI-X capability, introduce a new tunable hw.re.intr_filter. If the tunable is set to non-zero value, driver will use interrupt taskqueue. The default value of the tunable is 0. This tunable has no effect on controllers that has no MSI/MSI-X capability or if MSI/MSI-X is explicitly disabled by administrator. While I'm here cleanup interrupt setup/teardown since re(4) uses single MSI/MSI-X message at this moment.
2011-01-26 20:25:40 +00:00
/* Timer interrupt register */
#define RL_TIMERINT_8169_VAL 0x00001FFF
#define RL_TIMER_MIN 0
#define RL_TIMER_MAX 65 /* 65.528us */
#define RL_TIMER_DEFAULT RL_TIMER_MAX
#define RL_TIMER_PCIE_CLK 125 /* 125MHZ */
#define RL_USECS(x) ((x) * RL_TIMER_PCIE_CLK)
/*
* Gigabit PHY access register (8169 only)
*/
#define RL_PHYAR_PHYDATA 0x0000FFFF
#define RL_PHYAR_PHYREG 0x001F0000
#define RL_PHYAR_BUSY 0x80000000
/*
* Gigabit media status (8169 only)
*/
#define RL_GMEDIASTAT_FDX 0x01 /* full duplex */
#define RL_GMEDIASTAT_LINK 0x02 /* link up */
#define RL_GMEDIASTAT_10MBPS 0x04 /* 10mps link */
#define RL_GMEDIASTAT_100MBPS 0x08 /* 100mbps link */
#define RL_GMEDIASTAT_1000MBPS 0x10 /* gigE link */
#define RL_GMEDIASTAT_RXFLOW 0x20 /* RX flow control on */
#define RL_GMEDIASTAT_TXFLOW 0x40 /* TX flow control on */
#define RL_GMEDIASTAT_TBI 0x80 /* TBI enabled */
/*
* The RealTek doesn't use a fragment-based descriptor mechanism.
* Instead, there are only four register sets, each or which represents
* one 'descriptor.' Basically, each TX descriptor is just a contiguous
* packet buffer (32-bit aligned!) and we place the buffer addresses in
* the registers so the chip knows where they are.
*
* We can sort of kludge together the same kind of buffer management
* used in previous drivers, but we have to do buffer copies almost all
* the time, so it doesn't really buy us much.
*
* For reception, there's just one large buffer where the chip stores
* all received packets.
*/
#define RL_RX_BUF_SZ RL_RXBUF_64
#define RL_RXBUFLEN (1 << ((RL_RX_BUF_SZ >> 11) + 13))
#define RL_TX_LIST_CNT 4
#define RL_MIN_FRAMELEN 60
Various bus_dma(9) fixes. - The hardware does not support DAC so limit DMA address space to 4GB. - Removed BUS_DMA_ALLOC_NOW flag. - Created separated Tx buffer and Rx buffer DMA tags. Previously it used to single DMA tag and it was not possible to specify different DMA restrictions. - Apply 4 bytes alignment limitation of Tx buffer. - Apply 8 bytes alignment limitation of Rx buffer. - Tx side bus_dmamap_load_mbuf_sg(9) support. - Preallocate Tx DMA maps as creating DMA maps take very long time on architectures that require real DMA maps. - Adjust guard buffer size to 1522 + 8 as it should include VLAN and additional reserved bytes in Rx buffer. - Plug memory leak in device detach. Previously wrong buffer address was used to free allocated memory. - Added rl_list_rx_init() to clear Rx buffer and cleared the buffer. - Don't destroy DMA maps in rl_txeof() as the DMA map should be reused. There is no reason to destroy/recreate the DMA maps in this driver. - Removed rl_dma_map_rxbuf()/rl_dma_map_txbuf() callbacks. - The hardware does not support descriptor based DMA on Tx side and the Tx buffer address should be aligned on 4 bytes boundary as well as manual padding for short frames. Because of this hardware limitation rl(4) always used to invoke m_defrag(9) to get a 4 bytes aligned single buffer. However m_defrag(9) takes a lot of CPU cycles on slow machines and not all packets need the help of m_defrag(9). Armed with the information, don't invoke m_defrag(9) if the following conditions are true. 1. Buffer is not fragmented. 2. Buffer is aligned on 4 bytes boundary. 3. Manual padding is not necessary. 4. Or padding is necessary but upper stack passed a writable buffer and the space needed for padding is satisfied. This change combined with preallocated DMA maps greatly increased Tx performance of driver on sparc64. - Moved bus_dmamap_sync(9) in rl_start_locked() to rl_encap() and corrected memory synchronization operation specifier of bus_dmamap_sync(9). - Removed bus_dmamap_unload(9) in rl_stop(). There is no need to reload/unload Rx buffer as rl(4) always have to copy from the buffer. It just needs proper bus_dmamap_sync(9) calls before copying the received frame. With this change rl(4) should work on systems with more than 4GB memory. PR: kern/128143
2008-10-25 02:36:08 +00:00
#define RL_TX_8139_BUF_ALIGN 4
#define RL_RX_8139_BUF_ALIGN 8
#define RL_RX_8139_BUF_RESERVE sizeof(int64_t)
#define RL_RX_8139_BUF_GUARD_SZ \
2010-11-08 22:05:11 +00:00
(ETHER_MAX_LEN + ETHER_VLAN_ENCAP_LEN + RL_RX_8139_BUF_RESERVE)
#define RL_TXTHRESH(x) ((x) << 11)
#define RL_TX_THRESH_INIT 96
#define RL_RX_FIFOTHRESH RL_RXFIFO_NOTHRESH
#define RL_RX_MAXDMA RL_RXDMA_UNLIMITED
#define RL_TX_MAXDMA RL_TXDMA_2048BYTES
#define RL_RXCFG_CONFIG (RL_RX_FIFOTHRESH|RL_RX_MAXDMA|RL_RX_BUF_SZ)
#define RL_TXCFG_CONFIG (RL_TXCFG_IFG|RL_TX_MAXDMA)
#define RL_ETHER_ALIGN 2
/*
* re(4) hardware ip4csum-tx could be mangled with 28 bytes or less IP packets.
*/
#define RL_IP4CSUMTX_MINLEN 28
#define RL_IP4CSUMTX_PADLEN (ETHER_HDR_LEN + RL_IP4CSUMTX_MINLEN)
struct rl_chain_data {
uint16_t cur_rx;
uint8_t *rl_rx_buf;
uint8_t *rl_rx_buf_ptr;
struct mbuf *rl_tx_chain[RL_TX_LIST_CNT];
bus_dmamap_t rl_tx_dmamap[RL_TX_LIST_CNT];
Various bus_dma(9) fixes. - The hardware does not support DAC so limit DMA address space to 4GB. - Removed BUS_DMA_ALLOC_NOW flag. - Created separated Tx buffer and Rx buffer DMA tags. Previously it used to single DMA tag and it was not possible to specify different DMA restrictions. - Apply 4 bytes alignment limitation of Tx buffer. - Apply 8 bytes alignment limitation of Rx buffer. - Tx side bus_dmamap_load_mbuf_sg(9) support. - Preallocate Tx DMA maps as creating DMA maps take very long time on architectures that require real DMA maps. - Adjust guard buffer size to 1522 + 8 as it should include VLAN and additional reserved bytes in Rx buffer. - Plug memory leak in device detach. Previously wrong buffer address was used to free allocated memory. - Added rl_list_rx_init() to clear Rx buffer and cleared the buffer. - Don't destroy DMA maps in rl_txeof() as the DMA map should be reused. There is no reason to destroy/recreate the DMA maps in this driver. - Removed rl_dma_map_rxbuf()/rl_dma_map_txbuf() callbacks. - The hardware does not support descriptor based DMA on Tx side and the Tx buffer address should be aligned on 4 bytes boundary as well as manual padding for short frames. Because of this hardware limitation rl(4) always used to invoke m_defrag(9) to get a 4 bytes aligned single buffer. However m_defrag(9) takes a lot of CPU cycles on slow machines and not all packets need the help of m_defrag(9). Armed with the information, don't invoke m_defrag(9) if the following conditions are true. 1. Buffer is not fragmented. 2. Buffer is aligned on 4 bytes boundary. 3. Manual padding is not necessary. 4. Or padding is necessary but upper stack passed a writable buffer and the space needed for padding is satisfied. This change combined with preallocated DMA maps greatly increased Tx performance of driver on sparc64. - Moved bus_dmamap_sync(9) in rl_start_locked() to rl_encap() and corrected memory synchronization operation specifier of bus_dmamap_sync(9). - Removed bus_dmamap_unload(9) in rl_stop(). There is no need to reload/unload Rx buffer as rl(4) always have to copy from the buffer. It just needs proper bus_dmamap_sync(9) calls before copying the received frame. With this change rl(4) should work on systems with more than 4GB memory. PR: kern/128143
2008-10-25 02:36:08 +00:00
bus_dma_tag_t rl_tx_tag;
bus_dma_tag_t rl_rx_tag;
bus_dmamap_t rl_rx_dmamap;
bus_addr_t rl_rx_buf_paddr;
uint8_t last_tx;
uint8_t cur_tx;
};
#define RL_INC(x) (x = (x + 1) % RL_TX_LIST_CNT)
#define RL_CUR_TXADDR(x) ((x->rl_cdata.cur_tx * 4) + RL_TXADDR0)
#define RL_CUR_TXSTAT(x) ((x->rl_cdata.cur_tx * 4) + RL_TXSTAT0)
#define RL_CUR_TXMBUF(x) (x->rl_cdata.rl_tx_chain[x->rl_cdata.cur_tx])
#define RL_CUR_DMAMAP(x) (x->rl_cdata.rl_tx_dmamap[x->rl_cdata.cur_tx])
#define RL_LAST_TXADDR(x) ((x->rl_cdata.last_tx * 4) + RL_TXADDR0)
#define RL_LAST_TXSTAT(x) ((x->rl_cdata.last_tx * 4) + RL_TXSTAT0)
#define RL_LAST_TXMBUF(x) (x->rl_cdata.rl_tx_chain[x->rl_cdata.last_tx])
#define RL_LAST_DMAMAP(x) (x->rl_cdata.rl_tx_dmamap[x->rl_cdata.last_tx])
struct rl_type {
uint16_t rl_vid;
uint16_t rl_did;
int rl_basetype;
const char *rl_name;
};
struct rl_hwrev {
uint32_t rl_rev;
int rl_type;
const char *rl_desc;
int rl_max_mtu;
};
#define RL_8129 1
#define RL_8139 2
#define RL_8139CPLUS 3
#define RL_8169 4
#define RL_ISCPLUS(x) ((x)->rl_type == RL_8139CPLUS || \
(x)->rl_type == RL_8169)
/*
* The 8139C+ and 8160 gigE chips support descriptor-based TX
* and RX. In fact, they even support TCP large send. Descriptors
* must be allocated in contiguous blocks that are aligned on a
* 256-byte boundary. The rings can hold a maximum of 64 descriptors.
*/
/*
* RX/TX descriptor definition. When large send mode is enabled, the
* lower 11 bits of the TX rl_cmd word are used to hold the MSS, and
* the checksum offload bits are disabled. The structure layout is
* the same for RX and TX descriptors
*/
struct rl_desc {
uint32_t rl_cmdstat;
uint32_t rl_vlanctl;
uint32_t rl_bufaddr_lo;
uint32_t rl_bufaddr_hi;
};
#define RL_TDESC_CMD_FRAGLEN 0x0000FFFF
#define RL_TDESC_CMD_TCPCSUM 0x00010000 /* TCP checksum enable */
#define RL_TDESC_CMD_UDPCSUM 0x00020000 /* UDP checksum enable */
#define RL_TDESC_CMD_IPCSUM 0x00040000 /* IP header checksum enable */
#define RL_TDESC_CMD_MSSVAL 0x07FF0000 /* Large send MSS value */
#define RL_TDESC_CMD_MSSVAL_SHIFT 16 /* Large send MSS value shift */
#define RL_TDESC_CMD_LGSEND 0x08000000 /* TCP large send enb */
#define RL_TDESC_CMD_EOF 0x10000000 /* end of frame marker */
#define RL_TDESC_CMD_SOF 0x20000000 /* start of frame marker */
#define RL_TDESC_CMD_EOR 0x40000000 /* end of ring marker */
#define RL_TDESC_CMD_OWN 0x80000000 /* chip owns descriptor */
#define RL_TDESC_VLANCTL_TAG 0x00020000 /* Insert VLAN tag */
#define RL_TDESC_VLANCTL_DATA 0x0000FFFF /* TAG data */
/* RTL8168C/RTL8168CP/RTL8111C/RTL8111CP */
#define RL_TDESC_CMD_UDPCSUMV2 0x80000000
2010-11-08 22:05:11 +00:00
#define RL_TDESC_CMD_TCPCSUMV2 0x40000000
#define RL_TDESC_CMD_IPCSUMV2 0x20000000
#define RL_TDESC_CMD_MSSVALV2 0x1FFC0000
#define RL_TDESC_CMD_MSSVALV2_SHIFT 18
/*
* Error bits are valid only on the last descriptor of a frame
* (i.e. RL_TDESC_CMD_EOF == 1)
*/
#define RL_TDESC_STAT_COLCNT 0x000F0000 /* collision count */
#define RL_TDESC_STAT_EXCESSCOL 0x00100000 /* excessive collisions */
#define RL_TDESC_STAT_LINKFAIL 0x00200000 /* link faulure */
#define RL_TDESC_STAT_OWINCOL 0x00400000 /* out-of-window collision */
#define RL_TDESC_STAT_TXERRSUM 0x00800000 /* transmit error summary */
#define RL_TDESC_STAT_UNDERRUN 0x02000000 /* TX underrun occured */
#define RL_TDESC_STAT_OWN 0x80000000
/*
* RX descriptor cmd/vlan definitions
*/
#define RL_RDESC_CMD_EOR 0x40000000
#define RL_RDESC_CMD_OWN 0x80000000
#define RL_RDESC_CMD_BUFLEN 0x00001FFF
#define RL_RDESC_STAT_OWN 0x80000000
#define RL_RDESC_STAT_EOR 0x40000000
#define RL_RDESC_STAT_SOF 0x20000000
#define RL_RDESC_STAT_EOF 0x10000000
#define RL_RDESC_STAT_FRALIGN 0x08000000 /* frame alignment error */
#define RL_RDESC_STAT_MCAST 0x04000000 /* multicast pkt received */
#define RL_RDESC_STAT_UCAST 0x02000000 /* unicast pkt received */
#define RL_RDESC_STAT_BCAST 0x01000000 /* broadcast pkt received */
#define RL_RDESC_STAT_BUFOFLOW 0x00800000 /* out of buffer space */
#define RL_RDESC_STAT_FIFOOFLOW 0x00400000 /* FIFO overrun */
#define RL_RDESC_STAT_GIANT 0x00200000 /* pkt > 4096 bytes */
#define RL_RDESC_STAT_RXERRSUM 0x00100000 /* RX error summary */
#define RL_RDESC_STAT_RUNT 0x00080000 /* runt packet received */
#define RL_RDESC_STAT_CRCERR 0x00040000 /* CRC error */
#define RL_RDESC_STAT_PROTOID 0x00030000 /* Protocol type */
#define RL_RDESC_STAT_UDP 0x00020000 /* UDP, 8168C/CP, 8111C/CP */
#define RL_RDESC_STAT_TCP 0x00010000 /* TCP, 8168C/CP, 8111C/CP */
#define RL_RDESC_STAT_IPSUMBAD 0x00008000 /* IP header checksum bad */
#define RL_RDESC_STAT_UDPSUMBAD 0x00004000 /* UDP checksum bad */
#define RL_RDESC_STAT_TCPSUMBAD 0x00002000 /* TCP checksum bad */
#define RL_RDESC_STAT_FRAGLEN 0x00001FFF /* RX'ed frame/frag len */
#define RL_RDESC_STAT_GFRAGLEN 0x00003FFF /* RX'ed frame/frag len */
#define RL_RDESC_STAT_ERRS (RL_RDESC_STAT_GIANT|RL_RDESC_STAT_RUNT| \
RL_RDESC_STAT_CRCERR)
#define RL_RDESC_VLANCTL_TAG 0x00010000 /* VLAN tag available
(rl_vlandata valid)*/
#define RL_RDESC_VLANCTL_DATA 0x0000FFFF /* TAG data */
/* RTL8168C/RTL8168CP/RTL8111C/RTL8111CP */
#define RL_RDESC_IPV6 0x80000000
#define RL_RDESC_IPV4 0x40000000
#define RL_PROTOID_NONIP 0x00000000
#define RL_PROTOID_TCPIP 0x00010000
#define RL_PROTOID_UDPIP 0x00020000
#define RL_PROTOID_IP 0x00030000
#define RL_TCPPKT(x) (((x) & RL_RDESC_STAT_PROTOID) == \
RL_PROTOID_TCPIP)
#define RL_UDPPKT(x) (((x) & RL_RDESC_STAT_PROTOID) == \
RL_PROTOID_UDPIP)
/*
* Statistics counter structure (8139C+ and 8169 only)
*/
struct rl_stats {
uint64_t rl_tx_pkts;
uint64_t rl_rx_pkts;
uint64_t rl_tx_errs;
uint32_t rl_rx_errs;
uint16_t rl_missed_pkts;
uint16_t rl_rx_framealign_errs;
uint32_t rl_tx_onecoll;
uint32_t rl_tx_multicolls;
uint64_t rl_rx_ucasts;
uint64_t rl_rx_bcasts;
uint32_t rl_rx_mcasts;
uint16_t rl_tx_aborts;
uint16_t rl_rx_underruns;
};
/*
* Rx/Tx descriptor parameters (8139C+ and 8169 only)
*
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
* 8139C+
* Number of descriptors supported : up to 64
* Descriptor alignment : 256 bytes
* Tx buffer : At least 4 bytes in length.
* Rx buffer : At least 8 bytes in length and 8 bytes alignment required.
2010-11-08 22:05:11 +00:00
*
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
* 8169
* Number of descriptors supported : up to 1024
* Descriptor alignment : 256 bytes
* Tx buffer : At least 4 bytes in length.
* Rx buffer : At least 8 bytes in length and 8 bytes alignment required.
*/
#ifndef __NO_STRICT_ALIGNMENT
#define RE_FIXUP_RX 1
#endif
#define RL_8169_TX_DESC_CNT 256
#define RL_8169_RX_DESC_CNT 256
#define RL_8139_TX_DESC_CNT 64
#define RL_8139_RX_DESC_CNT 64
#define RL_TX_DESC_CNT RL_8169_TX_DESC_CNT
#define RL_RX_DESC_CNT RL_8169_RX_DESC_CNT
#define RL_RX_JUMBO_DESC_CNT RL_RX_DESC_CNT
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
#define RL_NTXSEGS 32
#define RL_RING_ALIGN 256
#define RL_DUMP_ALIGN 64
#define RL_IFQ_MAXLEN 512
#define RL_TX_DESC_NXT(sc,x) ((x + 1) & ((sc)->rl_ldata.rl_tx_desc_cnt - 1))
#define RL_TX_DESC_PRV(sc,x) ((x - 1) & ((sc)->rl_ldata.rl_tx_desc_cnt - 1))
#define RL_RX_DESC_NXT(sc,x) ((x + 1) & ((sc)->rl_ldata.rl_rx_desc_cnt - 1))
#define RL_OWN(x) (le32toh((x)->rl_cmdstat) & RL_RDESC_STAT_OWN)
#define RL_RXBYTES(x) (le32toh((x)->rl_cmdstat) & sc->rl_rxlenmask)
#define RL_PKTSZ(x) ((x)/* >> 3*/)
#ifdef RE_FIXUP_RX
#define RE_ETHER_ALIGN sizeof(uint64_t)
#define RE_RX_DESC_BUFLEN (MCLBYTES - RE_ETHER_ALIGN)
#else
#define RE_ETHER_ALIGN 0
#define RE_RX_DESC_BUFLEN MCLBYTES
#endif
#define RL_MSI_MESSAGES 1
#define RL_ADDR_LO(y) ((uint64_t) (y) & 0xFFFFFFFF)
#define RL_ADDR_HI(y) ((uint64_t) (y) >> 32)
/*
* The number of bits reserved for MSS in RealTek controllers is
* 11bits. This limits the maximum interface MTU size in TSO case
* as upper stack should not generate TCP segments with MSS greater
* than the limit.
*/
#define RL_TSO_MTU (2047 - ETHER_HDR_LEN - ETHER_CRC_LEN)
/* see comment in dev/re/if_re.c */
#define RL_JUMBO_FRAMELEN 7440
#define RL_JUMBO_MTU \
(RL_JUMBO_FRAMELEN-ETHER_VLAN_ENCAP_LEN-ETHER_HDR_LEN-ETHER_CRC_LEN)
#define RL_JUMBO_MTU_6K \
((6 * 1024) - ETHER_VLAN_ENCAP_LEN - ETHER_HDR_LEN - ETHER_CRC_LEN)
#define RL_JUMBO_MTU_9K \
((9 * 1024) - ETHER_VLAN_ENCAP_LEN - ETHER_HDR_LEN - ETHER_CRC_LEN)
#define RL_MTU \
(ETHER_MAX_LEN + ETHER_VLAN_ENCAP_LEN - ETHER_HDR_LEN - ETHER_CRC_LEN)
Take the support for the 8139C+/8169/8169S/8110S chips out of the rl(4) driver and put it in a new re(4) driver. The re(4) driver shares the if_rlreg.h file with rl(4) but is a separate module. (Ultimately I may change this. For now, it's convenient.) rl(4) has been modified so that it will never attach to an 8139C+ chip, leaving it to re(4) instead. Only re(4) has the PCI IDs to match the 8169/8169S/8110S gigE chips. if_re.c contains the same basic code that was originally bolted onto if_rl.c, with the following updates: - Added support for jumbo frames. Currently, there seems to be a limit of approximately 6200 bytes for jumbo frames on transmit. (This was determined via experimentation.) The 8169S/8110S chips apparently are limited to 7.5K frames on transmit. This may require some more work, though the framework to handle jumbo frames on RX is in place: the re_rxeof() routine will gather up frames than span multiple 2K clusters into a single mbuf list. - Fixed bug in re_txeof(): if we reap some of the TX buffers, but there are still some pending, re-arm the timer before exiting re_txeof() so that another timeout interrupt will be generated, just in case re_start() doesn't do it for us. - Handle the 'link state changed' interrupt - Fix a detach bug. If re(4) is loaded as a module, and you do tcpdump -i re0, then you do 'kldunload if_re,' the system will panic after a few seconds. This happens because ether_ifdetach() ends up calling the BPF detach code, which notices the interface is in promiscuous mode and tries to switch promisc mode off while detaching the BPF listner. This ultimately results in a call to re_ioctl() (due to SIOCSIFFLAGS), which in turn calls re_init() to handle the IFF_PROMISC flag change. Unfortunately, calling re_init() here turns the chip back on and restarts the 1-second timeout loop that drives re_tick(). By the time the timeout fires, if_re.ko has been unloaded, which results in a call to invalid code and blows up the system. To fix this, I cleared the IFF_UP flag before calling ether_ifdetach(), which stops the ioctl routine from trying to reset the chip. - Modified comments in re_rxeof() relating to the difference in RX descriptor status bit layout between the 8139C+ and the gigE chips. The layout is different because the frame length field was expanded from 12 bits to 13, and they got rid of one of the status bits to make room. - Add diagnostic code (re_diag()) to test for the case where a user has installed a broken 32-bit 8169 PCI NIC in a 64-bit slot. Some NICs have the REQ64# and ACK64# lines connected even though the board is 32-bit only (in this case, they should be pulled high). This fools the chip into doing 64-bit DMA transfers even though there is no 64-bit data path. To detect this, re_diag() puts the chip into digital loopback mode and sets the receiver to promiscuous mode, then initiates a single 64-byte packet transmission. The frame is echoed back to the host, and if the frame contents are intact, we know DMA is working correctly, otherwise we complain loudly on the console and abort the device attach. (At the moment, I don't know of any way to work around the problem other than physically modifying the board, so until/unless I can think of a software workaround, this will have do to.) - Created re(4) man page - Modified rlphy.c to allow re(4) to attach as well as rl(4). Note that this code works for the sample 8169/Marvell 88E1000 NIC that I have, but probably won't work for the 8169S/8110S chips. RealTek has sent me some sample NICs, but they haven't arrived yet. I will probably need to add an rlgphy driver to handle the on-board PHY in the 8169S/8110S (it needs special DSP initialization).
2003-09-08 02:11:25 +00:00
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
struct rl_txdesc {
struct mbuf *tx_m;
bus_dmamap_t tx_dmamap;
};
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
struct rl_rxdesc {
struct mbuf *rx_m;
bus_dmamap_t rx_dmamap;
bus_size_t rx_size;
};
struct rl_list_data {
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
struct rl_txdesc rl_tx_desc[RL_TX_DESC_CNT];
struct rl_rxdesc rl_rx_desc[RL_RX_DESC_CNT];
struct rl_rxdesc rl_jrx_desc[RL_RX_JUMBO_DESC_CNT];
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
int rl_tx_desc_cnt;
int rl_rx_desc_cnt;
int rl_tx_prodidx;
int rl_rx_prodidx;
int rl_tx_considx;
int rl_tx_free;
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
bus_dma_tag_t rl_tx_mtag; /* mbuf TX mapping tag */
bus_dma_tag_t rl_rx_mtag; /* mbuf RX mapping tag */
bus_dma_tag_t rl_jrx_mtag; /* mbuf RX mapping tag */
Overhaul re(4). o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs because it's hard to push the hardware to the limit with default 64 descriptors. TSO requires large number of Tx descriptors to pass a full sized TCP segment(65535 bytes IP packet) to hardware. Previously it consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size, to send the TCP segment which means re(4) couldn't queue more than two full sized IP packets. For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware limitations. With this changes there are (very) small waste of memory for 8139C+ users but I don't think it would affect 8139C+ users for most cases. o Various bus_dma(9) fixes. - The hardware supports DAC so allow 64bit DMA operations. - Removed BUS_DMA_ALLOC_NOW flag. - Increased DMA segment size to 4096 from MCLBYTES because TSO consumes too many descriptors with MCLBYTES DMA segment size. - Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these changes the code is more readable than previous one and got a (slightly) better performance as it doesn't need to pass/ decode arguments to/from callback function. - Removed unnecessary callback function re_dmamap_desc() and nuked rl_dmaload_arg structure which was used in the callback. - Additional protection for DMA map load failure. In case of failure reuse current map instead of returning a bogus DMA map. - Deferred DMA map unloading/sync operation for maximum performance until we really need to load new DMA map. If we happen to reuse current map(e.g. input error) there is no need to sync/unload/load again. - The number of allowable Tx DMA segments for a mbuf chains are now 32 instead of magic nseg value. If the number of available Tx descriptors are short enough to send highly fragmented mbuf chains an optimized re_defrag() is called to collapse mbuf chains which is supposed to be much faster than m_defrag(9). re_defrag() was borrowed from ath(4). - Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA tag correctly uses DMA maps that were created with DMA alignment restriction(8bytes alignments). Tx DMA tag does not have such alignment limitation. - Added additional sanity checks for DMA ring map load failure. - Added additional spare Rx DMA map for graceful handling of Rx DMA map load failure. - Fixed misused bus_dmamap_sync(9) and added missing bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof(). o Enabled TSO again as re(4) have reasonable number of Tx descriptors. o Don't touch DMA address of a Tx descriptor in re_txeof(). It's not needed. o Fix incorrect update of if_ierrors counter. For Rx buffer shortage it should update if_qdrops as the buffer is reused. o Added checks for unsupported H/W revisions and return ENXIO for these hardwares. This is required to remove resource allocation code in re_probe as other drivers do in device probe routine. o Modified descriptor index manipulation macros as it's now possible to have different number of descriptors for Rx/Tx. o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap since we already know the number of available Tx descriptors in advance. o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD descriptors in Tx path. There is no such a limitation mentioned in 8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok without reserving RL_TX_DESC_THLD descriptors. o Fix a comment for RL_GTXSTART. The register is 8bits register. o Added comments for 8169/8139C+ hardware restrictions on descriptors. o Removed forward declaration for "struct rl_softc", it's not needed. o Added a new structure rl_txdesc for Tx descriptor managements and a structure rl_rxdesc for Rx descriptor managements. o Removed unused member variable rl_intlock in driver softc. There are still several unused member variables which are supposed to be used to access hardware statistics counters. But it seems that accessing hardware counters were not implemented yet.
2008-01-15 01:10:31 +00:00
bus_dmamap_t rl_rx_sparemap;
bus_dmamap_t rl_jrx_sparemap;
bus_dma_tag_t rl_stag; /* stats mapping tag */
bus_dmamap_t rl_smap; /* stats map */
struct rl_stats *rl_stats;
bus_addr_t rl_stats_addr;
bus_dma_tag_t rl_rx_list_tag;
bus_dmamap_t rl_rx_list_map;
struct rl_desc *rl_rx_list;
bus_addr_t rl_rx_list_addr;
bus_dma_tag_t rl_tx_list_tag;
bus_dmamap_t rl_tx_list_map;
struct rl_desc *rl_tx_list;
bus_addr_t rl_tx_list_addr;
};
enum rl_twist { DONE, CHK_LINK, FIND_ROW, SET_PARAM, RECHK_LONG, RETUNE };
struct rl_softc {
struct ifnet *rl_ifp; /* interface info */
bus_space_handle_t rl_bhandle; /* bus space handle */
bus_space_tag_t rl_btag; /* bus space tag */
device_t rl_dev;
struct resource *rl_res;
int rl_res_id;
int rl_res_type;
struct resource *rl_res_pba;
struct resource *rl_irq[RL_MSI_MESSAGES];
void *rl_intrhand[RL_MSI_MESSAGES];
device_t rl_miibus;
bus_dma_tag_t rl_parent_tag;
uint8_t rl_type;
const struct rl_hwrev *rl_hwrev;
int rl_eecmd_read;
int rl_eewidth;
int rl_expcap;
int rl_txthresh;
bus_size_t rl_cfg0;
bus_size_t rl_cfg1;
bus_size_t rl_cfg2;
bus_size_t rl_cfg3;
bus_size_t rl_cfg4;
bus_size_t rl_cfg5;
struct rl_chain_data rl_cdata;
struct rl_list_data rl_ldata;
struct callout rl_stat_callout;
2006-12-01 21:52:07 +00:00
int rl_watchdog_timer;
struct mtx rl_mtx;
Take the support for the 8139C+/8169/8169S/8110S chips out of the rl(4) driver and put it in a new re(4) driver. The re(4) driver shares the if_rlreg.h file with rl(4) but is a separate module. (Ultimately I may change this. For now, it's convenient.) rl(4) has been modified so that it will never attach to an 8139C+ chip, leaving it to re(4) instead. Only re(4) has the PCI IDs to match the 8169/8169S/8110S gigE chips. if_re.c contains the same basic code that was originally bolted onto if_rl.c, with the following updates: - Added support for jumbo frames. Currently, there seems to be a limit of approximately 6200 bytes for jumbo frames on transmit. (This was determined via experimentation.) The 8169S/8110S chips apparently are limited to 7.5K frames on transmit. This may require some more work, though the framework to handle jumbo frames on RX is in place: the re_rxeof() routine will gather up frames than span multiple 2K clusters into a single mbuf list. - Fixed bug in re_txeof(): if we reap some of the TX buffers, but there are still some pending, re-arm the timer before exiting re_txeof() so that another timeout interrupt will be generated, just in case re_start() doesn't do it for us. - Handle the 'link state changed' interrupt - Fix a detach bug. If re(4) is loaded as a module, and you do tcpdump -i re0, then you do 'kldunload if_re,' the system will panic after a few seconds. This happens because ether_ifdetach() ends up calling the BPF detach code, which notices the interface is in promiscuous mode and tries to switch promisc mode off while detaching the BPF listner. This ultimately results in a call to re_ioctl() (due to SIOCSIFFLAGS), which in turn calls re_init() to handle the IFF_PROMISC flag change. Unfortunately, calling re_init() here turns the chip back on and restarts the 1-second timeout loop that drives re_tick(). By the time the timeout fires, if_re.ko has been unloaded, which results in a call to invalid code and blows up the system. To fix this, I cleared the IFF_UP flag before calling ether_ifdetach(), which stops the ioctl routine from trying to reset the chip. - Modified comments in re_rxeof() relating to the difference in RX descriptor status bit layout between the 8139C+ and the gigE chips. The layout is different because the frame length field was expanded from 12 bits to 13, and they got rid of one of the status bits to make room. - Add diagnostic code (re_diag()) to test for the case where a user has installed a broken 32-bit 8169 PCI NIC in a 64-bit slot. Some NICs have the REQ64# and ACK64# lines connected even though the board is 32-bit only (in this case, they should be pulled high). This fools the chip into doing 64-bit DMA transfers even though there is no 64-bit data path. To detect this, re_diag() puts the chip into digital loopback mode and sets the receiver to promiscuous mode, then initiates a single 64-byte packet transmission. The frame is echoed back to the host, and if the frame contents are intact, we know DMA is working correctly, otherwise we complain loudly on the console and abort the device attach. (At the moment, I don't know of any way to work around the problem other than physically modifying the board, so until/unless I can think of a software workaround, this will have do to.) - Created re(4) man page - Modified rlphy.c to allow re(4) to attach as well as rl(4). Note that this code works for the sample 8169/Marvell 88E1000 NIC that I have, but probably won't work for the 8169S/8110S chips. RealTek has sent me some sample NICs, but they haven't arrived yet. I will probably need to add an rlgphy driver to handle the on-board PHY in the 8169S/8110S (it needs special DSP initialization).
2003-09-08 02:11:25 +00:00
struct mbuf *rl_head;
struct mbuf *rl_tail;
uint32_t rl_rxlenmask;
Take the support for the 8139C+/8169/8169S/8110S chips out of the rl(4) driver and put it in a new re(4) driver. The re(4) driver shares the if_rlreg.h file with rl(4) but is a separate module. (Ultimately I may change this. For now, it's convenient.) rl(4) has been modified so that it will never attach to an 8139C+ chip, leaving it to re(4) instead. Only re(4) has the PCI IDs to match the 8169/8169S/8110S gigE chips. if_re.c contains the same basic code that was originally bolted onto if_rl.c, with the following updates: - Added support for jumbo frames. Currently, there seems to be a limit of approximately 6200 bytes for jumbo frames on transmit. (This was determined via experimentation.) The 8169S/8110S chips apparently are limited to 7.5K frames on transmit. This may require some more work, though the framework to handle jumbo frames on RX is in place: the re_rxeof() routine will gather up frames than span multiple 2K clusters into a single mbuf list. - Fixed bug in re_txeof(): if we reap some of the TX buffers, but there are still some pending, re-arm the timer before exiting re_txeof() so that another timeout interrupt will be generated, just in case re_start() doesn't do it for us. - Handle the 'link state changed' interrupt - Fix a detach bug. If re(4) is loaded as a module, and you do tcpdump -i re0, then you do 'kldunload if_re,' the system will panic after a few seconds. This happens because ether_ifdetach() ends up calling the BPF detach code, which notices the interface is in promiscuous mode and tries to switch promisc mode off while detaching the BPF listner. This ultimately results in a call to re_ioctl() (due to SIOCSIFFLAGS), which in turn calls re_init() to handle the IFF_PROMISC flag change. Unfortunately, calling re_init() here turns the chip back on and restarts the 1-second timeout loop that drives re_tick(). By the time the timeout fires, if_re.ko has been unloaded, which results in a call to invalid code and blows up the system. To fix this, I cleared the IFF_UP flag before calling ether_ifdetach(), which stops the ioctl routine from trying to reset the chip. - Modified comments in re_rxeof() relating to the difference in RX descriptor status bit layout between the 8139C+ and the gigE chips. The layout is different because the frame length field was expanded from 12 bits to 13, and they got rid of one of the status bits to make room. - Add diagnostic code (re_diag()) to test for the case where a user has installed a broken 32-bit 8169 PCI NIC in a 64-bit slot. Some NICs have the REQ64# and ACK64# lines connected even though the board is 32-bit only (in this case, they should be pulled high). This fools the chip into doing 64-bit DMA transfers even though there is no 64-bit data path. To detect this, re_diag() puts the chip into digital loopback mode and sets the receiver to promiscuous mode, then initiates a single 64-byte packet transmission. The frame is echoed back to the host, and if the frame contents are intact, we know DMA is working correctly, otherwise we complain loudly on the console and abort the device attach. (At the moment, I don't know of any way to work around the problem other than physically modifying the board, so until/unless I can think of a software workaround, this will have do to.) - Created re(4) man page - Modified rlphy.c to allow re(4) to attach as well as rl(4). Note that this code works for the sample 8169/Marvell 88E1000 NIC that I have, but probably won't work for the 8169S/8110S chips. RealTek has sent me some sample NICs, but they haven't arrived yet. I will probably need to add an rlgphy driver to handle the on-board PHY in the 8169S/8110S (it needs special DSP initialization).
2003-09-08 02:11:25 +00:00
int rl_testmode;
int rl_if_flags;
int rl_twister_enable;
enum rl_twist rl_twister;
int rl_twist_row;
int rl_twist_col;
int suspended; /* 0 = normal 1 = suspended */
#ifdef DEVICE_POLLING
int rxcycles;
#endif
struct task rl_inttask;
int rl_txstart;
Do not use interrupt taskqueue on controllers with MSI/MSI-X capability. One of reason using interrupt taskqueue in re(4) was to reduce number of TX/RX interrupts under load because re(4) controllers have no good TX/RX interrupt moderation mechanism. Basic TX interrupt moderation is done by hardware for most controllers but RX interrupt moderation through undocumented register showed poor RX performance so it was disabled in r215025. Using taskqueue to handle RX interrupt greatly reduced number of interrupts but re(4) consumed all available CPU cycles to run the taskqueue under high TX/RX network load. This can happen even with RTL810x fast ethernet controller and I believe this is not acceptable for most systems. To mitigate the issue, use one-shot timer register to moderate RX interrupts. The timer register provides programmable one-shot timer and can be used to suppress interrupt generation. The timer runs at 125MHZ on PCIe controllers so the minimum time allowed for the timer is 8ns. Data sheet says the register is 32 bits but experimentation shows only lower 13 bits are valid so maximum time that can be programmed is 65.528us. This yields theoretical maximum number of RX interrupts that could be generated per second is about 15260. Combined with TX completion interrupts re(4) shall generate less than 20k interrupts. This number is still slightly high compared to other intelligent ethernet controllers but system is very responsive even under high network load. Introduce sysctl variable dev.re.%d.int_rx_mod that controls amount of time to delay RX interrupt processing in units of us. Value 0 completely disables RX interrupt moderation. To provide old behavior for controllers that have MSI/MSI-X capability, introduce a new tunable hw.re.intr_filter. If the tunable is set to non-zero value, driver will use interrupt taskqueue. The default value of the tunable is 0. This tunable has no effect on controllers that has no MSI/MSI-X capability or if MSI/MSI-X is explicitly disabled by administrator. While I'm here cleanup interrupt setup/teardown since re(4) uses single MSI/MSI-X message at this moment.
2011-01-26 20:25:40 +00:00
int rl_int_rx_act;
int rl_int_rx_mod;
uint32_t rl_flags;
#define RL_FLAG_MSI 0x00000001
#define RL_FLAG_AUTOPAD 0x00000002
#define RL_FLAG_PHYWAKE_PM 0x00000004
#define RL_FLAG_PHYWAKE 0x00000008
#define RL_FLAG_JUMBOV2 0x00000010
#define RL_FLAG_PAR 0x00000020
#define RL_FLAG_DESCV2 0x00000040
#define RL_FLAG_MACSTAT 0x00000080
#define RL_FLAG_FASTETHER 0x00000100
#define RL_FLAG_CMDSTOP 0x00000200
#define RL_FLAG_MACRESET 0x00000400
#define RL_FLAG_MSIX 0x00000800
#define RL_FLAG_WOLRXENB 0x00001000
#define RL_FLAG_MACSLEEP 0x00002000
#define RL_FLAG_WAIT_TXPOLL 0x00004000
#define RL_FLAG_CMDSTOP_WAIT_TXQ 0x00008000
#define RL_FLAG_WOL_MANLINK 0x00010000
#define RL_FLAG_PCIE 0x40000000
#define RL_FLAG_LINK 0x80000000
};
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
#define RL_LOCK(_sc) mtx_lock(&(_sc)->rl_mtx)
#define RL_UNLOCK(_sc) mtx_unlock(&(_sc)->rl_mtx)
#define RL_LOCK_ASSERT(_sc) mtx_assert(&(_sc)->rl_mtx, MA_OWNED)
/*
* register space access macros
*/
#define CSR_WRITE_STREAM_4(sc, reg, val) \
bus_space_write_stream_4(sc->rl_btag, sc->rl_bhandle, reg, val)
#define CSR_WRITE_4(sc, reg, val) \
bus_space_write_4(sc->rl_btag, sc->rl_bhandle, reg, val)
#define CSR_WRITE_2(sc, reg, val) \
bus_space_write_2(sc->rl_btag, sc->rl_bhandle, reg, val)
#define CSR_WRITE_1(sc, reg, val) \
bus_space_write_1(sc->rl_btag, sc->rl_bhandle, reg, val)
#define CSR_READ_4(sc, reg) \
bus_space_read_4(sc->rl_btag, sc->rl_bhandle, reg)
#define CSR_READ_2(sc, reg) \
bus_space_read_2(sc->rl_btag, sc->rl_bhandle, reg)
#define CSR_READ_1(sc, reg) \
bus_space_read_1(sc->rl_btag, sc->rl_bhandle, reg)
#define CSR_BARRIER(sc, reg, length, flags) \
bus_space_barrier(sc->rl_btag, sc->rl_bhandle, reg, length, flags)
#define CSR_SETBIT_1(sc, offset, val) \
CSR_WRITE_1(sc, offset, CSR_READ_1(sc, offset) | (val))
#define CSR_CLRBIT_1(sc, offset, val) \
CSR_WRITE_1(sc, offset, CSR_READ_1(sc, offset) & ~(val))
#define CSR_SETBIT_2(sc, offset, val) \
CSR_WRITE_2(sc, offset, CSR_READ_2(sc, offset) | (val))
#define CSR_CLRBIT_2(sc, offset, val) \
CSR_WRITE_2(sc, offset, CSR_READ_2(sc, offset) & ~(val))
#define CSR_SETBIT_4(sc, offset, val) \
CSR_WRITE_4(sc, offset, CSR_READ_4(sc, offset) | (val))
#define CSR_CLRBIT_4(sc, offset, val) \
CSR_WRITE_4(sc, offset, CSR_READ_4(sc, offset) & ~(val))
#define RL_TIMEOUT 1000
#define RL_PHY_TIMEOUT 2000
/*
* General constants that are fun to know.
*
* RealTek PCI vendor ID
*/
#define RT_VENDORID 0x10EC
/*
* RealTek chip device IDs.
*/
#define RT_DEVICEID_8139D 0x8039
#define RT_DEVICEID_8129 0x8129
#define RT_DEVICEID_8101E 0x8136
#define RT_DEVICEID_8138 0x8138
#define RT_DEVICEID_8139 0x8139
#define RT_DEVICEID_8169SC 0x8167
#define RT_DEVICEID_8168 0x8168
#define RT_DEVICEID_8169 0x8169
#define RT_DEVICEID_8100 0x8100
#define RT_REVID_8139CPLUS 0x20
/*
* Accton PCI vendor ID
*/
#define ACCTON_VENDORID 0x1113
/*
* Accton MPX 5030/5038 device ID.
*/
#define ACCTON_DEVICEID_5030 0x1211
/*
* Nortel PCI vendor ID
*/
#define NORTEL_VENDORID 0x126C
/*
* Delta Electronics Vendor ID.
*/
#define DELTA_VENDORID 0x1500
/*
* Delta device IDs.
*/
#define DELTA_DEVICEID_8139 0x1360
/*
* Addtron vendor ID.
*/
#define ADDTRON_VENDORID 0x4033
/*
* Addtron device IDs.
*/
#define ADDTRON_DEVICEID_8139 0x1360
/*
* D-Link vendor ID.
*/
#define DLINK_VENDORID 0x1186
/*
* D-Link DFE-530TX+ device ID
*/
#define DLINK_DEVICEID_530TXPLUS 0x1300
/*
* D-Link DFE-5280T device ID
*/
#define DLINK_DEVICEID_528T 0x4300
#define DLINK_DEVICEID_530T_REVC 0x4302
/*
* D-Link DFE-690TXD device ID
*/
#define DLINK_DEVICEID_690TXD 0x1340
/*
* Corega K.K vendor ID
*/
#define COREGA_VENDORID 0x1259
/*
* Corega FEther CB-TXD device ID
*/
#define COREGA_DEVICEID_FETHERCBTXD 0xa117
/*
* Corega FEtherII CB-TXD device ID
*/
#define COREGA_DEVICEID_FETHERIICBTXD 0xa11e
/*
* Corega CG-LAPCIGT device ID
*/
#define COREGA_DEVICEID_CGLAPCIGT 0xc107
/*
* Linksys vendor ID
*/
#define LINKSYS_VENDORID 0x1737
/*
* Linksys EG1032 device ID
*/
#define LINKSYS_DEVICEID_EG1032 0x1032
/*
* Linksys EG1032 rev 3 sub-device ID
*/
#define LINKSYS_SUBDEVICE_EG1032_REV3 0x0024
/*
* Peppercon vendor ID
*/
#define PEPPERCON_VENDORID 0x1743
/*
* Peppercon ROL-F device ID
*/
#define PEPPERCON_DEVICEID_ROLF 0x8139
/*
* Planex Communications, Inc. vendor ID
*/
#define PLANEX_VENDORID 0x14ea
/*
* Planex FNW-3603-TX device ID
*/
#define PLANEX_DEVICEID_FNW3603TX 0xab06
/*
* Planex FNW-3800-TX device ID
*/
#define PLANEX_DEVICEID_FNW3800TX 0xab07
/*
* LevelOne vendor ID
*/
#define LEVEL1_VENDORID 0x018A
/*
* LevelOne FPC-0106TX devide ID
*/
#define LEVEL1_DEVICEID_FPC0106TX 0x0106
/*
* Compaq vendor ID
*/
#define CP_VENDORID 0x021B
/*
* Edimax vendor ID
*/
#define EDIMAX_VENDORID 0x13D1
/*
* Edimax EP-4103DL cardbus device ID
*/
#define EDIMAX_DEVICEID_EP4103DL 0xAB06
Another small update to the re(4) driver: - Change the workaround for the autopad/checksum offload bug so that instead of lying about the map size, we actually create a properly padded mbuf and map it as usual. The other trick works, but is ugly. This approach also gives us a chance to zero the pad space to avoid possibly leaking data. - With the PCIe devices, it looks issuing a TX command while there's already a transmission in progress doesn't have any effect. In other words, if you send two packets in rapid succession, the second one may end up sitting in the TX DMA ring until another transmit command is issued later in the future. Basically, if re_txeof() sees that there are still descriptors outstanding, it needs to manually resume the TX DMA channel by issuing another TX command to make sure all transmissions are flushed out. (The PCI devices seem to keep the TX channel moving until all descriptors have been consumed. I'm not sure why the PCIe devices behave differently.) (You can see this issue if you do the following test: plug an re(4) interface into another host via crossover cable, and from the other host do 'ping -c 2 <host with re(4) NIC>' to prime the ARP cache, then do 'ping -c 1 -s 1473 <host with re(4) NIC>'. You're supposed to see two packets sent in response, but you may only see one. If you do 'ping -c 1 -s 1473 <host with re(4) NIC>' again, you'll see two packets, but one will be the missing fragment from the last ping, followed by one of the fragments from this ping.) - Add the PCI ID for the US Robotics 997902 NIC, which is based on the RTL8169S. - Add a tsleep() of 1 second in re_detach() after the interrupt handler is disconnected. This should allow any tasks queued up by the ISR to drain. Now, I know you're supposed to use taskqueue_drain() for this, but something about the way taskqueue_drain() works with taskqueue_fast queues doesn't seem quite right, and I refuse to be tricked into fixing it.
2006-08-01 17:18:25 +00:00
/* US Robotics vendor ID */
#define USR_VENDORID 0x16EC
Another small update to the re(4) driver: - Change the workaround for the autopad/checksum offload bug so that instead of lying about the map size, we actually create a properly padded mbuf and map it as usual. The other trick works, but is ugly. This approach also gives us a chance to zero the pad space to avoid possibly leaking data. - With the PCIe devices, it looks issuing a TX command while there's already a transmission in progress doesn't have any effect. In other words, if you send two packets in rapid succession, the second one may end up sitting in the TX DMA ring until another transmit command is issued later in the future. Basically, if re_txeof() sees that there are still descriptors outstanding, it needs to manually resume the TX DMA channel by issuing another TX command to make sure all transmissions are flushed out. (The PCI devices seem to keep the TX channel moving until all descriptors have been consumed. I'm not sure why the PCIe devices behave differently.) (You can see this issue if you do the following test: plug an re(4) interface into another host via crossover cable, and from the other host do 'ping -c 2 <host with re(4) NIC>' to prime the ARP cache, then do 'ping -c 1 -s 1473 <host with re(4) NIC>'. You're supposed to see two packets sent in response, but you may only see one. If you do 'ping -c 1 -s 1473 <host with re(4) NIC>' again, you'll see two packets, but one will be the missing fragment from the last ping, followed by one of the fragments from this ping.) - Add the PCI ID for the US Robotics 997902 NIC, which is based on the RTL8169S. - Add a tsleep() of 1 second in re_detach() after the interrupt handler is disconnected. This should allow any tasks queued up by the ISR to drain. Now, I know you're supposed to use taskqueue_drain() for this, but something about the way taskqueue_drain() works with taskqueue_fast queues doesn't seem quite right, and I refuse to be tricked into fixing it.
2006-08-01 17:18:25 +00:00
/* US Robotics 997902 device ID */
#define USR_DEVICEID_997902 0x0116