About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Documentation / networking / stmmac.txt

Custom Search

Based on kernel version 4.1. Page generated on 2015-06-28 12:13 EST.

1	       STMicroelectronics 10/100/1000 Synopsys Ethernet driver
3	Copyright (C) 2007-2014  STMicroelectronics Ltd
4	Author: Giuseppe Cavallaro <peppe.cavallaro@st.com>
6	This is the driver for the MAC 10/100/1000 on-chip Ethernet controllers
7	(Synopsys IP blocks).
9	Currently this network device driver is for all STi embedded MAC/GMAC
10	(i.e. 7xxx/5xxx SoCs), SPEAr (arm), Loongson1B (mips) and XLINX XC2V3000
11	FF1152AMT0221 D1215994A VIRTEX FPGA board.
13	DWC Ether MAC 10/100/1000 Universal version 3.70a (and older) and DWC Ether
14	MAC 10/100 Universal version 4.0 have been used for developing this driver.
16	This driver supports both the platform bus and PCI.
18	Please, for more information also visit: www.stlinux.com
20	1) Kernel Configuration
21	The kernel configuration option is STMMAC_ETH:
22	 Device Drivers ---> Network device support ---> Ethernet (1000 Mbit) --->
23	 STMicroelectronics 10/100/1000 Ethernet driver (STMMAC_ETH)
25	CONFIG_STMMAC_PLATFORM: is to enable the platform driver.
26	CONFIG_STMMAC_PCI: is to enable the pci driver.
28	2) Driver parameters list:
29		debug: message level (0: no output, 16: all);
30		phyaddr: to manually provide the physical address to the PHY device;
31		dma_rxsize: DMA rx ring size;
32		dma_txsize: DMA tx ring size;
33		buf_sz: DMA buffer size;
34		tc: control the HW FIFO threshold;
35		watchdog: transmit timeout (in milliseconds);
36		flow_ctrl: Flow control ability [on/off];
37		pause: Flow Control Pause Time;
38		eee_timer: tx EEE timer;
39		chain_mode: select chain mode instead of ring.
41	3) Command line options
42	Driver parameters can be also passed in command line by using:
43		stmmaceth=dma_rxsize:128,dma_txsize:512
45	4) Driver information and notes
47	4.1) Transmit process
48	The xmit method is invoked when the kernel needs to transmit a packet; it sets
49	the descriptors in the ring and informs the DMA engine that there is a packet
50	ready to be transmitted.
51	By default, the driver sets the NETIF_F_SG bit in the features field of the
52	net_device structure enabling the scatter-gather feature. This is true on
53	chips and configurations where the checksum can be done in hardware.
54	Once the controller has finished transmitting the packet, napi will be
55	scheduled to release the transmit resources.
57	4.2) Receive process
58	When one or more packets are received, an interrupt happens. The interrupts
59	are not queued so the driver has to scan all the descriptors in the ring during
60	the receive process.
61	This is based on NAPI so the interrupt handler signals only if there is work
62	to be done, and it exits.
63	Then the poll method will be scheduled at some future point.
64	The incoming packets are stored, by the DMA, in a list of pre-allocated socket
65	buffers in order to avoid the memcpy (zero-copy).
67	4.3) Interrupt Mitigation
68	The driver is able to mitigate the number of its DMA interrupts
69	using NAPI for the reception on chips older than the 3.50.
70	New chips have an HW RX-Watchdog used for this mitigation.
71	Mitigation parameters can be tuned by ethtool.
73	4.4) WOL
74	Wake up on Lan feature through Magic and Unicast frames are supported for the
75	GMAC core.
77	4.5) DMA descriptors
78	Driver handles both normal and alternate descriptors. The latter has been only
79	tested on DWC Ether MAC 10/100/1000 Universal version 3.41a and later.
81	STMMAC supports DMA descriptor to operate both in dual buffer (RING)
82	and linked-list(CHAINED) mode. In RING each descriptor points to two
83	data buffer pointers whereas in CHAINED mode they point to only one data
84	buffer pointer. RING mode is the default.
86	In CHAINED mode each descriptor will have pointer to next descriptor in
87	the list, hence creating the explicit chaining in the descriptor itself,
88	whereas such explicit chaining is not possible in RING mode.
90	4.5.1) Extended descriptors
91		The extended descriptors give us information about the Ethernet payload
92		when it is carrying PTP packets or TCP/UDP/ICMP over IP.
93		These are not available on GMAC Synopsys chips older than the 3.50.
94		At probe time the driver will decide if these can be actually used.
95		This support also is mandatory for PTPv2 because the extra descriptors
96		are used for saving the hardware timestamps and Extended Status.
98	4.6) Ethtool support
99	Ethtool is supported.
101	For example, driver statistics (including RMON), internal errors can be taken
102	using:
103	  # ethtool -S ethX command
105	4.7) Jumbo and Segmentation Offloading
106	Jumbo frames are supported and tested for the GMAC.
107	The GSO has been also added but it's performed in software.
108	LRO is not supported.
110	4.8) Physical
111	The driver is compatible with Physical Abstraction Layer to be connected with
112	PHY and GPHY devices.
114	4.9) Platform information
115	Several information can be passed through the platform and device-tree.
117	struct plat_stmmacenet_data {
118		char *phy_bus_name;
119		int bus_id;
120		int phy_addr;
121		int interface;
122		struct stmmac_mdio_bus_data *mdio_bus_data;
123		struct stmmac_dma_cfg *dma_cfg;
124		int clk_csr;
125		int has_gmac;
126		int enh_desc;
127		int tx_coe;
128		int rx_coe;
129		int bugged_jumbo;
130		int pmt;
131		int force_sf_dma_mode;
132		int force_thresh_dma_mode;
133		int riwt_off;
134		int max_speed;
135		int maxmtu;
136		void (*fix_mac_speed)(void *priv, unsigned int speed);
137		void (*bus_setup)(void __iomem *ioaddr);
138		void *(*setup)(struct platform_device *pdev);
139		void (*free)(struct platform_device *pdev, void *priv);
140		int (*init)(struct platform_device *pdev, void *priv);
141		void (*exit)(struct platform_device *pdev, void *priv);
142		void *custom_cfg;
143		void *custom_data;
144		void *bsp_priv;
145	};
147	Where:
148	 o phy_bus_name: phy bus name to attach to the stmmac.
149	 o bus_id: bus identifier.
150	 o phy_addr: the physical address can be passed from the platform.
151		    If it is set to -1 the driver will automatically
152		    detect it at run-time by probing all the 32 addresses.
153	 o interface: PHY device's interface.
154	 o mdio_bus_data: specific platform fields for the MDIO bus.
155	 o dma_cfg: internal DMA parameters
156	   o pbl: the Programmable Burst Length is maximum number of beats to
157	       be transferred in one DMA transaction.
158	       GMAC also enables the 4xPBL by default.
159	   o fixed_burst/mixed_burst/burst_len
160	 o clk_csr: fixed CSR Clock range selection.
161	 o has_gmac: uses the GMAC core.
162	 o enh_desc: if sets the MAC will use the enhanced descriptor structure.
163	 o tx_coe: core is able to perform the tx csum in HW.
164	 o rx_coe: the supports three check sum offloading engine types:
165		   type_1, type_2 (full csum) and no RX coe.
166	 o bugged_jumbo: some HWs are not able to perform the csum in HW for
167			over-sized frames due to limited buffer sizes.
168			Setting this flag the csum will be done in SW on
169			JUMBO frames.
170	 o pmt: core has the embedded power module (optional).
171	 o force_sf_dma_mode: force DMA to use the Store and Forward mode
172			     instead of the Threshold.
173	 o force_thresh_dma_mode: force DMA to use the Threshold mode other than
174			     the Store and Forward mode.
175	 o riwt_off: force to disable the RX watchdog feature and switch to NAPI mode.
176	 o fix_mac_speed: this callback is used for modifying some syscfg registers
177			 (on ST SoCs) according to the link speed negotiated by the
178			 physical layer .
179	 o bus_setup: perform HW setup of the bus. For example, on some ST platforms
180		     this field is used to configure the AMBA  bridge to generate more
181		     efficient STBus traffic.
182	 o setup/init/exit: callbacks used for calling a custom initialization;
183		     this is sometime necessary on some platforms (e.g. ST boxes)
184		     where the HW needs to have set some PIO lines or system cfg
185		     registers. setup should return a pointer to private data,
186		     which will be stored in bsp_priv, and then passed to init and
187		     exit callbacks. init/exit callbacks should not use or modify
188		     platform data.
189	 o custom_cfg/custom_data: this is a custom configuration that can be passed
190				   while initializing the resources.
191	 o bsp_priv: another private pointer.
193	For MDIO bus The we have:
195	 struct stmmac_mdio_bus_data {
196		int (*phy_reset)(void *priv);
197		unsigned int phy_mask;
198		int *irqs;
199		int probed_phy_irq;
200	 };
202	Where:
203	 o phy_reset: hook to reset the phy device attached to the bus.
204	 o phy_mask: phy mask passed when register the MDIO bus within the driver.
205	 o irqs: list of IRQs, one per PHY.
206	 o probed_phy_irq: if irqs is NULL, use this for probed PHY.
208	For DMA engine we have the following internal fields that should be
209	tuned according to the HW capabilities.
211	struct stmmac_dma_cfg {
212		int pbl;
213		int fixed_burst;
214		int burst_len_supported;
215	};
217	Where:
218	 o pbl: Programmable Burst Length
219	 o fixed_burst: program the DMA to use the fixed burst mode
220	 o burst_len: this is the value we put in the register
221		      supported values are provided as macros in
222		      linux/stmmac.h header file.
224	---
226	Below an example how the structures above are using on ST platforms.
228	 static struct plat_stmmacenet_data stxYYY_ethernet_platform_data = {
229		.has_gmac = 0,
230		.enh_desc = 0,
231		.fix_mac_speed = stxYYY_ethernet_fix_mac_speed,
232					|
233					|-> to write an internal syscfg
234					|   on this platform when the
235					|   link speed changes from 10 to
236					|   100 and viceversa
237		.init = &stmmac_claim_resource,
238					|
239					|-> On ST SoC this calls own "PAD"
240					|   manager framework to claim
241					|   all the resources necessary
242					|   (GPIO ...). The .custom_cfg field
243					|   is used to pass a custom config.
244	};
246	Below the usage of the stmmac_mdio_bus_data: on this SoC, in fact,
247	there are two MAC cores: one MAC is for MDIO Bus/PHY emulation
248	with fixed_link support.
250	static struct stmmac_mdio_bus_data stmmac1_mdio_bus = {
251		.phy_reset = phy_reset;
252			|
253			|-> function to provide the phy_reset on this board
254		.phy_mask = 0,
255	};
257	static struct fixed_phy_status stmmac0_fixed_phy_status = {
258		.link = 1,
259		.speed = 100,
260		.duplex = 1,
261	};
263	During the board's device_init we can configure the first
264	MAC for fixed_link by calling:
265	  fixed_phy_add(PHY_POLL, 1, &stmmac0_fixed_phy_status));)
266	and the second one, with a real PHY device attached to the bus,
267	by using the stmmac_mdio_bus_data structure (to provide the id, the
268	reset procedure etc).
270	Note that, starting from new chips, where it is available the HW capability
271	register, many configurations are discovered at run-time for example to
272	understand if EEE, HW csum, PTP, enhanced descriptor etc are actually
273	available. As strategy adopted in this driver, the information from the HW
274	capability register can replace what has been passed from the platform.
276	4.10) Device-tree support.
278	Please see the following document:
279		Documentation/devicetree/bindings/net/stmmac.txt
281	and the stmmac_of_data structure inside the include/linux/stmmac.h header file.
283	4.11) This is a summary of the content of some relevant files:
284	 o stmmac_main.c: to implement the main network device driver;
285	 o stmmac_mdio.c: to provide mdio functions;
286	 o stmmac_pci: this the PCI driver;
287	 o stmmac_platform.c: this the platform driver (OF supported)
288	 o stmmac_ethtool.c: to implement the ethtool support;
289	 o stmmac.h: private driver structure;
290	 o common.h: common definitions and VFTs;
291	 o descs.h: descriptor structure definitions;
292	 o dwmac1000_core.c: dwmac GiGa core functions;
293	 o dwmac1000_dma.c: dma functions for the GMAC chip;
294	 o dwmac1000.h: specific header file for the dwmac GiGa;
295	 o dwmac100_core: dwmac 100 core code;
296	 o dwmac100_dma.c: dma functions for the dwmac 100 chip;
297	 o dwmac1000.h: specific header file for the MAC;
298	 o dwmac_lib.c: generic DMA functions;
299	 o enh_desc.c: functions for handling enhanced descriptors;
300	 o norm_desc.c: functions for handling normal descriptors;
301	 o chain_mode.c/ring_mode.c:: functions to manage RING/CHAINED modes;
302	 o mmc_core.c/mmc.h: Management MAC Counters;
303	 o stmmac_hwtstamp.c: HW timestamp support for PTP;
304	 o stmmac_ptp.c: PTP 1588 clock;
305	 o dwmac-<XXX>.c: these are for the platform glue-logic file; e.g. dwmac-sti.c
306	   for STMicroelectronics SoCs.
308	5) Debug Information
310	The driver exports many information i.e. internal statistics,
311	debug information, MAC and DMA registers etc.
313	These can be read in several ways depending on the
314	type of the information actually needed.
316	For example a user can be use the ethtool support
317	to get statistics: e.g. using: ethtool -S ethX
318	(that shows the Management counters (MMC) if supported)
319	or sees the MAC/DMA registers: e.g. using: ethtool -d ethX
321	Compiling the Kernel with CONFIG_DEBUG_FS the driver will export the following
322	debugfs entries:
324	/sys/kernel/debug/stmmaceth/descriptors_status
325	  To show the DMA TX/RX descriptor rings
327	Developer can also use the "debug" module parameter to get further debug
328	information (please see: NETIF Msg Level).
330	6) Energy Efficient Ethernet
332	Energy Efficient Ethernet(EEE) enables IEEE 802.3 MAC sublayer along
333	with a family of Physical layer to operate in the Low power Idle(LPI)
334	mode. The EEE mode supports the IEEE 802.3 MAC operation at 100Mbps,
335	1000Mbps & 10Gbps.
337	The LPI mode allows power saving by switching off parts of the
338	communication device functionality when there is no data to be
339	transmitted & received. The system on both the side of the link can
340	disable some functionalities & save power during the period of low-link
341	utilization. The MAC controls whether the system should enter or exit
342	the LPI mode & communicate this to PHY.
344	As soon as the interface is opened, the driver verifies if the EEE can
345	be supported. This is done by looking at both the DMA HW capability
346	register and the PHY devices MCD registers.
347	To enter in Tx LPI mode the driver needs to have a software timer
348	that enable and disable the LPI mode when there is nothing to be
349	transmitted.
351	7) Precision Time Protocol (PTP)
352	The driver supports the IEEE 1588-2002, Precision Time Protocol (PTP),
353	which enables precise synchronization of clocks in measurement and
354	control systems implemented with technologies such as network
355	communication.
357	In addition to the basic timestamp features mentioned in IEEE 1588-2002
358	Timestamps, new GMAC cores support the advanced timestamp features.
359	IEEE 1588-2008 that can be enabled when configure the Kernel.
361	8) SGMII/RGMII supports
362	New GMAC devices provide own way to manage RGMII/SGMII.
363	This information is available at run-time by looking at the
364	HW capability register. This means that the stmmac can manage
365	auto-negotiation and link status w/o using the PHYLIB stuff
366	In fact, the HW provides a subset of extended registers to
367	restart the ANE, verify Full/Half duplex mode and Speed.
368	Also thanks to these registers it is possible to look at the
369	Auto-negotiated Link Parter Ability.
Hide Line Numbers
About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Information is copyright its respective author. All material is available from the Linux Kernel Source distributed under a GPL License. This page is provided as a free service by mjmwired.net.