diff mbox

[v7] ethernet/arc/arc_emac - Add new driver

Message ID 1371799241-27771-1-git-send-email-abrodkin@synopsys.com
State Superseded, archived
Delegated to: David Miller
Headers show

Commit Message

Alexey Brodkin June 21, 2013, 7:20 a.m. UTC
Driver for non-standard on-chip ethernet device ARC EMAC 10/100,
instantiated in some legacy ARC (Synopsys) FPGA Boards such as
ARCAngel4/ML50x.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Francois Romieu <romieu@fr.zoreil.com>
Cc: Joe Perches <joe@perches.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Mischa Jonker <mjonker@synopsys.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: linux-kernel@vger.kernel.org
Cc: devicetree-discuss@lists.ozlabs.org
Cc: Florian Fainelli <florian@openwrt.org>

---
Hi all,

please take a look at 7th re-spin of the patch.

Here I addressed another set of comments from Andy Shevchenko.

1. Re-use FIRST_OR_LAST_MASK more extensively.
2. Use write-specific barrier "wmb()" instead of generic "mb()".
3. Do netdev allocation as late as possible.
---
 Documentation/devicetree/bindings/net/arc_emac.txt |   38 +
 drivers/net/ethernet/Kconfig                       |    1 +
 drivers/net/ethernet/Makefile                      |    1 +
 drivers/net/ethernet/arc/Kconfig                   |   31 +
 drivers/net/ethernet/arc/Makefile                  |    6 +
 drivers/net/ethernet/arc/emac.h                    |  214 ++++++
 drivers/net/ethernet/arc/emac_main.c               |  806 ++++++++++++++++++++
 drivers/net/ethernet/arc/emac_mdio.c               |  152 ++++
 8 files changed, 1249 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/arc_emac.txt
 create mode 100644 drivers/net/ethernet/arc/Kconfig
 create mode 100644 drivers/net/ethernet/arc/Makefile
 create mode 100644 drivers/net/ethernet/arc/emac.h
 create mode 100644 drivers/net/ethernet/arc/emac_main.c
 create mode 100644 drivers/net/ethernet/arc/emac_mdio.c

Comments

Joe Perches June 21, 2013, 10:32 a.m. UTC | #1
On Fri, 2013-06-21 at 11:20 +0400, Alexey Brodkin wrote:
> Driver for non-standard on-chip ethernet device ARC EMAC 10/100,
> instantiated in some legacy ARC (Synopsys) FPGA Boards such as
> ARCAngel4/ML50x.

Looks fine to me.

One nit you could fix later and a question.

> diff --git a/drivers/net/ethernet/arc/emac_main.c b/drivers/net/ethernet/arc/emac_main.c
[]
> +static int arc_emac_rx(struct net_device *ndev, int budget)
> +{
[]
> +			if (net_ratelimit())
> +				netdev_err(ndev, "incomplete packed received\n");

s/packed/packet/

> +		rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);

32 bit only.  Should the Kconfig block have some arch_arc dependency
so it can't get compiled for 64 bit systems?

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Alexey Brodkin June 21, 2013, 10:53 a.m. UTC | #2
On 06/21/2013 02:32 PM, Joe Perches wrote:
> On Fri, 2013-06-21 at 11:20 +0400, Alexey Brodkin wrote:
>> Driver for non-standard on-chip ethernet device ARC EMAC 10/100,
>> instantiated in some legacy ARC (Synopsys) FPGA Boards such as
>> ARCAngel4/ML50x.
>
> Looks fine to me.
>
> One nit you could fix later and a question.
>
>> diff --git a/drivers/net/ethernet/arc/emac_main.c b/drivers/net/ethernet/arc/emac_main.c
> []
>> +static int arc_emac_rx(struct net_device *ndev, int budget)
>> +{
> []
>> +			if (net_ratelimit())
>> +				netdev_err(ndev, "incomplete packed received\n");
>
> s/packed/packet/

Ooops. Thanks for noticing this subtle misspelling.

>> +		rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);
>
> 32 bit only.  Should the Kconfig block have some arch_arc dependency
> so it can't get compiled for 64 bit systems?

Well, initially I add dependency on ARC in Kconfig.
But later removed it because in driver itself there're really no 
ARC-specific things. So why don't make it available for any other arch?
For example I re-used Xilinx SystemACE driver on ARC even though it was 
only targeted for PPC/Microblaze. And now I need to add another 
dependency (+ || ARC) in Kconfig for all block devices.

But your comment regarding 64-bit arches makes perfect sense. But still 
I don't know if there's a way to restrict usage on only 32-bit arches 
(independent of being x86/arc/mips/etc). I may suppose that only strict 
specification of all virtually supported 32-bit arches - which is 
nonsense I think.

Also independence of architecture makes it possible to compile driver on 
x86 for x86 with native toolchain. In other words anybody may try to 
compile it. And I thought it's if not mandatory but at least very 
welcome - if maintainer or reviewer may compile stuff and teach me how 
to eliminate all compile-time warnings/errors.

But during compilation on x86 one needs to enable OF. Which will lead to 
compilation errors in other things.

So for now I may easily add dependency on ARC if it makes acceptance of 
driver easier.

-Alexey
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andy Shevchenko June 21, 2013, 12:15 p.m. UTC | #3
On Fri, Jun 21, 2013 at 1:53 PM, Alexey Brodkin
<Alexey.Brodkin@synopsys.com> wrote:

[]

> But during compilation on x86 one needs to enable OF. Which will lead to
> compilation errors in other things.

JFYI, x86 has at least one platform with official OF dependency.
Once I actually tried to build kernel for Intel MID with OF enabled
(it required a hack in the arch Kconfig, of course), compiled fine.

--
With Best Regards,
Andy Shevchenko
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Alexey Brodkin June 21, 2013, 12:25 p.m. UTC | #4
On 06/21/2013 04:15 PM, Andy Shevchenko wrote:
> On Fri, Jun 21, 2013 at 1:53 PM, Alexey Brodkin
> <Alexey.Brodkin@synopsys.com> wrote:
>
> []
>
>> But during compilation on x86 one needs to enable OF. Which will lead to
>> compilation errors in other things.
>
> JFYI, x86 has at least one platform with official OF dependency.
> Once I actually tried to build kernel for Intel MID with OF enabled
> (it required a hack in the arch Kconfig, of course), compiled fine.

Ok, good to know it.

Any other comments on patch itself?

-Alexey
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andy Shevchenko June 21, 2013, 12:40 p.m. UTC | #5
On Fri, Jun 21, 2013 at 3:25 PM, Alexey Brodkin
<Alexey.Brodkin@synopsys.com> wrote:
> Any other comments on patch itself?

Looks fine to me.

--
With Best Regards,
Andy Shevchenko
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Joe Perches June 21, 2013, 2:43 p.m. UTC | #6
On Fri, 2013-06-21 at 10:53 +0000, Alexey Brodkin wrote:
> On 06/21/2013 02:32 PM, Joe Perches wrote:
> > On Fri, 2013-06-21 at 11:20 +0400, Alexey Brodkin wrote:
> >> Driver for non-standard on-chip ethernet device ARC EMAC 10/100,
> >> instantiated in some legacy ARC (Synopsys) FPGA Boards such as
> >> ARCAngel4/ML50x.
[]
> >> +		rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);
> >
> > 32 bit only.  Should the Kconfig block have some arch_arc dependency
> > so it can't get compiled for 64 bit systems?
[]
> So for now I may easily add dependency on ARC if it makes acceptance of 
> driver easier.

I don't think it's a big problem.

Maybe add a Kconfig "depends on !64BIT".

Another thing to do would be to run it through sparse
with make C=1


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Arnd Bergmann June 21, 2013, 3:58 p.m. UTC | #7
On Friday 21 June 2013, Joe Perches wrote:
> On Fri, 2013-06-21 at 10:53 +0000, Alexey Brodkin wrote:
> > On 06/21/2013 02:32 PM, Joe Perches wrote:
> > > On Fri, 2013-06-21 at 11:20 +0400, Alexey Brodkin wrote:
> > >> Driver for non-standard on-chip ethernet device ARC EMAC 10/100,
> > >> instantiated in some legacy ARC (Synopsys) FPGA Boards such as
> > >> ARCAngel4/ML50x.
> []
> > >> +          rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);
> > >
> > > 32 bit only.  Should the Kconfig block have some arch_arc dependency
> > > so it can't get compiled for 64 bit systems?
> []
> > So for now I may easily add dependency on ARC if it makes acceptance of 
> > driver easier.
> 
> I don't think it's a big problem.
> 
> Maybe add a Kconfig "depends on !64BIT".
> 
> Another thing to do would be to run it through sparse
> with make C=1

No, I think the driver should just be made 64-bit clean. Not because
anyone is going to need that of course, but to avoid having
to add silly dependencies.

The problem is that rxbd->data is the wrong type. It is declared
as a void*, but it is not a kernel pointer at all, it is a DMA
pointer. Look at this code in context:


+               addr = dma_map_single(&ndev->dev, (void *)rx_buff->skb->data,
+                                     buflen, DMA_FROM_DEVICE);
+               if (dma_mapping_error(&ndev->dev, addr)) {
+                       if (net_ratelimit())
+                               netdev_err(ndev, "cannot dma map\n");
+                       dev_kfree_skb(rx_buff->skb);
+                       stats->rx_errors++;
+                       continue;
+               }
+               dma_unmap_addr_set(&rx_buff, mapping, addr);
+               dma_unmap_len_set(&rx_buff, len, buflen);
+
+               rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);

the 'addr' returned by dma_map_single is what the device really
needs, although it is the same as rx_buff->skb->data with the
trivial definition of dma_map_single.

The last line here needs to be

		rxbd->data = cpu_to_le32(addr);

which fixes the bug, and has no dependency on a 32 bit CPU.

	Arnd
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Joe Perches June 21, 2013, 4:05 p.m. UTC | #8
On Fri, 2013-06-21 at 17:58 +0200, Arnd Bergmann wrote:
> On Friday 21 June 2013, Joe Perches wrote:
> > On Fri, 2013-06-21 at 10:53 +0000, Alexey Brodkin wrote:
> > > On 06/21/2013 02:32 PM, Joe Perches wrote:
> > > > On Fri, 2013-06-21 at 11:20 +0400, Alexey Brodkin wrote:
> > > >> Driver for non-standard on-chip ethernet device ARC EMAC 10/100,
> > > >> instantiated in some legacy ARC (Synopsys) FPGA Boards such as
> > > >> ARCAngel4/ML50x.
> > []
> > > >> +          rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);
> > > >
> > > > 32 bit only.  Should the Kconfig block have some arch_arc dependency
> > > > so it can't get compiled for 64 bit systems?
> > []
> > > So for now I may easily add dependency on ARC if it makes acceptance of 
> > > driver easier.
> > 
> > I don't think it's a big problem.
> > 
> > Maybe add a Kconfig "depends on !64BIT".
> > 
> > Another thing to do would be to run it through sparse
> > with make C=1
> 
> No, I think the driver should just be made 64-bit clean. Not because
> anyone is going to need that of course, but to avoid having
> to add silly dependencies.
> 
> The problem is that rxbd->data is the wrong type. It is declared
> as a void*, but it is not a kernel pointer at all, it is a DMA
> pointer. Look at this code in context:

So it should be declared dma_addr_t then,

> +               addr = dma_map_single(&ndev->dev, (void *)rx_buff->skb->data,
> +                                     buflen, DMA_FROM_DEVICE);
> +               if (dma_mapping_error(&ndev->dev, addr)) {
> +                       if (net_ratelimit())
> +                               netdev_err(ndev, "cannot dma map\n");
> +                       dev_kfree_skb(rx_buff->skb);
> +                       stats->rx_errors++;
> +                       continue;
> +               }
> +               dma_unmap_addr_set(&rx_buff, mapping, addr);
> +               dma_unmap_len_set(&rx_buff, len, buflen);
> +
> +               rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);
> 
> the 'addr' returned by dma_map_single is what the device really
> needs, although it is the same as rx_buff->skb->data with the
> trivial definition of dma_map_single.
> 
> The last line here needs to be
>                 rxbd->data = cpu_to_le32(addr);
> 
> which fixes the bug, and has no dependency on a 32 bit CPU.

It still has a dependency on dma_addr_t size being 32 bit



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Laight June 21, 2013, 4:22 p.m. UTC | #9
> So it should be declared dma_addr_t then,
> 
> > +               addr = dma_map_single(&ndev->dev, (void *)rx_buff->skb->data,
> > +                                     buflen, DMA_FROM_DEVICE);
> > +               if (dma_mapping_error(&ndev->dev, addr)) {
> > +                       if (net_ratelimit())
> > +                               netdev_err(ndev, "cannot dma map\n");
> > +                       dev_kfree_skb(rx_buff->skb);
> > +                       stats->rx_errors++;
> > +                       continue;
> > +               }
> > +               dma_unmap_addr_set(&rx_buff, mapping, addr);
> > +               dma_unmap_len_set(&rx_buff, len, buflen);
> > +
> > +               rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);
> >
> > the 'addr' returned by dma_map_single is what the device really
> > needs, although it is the same as rx_buff->skb->data with the
> > trivial definition of dma_map_single.
> >
> > The last line here needs to be
> >                 rxbd->data = cpu_to_le32(addr);
> >
> > which fixes the bug, and has no dependency on a 32 bit CPU.
> 
> It still has a dependency on dma_addr_t size being 32 bit

No - just a dependency on memory being mapped to a 32bit
physical address accessible from the device.
Many 64 bit systems have devices that can only access 32bit
address space, either the memory has to be allocated from the
correct arena, of DMA bounce buffers are used.

	David



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller June 21, 2013, 6:27 p.m. UTC | #10
From: Joe Perches <joe@perches.com>
Date: Fri, 21 Jun 2013 07:43:33 -0700

> Maybe add a Kconfig "depends on !64BIT".

Please, no.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/Documentation/devicetree/bindings/net/arc_emac.txt b/Documentation/devicetree/bindings/net/arc_emac.txt
new file mode 100644
index 0000000..bcbc3f0
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/arc_emac.txt
@@ -0,0 +1,38 @@ 
+* Synopsys ARC EMAC 10/100 Ethernet driver (EMAC)
+
+Required properties:
+- compatible: Should be "snps,arc-emac"
+- reg: Address and length of the register set for the device
+- interrupts: Should contain the EMAC interrupts
+- clock-frequency: CPU frequency. It is needed to calculate and set polling
+period of EMAC.
+- max-speed: Maximum supported data-rate in Mbit/s. In some HW configurations
+bandwidth of external memory controller might be a limiting factor. That's why
+it's required to specify which data-rate is supported on current SoC or FPGA.
+For example if only 10 Mbit/s is supported (10BASE-T) set "10". If 100 Mbit/s is
+supported (100BASE-TX) set "100".
+- phy: PHY device attached to the EMAC via MDIO bus
+
+Child nodes of the driver are the individual PHY devices connected to the
+MDIO bus. They must have a "reg" property given the PHY address on the MDIO bus.
+
+Optional properties:
+- mac-address: 6 bytes, mac address
+
+Examples:
+
+	ethernet@c0fc2000 {
+		compatible = "snps,arc-emac";
+		reg = <0xc0fc2000 0x3c>;
+		interrupts = <6>;
+		mac-address = [ 00 11 22 33 44 55 ];
+		clock-frequency = <80000000>;
+		max-speed = <100>;
+		phy = <&phy0>;
+
+		#address-cells = <1>;
+		#size-cells = <0>;
+		phy0: ethernet-phy@0 {
+			reg = <1>;
+		};
+	};
diff --git a/drivers/net/ethernet/Kconfig b/drivers/net/ethernet/Kconfig
index 4bedae2..98eec48 100644
--- a/drivers/net/ethernet/Kconfig
+++ b/drivers/net/ethernet/Kconfig
@@ -23,6 +23,7 @@  source "drivers/net/ethernet/aeroflex/Kconfig"
 source "drivers/net/ethernet/alteon/Kconfig"
 source "drivers/net/ethernet/amd/Kconfig"
 source "drivers/net/ethernet/apple/Kconfig"
+source "drivers/net/ethernet/arc/Kconfig"
 source "drivers/net/ethernet/atheros/Kconfig"
 source "drivers/net/ethernet/cadence/Kconfig"
 source "drivers/net/ethernet/adi/Kconfig"
diff --git a/drivers/net/ethernet/Makefile b/drivers/net/ethernet/Makefile
index 183e3f4..b764c73 100644
--- a/drivers/net/ethernet/Makefile
+++ b/drivers/net/ethernet/Makefile
@@ -9,6 +9,7 @@  obj-$(CONFIG_GRETH) += aeroflex/
 obj-$(CONFIG_NET_VENDOR_ALTEON) += alteon/
 obj-$(CONFIG_NET_VENDOR_AMD) += amd/
 obj-$(CONFIG_NET_VENDOR_APPLE) += apple/
+obj-$(CONFIG_NET_VENDOR_ARC) += arc/
 obj-$(CONFIG_NET_VENDOR_ATHEROS) += atheros/
 obj-$(CONFIG_NET_CADENCE) += cadence/
 obj-$(CONFIG_NET_BFIN) += adi/
diff --git a/drivers/net/ethernet/arc/Kconfig b/drivers/net/ethernet/arc/Kconfig
new file mode 100644
index 0000000..514c57f
--- /dev/null
+++ b/drivers/net/ethernet/arc/Kconfig
@@ -0,0 +1,31 @@ 
+#
+# ARC EMAC network device configuration
+#
+
+config NET_VENDOR_ARC
+	bool "ARC devices"
+	default y
+	---help---
+	  If you have a network (Ethernet) card belonging to this class, say Y
+	  and read the Ethernet-HOWTO, available from
+	  <http://www.tldp.org/docs.html#howto>.
+
+	  Note that the answer to this question doesn't directly affect the
+	  kernel: saying N will just cause the configurator to skip all
+	  the questions about ARC cards. If you say Y, you will be asked for
+	  your specific card in the following questions.
+
+if NET_VENDOR_ARC
+
+config ARC_EMAC
+	tristate "ARC EMAC support"
+	select MII
+	select PHYLIB
+	depends on OF_IRQ
+	depends on OF_NET
+	---help---
+	  On some legacy ARC (Synopsys) FPGA boards such as ARCAngel4/ML50x
+	  non-standard on-chip ethernet device ARC EMAC 10/100 is used.
+	  Say Y here if you have such a board.  If unsure, say N.
+
+endif # NET_VENDOR_ARC
diff --git a/drivers/net/ethernet/arc/Makefile b/drivers/net/ethernet/arc/Makefile
new file mode 100644
index 0000000..00c8657
--- /dev/null
+++ b/drivers/net/ethernet/arc/Makefile
@@ -0,0 +1,6 @@ 
+#
+# Makefile for the ARC network device drivers.
+#
+
+arc_emac-objs := emac_main.o emac_mdio.o
+obj-$(CONFIG_ARC_EMAC) += arc_emac.o
diff --git a/drivers/net/ethernet/arc/emac.h b/drivers/net/ethernet/arc/emac.h
new file mode 100644
index 0000000..99adf0f
--- /dev/null
+++ b/drivers/net/ethernet/arc/emac.h
@@ -0,0 +1,214 @@ 
+/*
+ * Copyright (C) 2004-2013 Synopsys, Inc. (www.synopsys.com)
+ *
+ * Registers and bits definitions of ARC EMAC
+ */
+
+#ifndef ARC_EMAC_H
+#define ARC_EMAC_H
+
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/netdevice.h>
+#include <linux/phy.h>
+
+/* STATUS and ENABLE Register bit masks */
+#define TXINT_MASK	(1<<0)	/* Transmit interrupt */
+#define RXINT_MASK	(1<<1)	/* Receive interrupt */
+#define ERR_MASK	(1<<2)	/* Error interrupt */
+#define TXCH_MASK	(1<<3)	/* Transmit chaining error interrupt */
+#define MSER_MASK	(1<<4)	/* Missed packet counter error */
+#define RXCR_MASK	(1<<8)	/* RXCRCERR counter rolled over  */
+#define RXFR_MASK	(1<<9)	/* RXFRAMEERR counter rolled over */
+#define RXFL_MASK	(1<<10)	/* RXOFLOWERR counter rolled over */
+#define MDIO_MASK	(1<<12)	/* MDIO complete interrupt */
+#define TXPL_MASK	(1<<31)	/* Force polling of BD by EMAC */
+
+/* CONTROL Register bit masks */
+#define EN_MASK		(1<<0)	/* VMAC enable */
+#define TXRN_MASK	(1<<3)	/* TX enable */
+#define RXRN_MASK	(1<<4)	/* RX enable */
+#define DSBC_MASK	(1<<8)	/* Disable receive broadcast */
+#define ENFL_MASK	(1<<10)	/* Enable Full-duplex */
+#define PROM_MASK	(1<<11)	/* Promiscuous mode */
+
+/* Buffer descriptor INFO bit masks */
+#define OWN_MASK	(1<<31)	/* 0-CPU owns buffer, 1-EMAC owns buffer */
+#define FIRST_MASK	(1<<16)	/* First buffer in chain */
+#define LAST_MASK	(1<<17)	/* Last buffer in chain */
+#define LEN_MASK	0x000007FF	/* last 11 bits */
+#define CRLS		(1<<21)
+#define DEFR		(1<<22)
+#define DROP		(1<<23)
+#define RTRY		(1<<24)
+#define LTCL		(1<<28)
+#define UFLO		(1<<29)
+
+#define FOR_EMAC	OWN_MASK
+#define FOR_CPU		0
+
+/* ARC EMAC register set combines entries for MAC and MDIO */
+enum {
+	R_ID = 0,
+	R_STATUS,
+	R_ENABLE,
+	R_CTRL,
+	R_POLLRATE,
+	R_RXERR,
+	R_MISS,
+	R_TX_RING,
+	R_RX_RING,
+	R_ADDRL,
+	R_ADDRH,
+	R_LAFL,
+	R_LAFH,
+	R_MDIO,
+};
+
+#define TX_TIMEOUT		(400*HZ/1000)	/* Transmission timeout */
+
+#define ARC_EMAC_NAPI_WEIGHT	40		/* Workload for NAPI */
+
+#define EMAC_BUFFER_SIZE	1536		/* EMAC buffer size */
+
+/**
+ * struct arc_emac_bd - EMAC buffer descriptor (BD).
+ *
+ * @info:	Contains status information on the buffer itself.
+ * @data:	32-bit byte addressable pointer to the packet data.
+ */
+struct arc_emac_bd {
+	__le32 info;
+	unsigned char *data;
+};
+
+/* Number of Rx/Tx BD's */
+#define RX_BD_NUM	128
+#define TX_BD_NUM	128
+
+#define RX_RING_SZ	(RX_BD_NUM * sizeof(struct arc_emac_bd))
+#define TX_RING_SZ	(TX_BD_NUM * sizeof(struct arc_emac_bd))
+
+/**
+ * struct buffer_state - Stores Rx/Tx buffer state.
+ * @sk_buff:	Pointer to socket buffer.
+ * @addr:	Start address of DMA-mapped memory region.
+ * @len:	Length of DMA-mapped memory region.
+ */
+struct buffer_state {
+	struct sk_buff *skb;
+	DEFINE_DMA_UNMAP_ADDR(addr);
+	DEFINE_DMA_UNMAP_LEN(len);
+};
+
+/**
+ * struct arc_emac_priv - Storage of EMAC's private information.
+ * @dev:	Pointer to the current device.
+ * @ndev:	Pointer to the current network device.
+ * @phy_dev:	Pointer to attached PHY device.
+ * @bus:	Pointer to the current MII bus.
+ * @regs:	Base address of EMAC memory-mapped control registers.
+ * @napi:	Structure for NAPI.
+ * @stats:	Network device statistics.
+ * @rxbd:	Pointer to Rx BD ring.
+ * @txbd:	Pointer to Tx BD ring.
+ * @rxbd_dma:	DMA handle for Rx BD ring.
+ * @txbd_dma:	DMA handle for Tx BD ring.
+ * @rx_buff:	Storage for Rx buffers states.
+ * @tx_buff:	Storage for Tx buffers states.
+ * @txbd_curr:	Index of Tx BD to use on the next "ndo_start_xmit".
+ * @txbd_dirty:	Index of Tx BD to free on the next Tx interrupt.
+ * @last_rx_bd:	Index of the last Rx BD we've got from EMAC.
+ * @link:	PHY's last seen link state.
+ * @duplex:	PHY's last set duplex mode.
+ * @speed:	PHY's last set speed.
+ * @max_speed:	Maximum supported by current system network data-rate.
+ */
+struct arc_emac_priv {
+	/* Devices */
+	struct device *dev;
+	struct net_device *ndev;
+	struct phy_device *phy_dev;
+	struct mii_bus *bus;
+
+	void __iomem *regs;
+
+	struct napi_struct napi;
+	struct net_device_stats stats;
+
+	struct arc_emac_bd *rxbd;
+	struct arc_emac_bd *txbd;
+
+	dma_addr_t rxbd_dma;
+	dma_addr_t txbd_dma;
+
+	struct buffer_state rx_buff[RX_BD_NUM];
+	struct buffer_state tx_buff[TX_BD_NUM];
+	unsigned int txbd_curr;
+	unsigned int txbd_dirty;
+
+	unsigned int last_rx_bd;
+
+	unsigned int link;
+	unsigned int duplex;
+	unsigned int speed;
+	unsigned int max_speed;
+};
+
+/**
+ * arc_reg_set - Sets EMAC register with provided value.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ * @value:	Value to set in register.
+ */
+static inline void arc_reg_set(struct arc_emac_priv *priv, int reg, int value)
+{
+	iowrite32(value, priv->regs + reg * sizeof(int));
+}
+
+/**
+ * arc_reg_get - Gets value of specified EMAC register.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ *
+ * returns:	Value of requested register.
+ */
+static inline unsigned int arc_reg_get(struct arc_emac_priv *priv, int reg)
+{
+	return ioread32(priv->regs + reg * sizeof(int));
+}
+
+/**
+ * arc_reg_or - Applies mask to specified EMAC register - ("reg" | "mask").
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ * @mask:	Mask to apply to specified register.
+ *
+ * This function reads initial register value, then applies provided mask
+ * to it and then writes register back.
+ */
+static inline void arc_reg_or(struct arc_emac_priv *priv, int reg, int mask)
+{
+	unsigned int value = arc_reg_get(priv, reg);
+	arc_reg_set(priv, reg, value | mask);
+}
+
+/**
+ * arc_reg_clr - Applies mask to specified EMAC register - ("reg" & ~"mask").
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ * @mask:	Mask to apply to specified register.
+ *
+ * This function reads initial register value, then applies provided mask
+ * to it and then writes register back.
+ */
+static inline void arc_reg_clr(struct arc_emac_priv *priv, int reg, int mask)
+{
+	unsigned int value = arc_reg_get(priv, reg);
+	arc_reg_set(priv, reg, value & ~mask);
+}
+
+int arc_mdio_probe(struct platform_device *pdev, struct arc_emac_priv *priv);
+int arc_mdio_remove(struct arc_emac_priv *priv);
+
+#endif /* ARC_EMAC_H */
diff --git a/drivers/net/ethernet/arc/emac_main.c b/drivers/net/ethernet/arc/emac_main.c
new file mode 100644
index 0000000..1606cfb
--- /dev/null
+++ b/drivers/net/ethernet/arc/emac_main.c
@@ -0,0 +1,806 @@ 
+/*
+ * Copyright (C) 2004-2013 Synopsys, Inc. (www.synopsys.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Driver for the ARC EMAC 10100 (hardware revision 5)
+ *
+ * Contributors:
+ *		Amit Bhor
+ *		Sameer Dhavale
+ *		Vineet Gupta
+ */
+
+#include <linux/etherdevice.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_mdio.h>
+#include <linux/of_net.h>
+#include <linux/of_platform.h>
+
+#include "emac.h"
+
+#define DRV_NAME	"arc_emac"
+#define DRV_VERSION	"1.0"
+
+/**
+ * arc_emac_adjust_link - Adjust the PHY link duplex.
+ * @ndev:	Pointer to the net_device structure.
+ *
+ * This function is called to change the duplex setting after auto negotiation
+ * is done by the PHY.
+ */
+static void arc_emac_adjust_link(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct phy_device *phy_dev = priv->phy_dev;
+	unsigned int reg, state_changed = 0;
+
+	if (priv->link != phy_dev->link) {
+		priv->link = phy_dev->link;
+		state_changed = 1;
+	}
+
+	if (priv->speed != phy_dev->speed) {
+		priv->speed = phy_dev->speed;
+		state_changed = 1;
+	}
+
+	if (priv->duplex != phy_dev->duplex) {
+		reg = arc_reg_get(priv, R_CTRL);
+
+		if (DUPLEX_FULL == phy_dev->duplex)
+			reg |= ENFL_MASK;
+		else
+			reg &= ~ENFL_MASK;
+
+		arc_reg_set(priv, R_CTRL, reg);
+		priv->duplex = phy_dev->duplex;
+		state_changed = 1;
+	}
+
+	if (state_changed)
+		phy_print_status(phy_dev);
+}
+
+/**
+ * arc_emac_get_settings - Get PHY settings.
+ * @ndev:	Pointer to net_device structure.
+ * @cmd:	Pointer to ethtool_cmd structure.
+ *
+ * This implements ethtool command for getting PHY settings. If PHY could
+ * not be found, the function returns -ENODEV. This function calls the
+ * relevant PHY ethtool API to get the PHY settings.
+ * Issue "ethtool ethX" under linux prompt to execute this function.
+ */
+static int arc_emac_get_settings(struct net_device *ndev,
+				 struct ethtool_cmd *cmd)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	return phy_ethtool_gset(priv->phy_dev, cmd);
+}
+
+/**
+ * arc_emac_set_settings - Set PHY settings as passed in the argument.
+ * @ndev:	Pointer to net_device structure.
+ * @cmd:	Pointer to ethtool_cmd structure.
+ *
+ * This implements ethtool command for setting various PHY settings. If PHY
+ * could not be found, the function returns -ENODEV. This function calls the
+ * relevant PHY ethtool API to set the PHY.
+ * Issue e.g. "ethtool -s ethX speed 1000" under linux prompt to execute this
+ * function.
+ */
+static int arc_emac_set_settings(struct net_device *ndev,
+				 struct ethtool_cmd *cmd)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	if (!capable(CAP_NET_ADMIN))
+		return -EPERM;
+
+	return phy_ethtool_sset(priv->phy_dev, cmd);
+}
+
+/**
+ * arc_emac_get_drvinfo - Get EMAC driver information.
+ * @ndev:	Pointer to net_device structure.
+ * @info:	Pointer to ethtool_drvinfo structure.
+ *
+ * This implements ethtool command for getting the driver information.
+ * Issue "ethtool -i ethX" under linux prompt to execute this function.
+ */
+static void arc_emac_get_drvinfo(struct net_device *ndev,
+				 struct ethtool_drvinfo *info)
+{
+	strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
+	strlcpy(info->version, DRV_VERSION, sizeof(info->version));
+}
+
+static const struct ethtool_ops arc_emac_ethtool_ops = {
+	.get_settings	= arc_emac_get_settings,
+	.set_settings	= arc_emac_set_settings,
+	.get_drvinfo	= arc_emac_get_drvinfo,
+	.get_link	= ethtool_op_get_link,
+};
+
+#define FIRST_OR_LAST_MASK	(FIRST_MASK | LAST_MASK)
+
+/**
+ * arc_emac_tx_clean - clears processed by EMAC Tx BDs.
+ * @ndev:	Pointer to the network device.
+ */
+static void arc_emac_tx_clean(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct net_device_stats *stats = &priv->stats;
+	unsigned int i;
+
+	for (i = 0; i < TX_BD_NUM; i++) {
+		unsigned int *txbd_dirty = &priv->txbd_dirty;
+		struct arc_emac_bd *txbd = &priv->txbd[*txbd_dirty];
+		struct buffer_state *tx_buff = &priv->tx_buff[*txbd_dirty];
+		struct sk_buff *skb = tx_buff->skb;
+		unsigned int info = le32_to_cpu(txbd->info);
+
+		*txbd_dirty = (*txbd_dirty + 1) % TX_BD_NUM;
+
+		if ((info & FOR_EMAC) || !txbd->data)
+			break;
+
+		if (unlikely(info & (DROP | DEFR | LTCL | UFLO))) {
+			stats->tx_errors++;
+			stats->tx_dropped++;
+
+			if (info & DEFR)
+				stats->tx_carrier_errors++;
+
+			if (info & LTCL)
+				stats->collisions++;
+
+			if (info & UFLO)
+				stats->tx_fifo_errors++;
+		} else if (likely(info & FIRST_OR_LAST_MASK)) {
+			stats->tx_packets++;
+			stats->tx_bytes += skb->len;
+		}
+
+		dma_unmap_single(&ndev->dev, dma_unmap_addr(&tx_buff, addr),
+				 dma_unmap_len(&tx_buff, len), DMA_TO_DEVICE);
+
+		/* return the sk_buff to system */
+		dev_kfree_skb_irq(skb);
+
+		txbd->data = 0;
+		txbd->info = 0;
+
+		if (netif_queue_stopped(ndev))
+			netif_wake_queue(ndev);
+	}
+}
+
+/**
+ * arc_emac_rx - processing of Rx packets.
+ * @ndev:	Pointer to the network device.
+ * @budget:	How many BDs to process on 1 call.
+ *
+ * returns:	Number of processed BDs
+ *
+ * Iterate through Rx BDs and deliver received packages to upper layer.
+ */
+static int arc_emac_rx(struct net_device *ndev, int budget)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	unsigned int work_done;
+
+	for (work_done = 0; work_done <= budget; work_done++) {
+		unsigned int *last_rx_bd = &priv->last_rx_bd;
+		struct net_device_stats *stats = &priv->stats;
+		struct buffer_state *rx_buff = &priv->rx_buff[*last_rx_bd];
+		struct arc_emac_bd *rxbd = &priv->rxbd[*last_rx_bd];
+		unsigned int buflen = EMAC_BUFFER_SIZE;
+		unsigned int pktlen, info = le32_to_cpu(rxbd->info);
+		struct sk_buff *skb;
+		dma_addr_t addr;
+
+		if (unlikely((info & OWN_MASK) == FOR_EMAC))
+			break;
+
+		/* Make a note that we saw a packet at this BD.
+		 * So next time, driver starts from this + 1
+		 */
+		*last_rx_bd = (*last_rx_bd + 1) % RX_BD_NUM;
+
+		if (unlikely((info & FIRST_OR_LAST_MASK) !=
+			     FIRST_OR_LAST_MASK)) {
+			/* We pre-allocate buffers of MTU size so incoming
+			 * packets won't be split/chained.
+			 */
+			if (net_ratelimit())
+				netdev_err(ndev, "incomplete packed received\n");
+
+			/* Return ownership to EMAC */
+			rxbd->info = cpu_to_le32(FOR_EMAC | buflen);
+			stats->rx_errors++;
+			stats->rx_length_errors++;
+			continue;
+		}
+
+		pktlen = info & LEN_MASK;
+		stats->rx_packets++;
+		stats->rx_bytes += pktlen;
+		skb = rx_buff->skb;
+		skb_put(skb, pktlen);
+		skb->dev = ndev;
+		skb->protocol = eth_type_trans(skb, ndev);
+
+		dma_unmap_single(&ndev->dev, dma_unmap_addr(&rx_buff, addr),
+				 dma_unmap_len(&rx_buff, len), DMA_FROM_DEVICE);
+
+		/* Prepare the BD for next cycle */
+		rx_buff->skb = netdev_alloc_skb_ip_align(ndev, buflen);
+		if (unlikely(!rx_buff->skb)) {
+			stats->rx_errors++;
+			/* Because receive_skb is below, increment rx_dropped */
+			stats->rx_dropped++;
+			continue;
+		}
+
+		/* receive_skb only if new skb was allocated to avoid holes */
+		netif_receive_skb(skb);
+
+		addr = dma_map_single(&ndev->dev, (void *)rx_buff->skb->data,
+				      buflen, DMA_FROM_DEVICE);
+		if (dma_mapping_error(&ndev->dev, addr)) {
+			if (net_ratelimit())
+				netdev_err(ndev, "cannot dma map\n");
+			dev_kfree_skb(rx_buff->skb);
+			stats->rx_errors++;
+			continue;
+		}
+		dma_unmap_addr_set(&rx_buff, mapping, addr);
+		dma_unmap_len_set(&rx_buff, len, buflen);
+
+		rxbd->data = (unsigned char *)cpu_to_le32(rx_buff->skb->data);
+
+		/* Make sure pointer to data buffer is set */
+		wmb();
+
+		/* Return ownership to EMAC */
+		rxbd->info = cpu_to_le32(FOR_EMAC | buflen);
+	}
+
+	return work_done;
+}
+
+/**
+ * arc_emac_poll - NAPI poll handler.
+ * @napi:	Pointer to napi_struct structure.
+ * @budget:	How many BDs to process on 1 call.
+ *
+ * returns:	Number of processed BDs
+ */
+static int arc_emac_poll(struct napi_struct *napi, int budget)
+{
+	struct net_device *ndev = napi->dev;
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	unsigned int work_done;
+
+	arc_emac_tx_clean(ndev);
+
+	work_done = arc_emac_rx(ndev, budget);
+	if (work_done < budget) {
+		napi_complete(napi);
+		arc_reg_or(priv, R_ENABLE, RXINT_MASK);
+	}
+
+	return work_done;
+}
+
+/**
+ * arc_emac_intr - Global interrupt handler for EMAC.
+ * @irq:		irq number.
+ * @dev_instance:	device instance.
+ *
+ * returns: IRQ_HANDLED for all cases.
+ *
+ * ARC EMAC has only 1 interrupt line, and depending on bits raised in
+ * STATUS register we may tell what is a reason for interrupt to fire.
+ */
+static irqreturn_t arc_emac_intr(int irq, void *dev_instance)
+{
+	struct net_device *ndev = dev_instance;
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct net_device_stats *stats = &priv->stats;
+	unsigned int status;
+
+	status = arc_reg_get(priv, R_STATUS);
+	status &= ~MDIO_MASK;
+
+	/* Reset all flags except "MDIO complete" */
+	arc_reg_set(priv, R_STATUS, status);
+
+	if (status & RXINT_MASK) {
+		if (likely(napi_schedule_prep(&priv->napi))) {
+			arc_reg_clr(priv, R_ENABLE, RXINT_MASK);
+			__napi_schedule(&priv->napi);
+		}
+	}
+
+	if (status & ERR_MASK) {
+		/* MSER/RXCR/RXFR/RXFL interrupt fires on corresponding
+		 * 8-bit error counter overrun.
+		 */
+
+		if (status & MSER_MASK) {
+			stats->rx_missed_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+
+		if (status & RXCR_MASK) {
+			stats->rx_crc_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+
+		if (status & RXFR_MASK) {
+			stats->rx_frame_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+
+		if (status & RXFL_MASK) {
+			stats->rx_over_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+/**
+ * arc_emac_open - Open the network device.
+ * @ndev:	Pointer to the network device.
+ *
+ * returns: 0, on success or non-zero error value on failure.
+ *
+ * This function sets the MAC address, requests and enables an IRQ
+ * for the EMAC device and starts the Tx queue.
+ * It also connects to the phy device.
+ */
+static int arc_emac_open(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct phy_device *phy_dev = priv->phy_dev;
+	struct arc_emac_bd *bd;
+	struct sk_buff *skb;
+	int i;
+
+	phy_dev->autoneg = AUTONEG_ENABLE;
+	phy_dev->speed = 0;
+	phy_dev->duplex = 0;
+	phy_dev->advertising = phy_dev->supported;
+
+	if (priv->max_speed > 100) {
+		phy_dev->advertising &= PHY_GBIT_FEATURES;
+	} else if (priv->max_speed <= 100) {
+		phy_dev->advertising &= PHY_BASIC_FEATURES;
+		if (priv->max_speed <= 10) {
+			phy_dev->advertising &= ~SUPPORTED_100baseT_Half;
+			phy_dev->advertising &= ~SUPPORTED_100baseT_Full;
+		}
+	}
+
+	/* Allocate and set buffers for Rx BD's */
+	bd = priv->rxbd;
+	for (i = 0; i < RX_BD_NUM; i++) {
+		skb = netdev_alloc_skb_ip_align(ndev, EMAC_BUFFER_SIZE);
+		if (unlikely(!skb))
+			return -ENOMEM;
+
+		priv->rx_buff[i].skb = skb;
+		bd->data = (unsigned char *)cpu_to_le32(skb->data);
+
+		/* Make sure pointer to data buffer is set */
+		wmb();
+
+		/* Set ownership to EMAC */
+		bd->info = cpu_to_le32(FOR_EMAC | EMAC_BUFFER_SIZE);
+		bd++;
+	}
+
+	priv->last_rx_bd = 0;
+
+	/* Clean Tx BD's */
+	memset(priv->txbd, 0, TX_RING_SZ);
+
+	/* Initialize logical address filter */
+	arc_reg_set(priv, R_LAFL, 0);
+	arc_reg_set(priv, R_LAFH, 0);
+
+	/* Set BD ring pointers for device side */
+	arc_reg_set(priv, R_RX_RING, (unsigned int)priv->rxbd_dma);
+	arc_reg_set(priv, R_TX_RING, (unsigned int)priv->txbd_dma);
+
+	/* Enable interrupts */
+	arc_reg_set(priv, R_ENABLE, RXINT_MASK | ERR_MASK);
+
+	/* Set CONTROL */
+	arc_reg_set(priv, R_CTRL,
+		     (RX_BD_NUM << 24) |	/* RX BD table length */
+		     (TX_BD_NUM << 16) |	/* TX BD table length */
+		     TXRN_MASK | RXRN_MASK);
+
+	napi_enable(&priv->napi);
+
+	/* Enable EMAC */
+	arc_reg_or(priv, R_CTRL, EN_MASK);
+
+	phy_start_aneg(priv->phy_dev);
+
+	netif_start_queue(ndev);
+
+	return 0;
+}
+
+/**
+ * arc_emac_stop - Close the network device.
+ * @ndev:	Pointer to the network device.
+ *
+ * This function stops the Tx queue, disables interrupts and frees the IRQ for
+ * the EMAC device.
+ * It also disconnects the PHY device associated with the EMAC device.
+ */
+static int arc_emac_stop(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	napi_disable(&priv->napi);
+	netif_stop_queue(ndev);
+
+	/* Disable interrupts */
+	arc_reg_clr(priv, R_ENABLE, RXINT_MASK | ERR_MASK);
+
+	/* Disable EMAC */
+	arc_reg_clr(priv, R_CTRL, EN_MASK);
+
+	return 0;
+}
+
+/**
+ * arc_emac_stats - Get system network statistics.
+ * @ndev:	Pointer to net_device structure.
+ *
+ * Returns the address of the device statistics structure.
+ * Statistics are updated in interrupt handler.
+ */
+static struct net_device_stats *arc_emac_stats(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct net_device_stats *stats = &priv->stats;
+	unsigned long miss, rxerr;
+	u8 rxcrc, rxfram, rxoflow;
+
+	rxerr = arc_reg_get(priv, R_RXERR);
+	miss = arc_reg_get(priv, R_MISS);
+
+	rxcrc = rxerr;
+	rxfram = rxerr >> 8;
+	rxoflow = rxerr >> 16;
+
+	stats->rx_errors += miss;
+	stats->rx_errors += rxcrc + rxfram + rxoflow;
+
+	stats->rx_over_errors += rxoflow;
+	stats->rx_frame_errors += rxfram;
+	stats->rx_crc_errors += rxcrc;
+	stats->rx_missed_errors += miss;
+
+	return stats;
+}
+
+/**
+ * arc_emac_tx - Starts the data transmission.
+ * @skb:	sk_buff pointer that contains data to be Transmitted.
+ * @ndev:	Pointer to net_device structure.
+ *
+ * returns: NETDEV_TX_OK, on success
+ *		NETDEV_TX_BUSY, if any of the descriptors are not free.
+ *
+ * This function is invoked from upper layers to initiate transmission.
+ */
+static int arc_emac_tx(struct sk_buff *skb, struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	unsigned int len, *txbd_curr = &priv->txbd_curr;
+	struct net_device_stats *stats = &priv->stats;
+	__le32 *info = &priv->txbd[*txbd_curr].info;
+	dma_addr_t addr;
+
+	if (skb_padto(skb, ETH_ZLEN))
+		return NETDEV_TX_OK;
+
+	len = max_t(unsigned int, ETH_ZLEN, skb->len);
+
+	/* EMAC still holds this buffer in its possession.
+	 * CPU must not modify this buffer descriptor
+	 */
+	if (unlikely((le32_to_cpu(*info) & OWN_MASK) == FOR_EMAC)) {
+		netif_stop_queue(ndev);
+		return NETDEV_TX_BUSY;
+	}
+
+	addr = dma_map_single(&ndev->dev, (void *)skb->data, len,
+			      DMA_TO_DEVICE);
+
+	if (unlikely(dma_mapping_error(&ndev->dev, addr))) {
+		stats->tx_dropped++;
+		stats->tx_errors++;
+		dev_kfree_skb(skb);
+		return NETDEV_TX_OK;
+	}
+	dma_unmap_addr_set(&priv->tx_buff[*txbd_curr], mapping, addr);
+	dma_unmap_len_set(&priv->tx_buff[*txbd_curr], len, len);
+
+	priv->tx_buff[*txbd_curr].skb = skb;
+	priv->txbd[*txbd_curr].data = (unsigned char *)cpu_to_le32(skb->data);
+
+	/* Make sure pointer to data buffer is set */
+	wmb();
+
+	*info = cpu_to_le32(FOR_EMAC | FIRST_OR_LAST_MASK | len);
+
+	/* Increment index to point to the next BD */
+	*txbd_curr = (*txbd_curr + 1) % TX_BD_NUM;
+
+	/* Get "info" of the next BD */
+	info = &priv->txbd[*txbd_curr].info;
+
+	/* Check if if Tx BD ring is full - next BD is still owned by EMAC */
+	if (unlikely((le32_to_cpu(*info) & OWN_MASK) == FOR_EMAC))
+		netif_stop_queue(ndev);
+
+	arc_reg_set(priv, R_STATUS, TXPL_MASK);
+
+	skb_tx_timestamp(skb);
+
+	return NETDEV_TX_OK;
+}
+
+/**
+ * arc_emac_set_address - Set the MAC address for this device.
+ * @ndev:	Pointer to net_device structure.
+ * @p:		6 byte Address to be written as MAC address.
+ *
+ * This function copies the HW address from the sockaddr structure to the
+ * net_device structure and updates the address in HW.
+ *
+ * returns:	-EBUSY if the net device is busy or 0 if the address is set
+ *		successfully.
+ */
+static int arc_emac_set_address(struct net_device *ndev, void *p)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct sockaddr *addr = p;
+	unsigned int addr_low, addr_hi;
+
+	if (netif_running(ndev))
+		return -EBUSY;
+
+	if (!is_valid_ether_addr(addr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	memcpy(ndev->dev_addr, addr->sa_data, ndev->addr_len);
+
+	addr_low = le32_to_cpu(*(__le32 *) &ndev->dev_addr[0]);
+	addr_hi = le16_to_cpu(*(__le16 *) &ndev->dev_addr[4]);
+
+	arc_reg_set(priv, R_ADDRL, addr_low);
+	arc_reg_set(priv, R_ADDRH, addr_hi);
+
+	return 0;
+}
+
+static const struct net_device_ops arc_emac_netdev_ops = {
+	.ndo_open		= arc_emac_open,
+	.ndo_stop		= arc_emac_stop,
+	.ndo_start_xmit		= arc_emac_tx,
+	.ndo_set_mac_address	= arc_emac_set_address,
+	.ndo_get_stats		= arc_emac_stats,
+};
+
+static int arc_emac_probe(struct platform_device *pdev)
+{
+	struct resource res_regs, res_irq;
+	struct device_node *phy_node;
+	struct arc_emac_priv *priv;
+	struct net_device *ndev;
+	const char *mac_addr;
+	unsigned int id, clock_frequency;
+	int err;
+
+	if (!pdev->dev.of_node)
+		return -ENODEV;
+
+	/* Get PHY from device tree */
+	phy_node = of_parse_phandle(pdev->dev.of_node, "phy", 0);
+	if (!phy_node) {
+		dev_err(&pdev->dev, "failed to retrieve phy description from device tree\n");
+		return -ENODEV;
+	}
+
+	/* Get EMAC registers base address from device tree */
+	err = of_address_to_resource(pdev->dev.of_node, 0, &res_regs);
+	if (err) {
+		dev_err(&pdev->dev, "failed to retrieve registers base from device tree\n");
+		return -ENODEV;
+	}
+
+	/* Get CPU clock frequency from device tree */
+	if (of_property_read_u32(pdev->dev.of_node, "clock-frequency",
+				 &clock_frequency)) {
+		dev_err(&pdev->dev, "failed to retrieve <clock-frequency> from device tree\n");
+		return -EINVAL;
+	}
+
+	/* Get IRQ from device tree */
+	err = of_irq_to_resource(pdev->dev.of_node, 0, &res_irq);
+	if (!err) {
+		dev_err(&pdev->dev, "failed to retrieve <irq> value from device tree\n");
+		return -ENODEV;
+	}
+
+	ndev = alloc_etherdev(sizeof(struct arc_emac_priv));
+	if (!ndev)
+		return -ENOMEM;
+
+	SET_NETDEV_DEV(ndev, &pdev->dev);
+
+	ndev->netdev_ops = &arc_emac_netdev_ops;
+	ndev->ethtool_ops = &arc_emac_ethtool_ops;
+	ndev->watchdog_timeo = TX_TIMEOUT;
+	/* FIXME :: no multicast support yet */
+	ndev->flags &= ~IFF_MULTICAST;
+
+	priv = netdev_priv(ndev);
+	priv->dev = &pdev->dev;
+	priv->ndev = ndev;
+
+	priv->regs = devm_ioremap_resource(&pdev->dev, &res_regs);
+	if (IS_ERR(priv->regs)) {
+		err = PTR_ERR(priv->regs);
+		goto out;
+	}
+	dev_dbg(&pdev->dev, "Registers base address is 0x%p\n", priv->regs);
+
+	id = arc_reg_get(priv, R_ID);
+
+	/* Check for EMAC revision 5 or 7, magic number */
+	if (!(id == 0x0005fd02 || id == 0x0007fd02)) {
+		dev_err(&pdev->dev, "ARC EMAC not detected, id=0x%x\n", id);
+		err = -ENODEV;
+		goto out;
+	}
+	dev_info(&pdev->dev, "ARC EMAC detected with id: 0x%x\n", id);
+
+	/* Set poll rate so that it polls every 1 ms */
+	arc_reg_set(priv, R_POLLRATE, clock_frequency / 1000000);
+
+	/* Get max speed of operation from device tree */
+	if (of_property_read_u32(pdev->dev.of_node, "max-speed",
+				 &priv->max_speed)) {
+		dev_err(&pdev->dev, "failed to retrieve <max-speed> from device tree\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	ndev->irq = res_irq.start;
+	dev_info(&pdev->dev, "IRQ is %d\n", ndev->irq);
+
+	/* Register interrupt handler for device */
+	err = devm_request_irq(&pdev->dev, ndev->irq, arc_emac_intr, 0,
+			       ndev->name, ndev);
+	if (err) {
+		dev_err(&pdev->dev, "could not allocate IRQ\n");
+		goto out;
+	}
+
+	/* Get MAC address from device tree */
+	mac_addr = of_get_mac_address(pdev->dev.of_node);
+
+	if (!mac_addr || !is_valid_ether_addr(mac_addr))
+		eth_hw_addr_random(ndev);
+	else
+		memcpy(ndev->dev_addr, mac_addr, ETH_ALEN);
+
+	dev_info(&pdev->dev, "MAC address is now %pM\n", ndev->dev_addr);
+
+	/* Do 1 allocation instead of 2 separate ones for Rx and Tx BD rings */
+	priv->rxbd = dmam_alloc_coherent(&pdev->dev, RX_RING_SZ + TX_RING_SZ,
+					 &priv->rxbd_dma, GFP_KERNEL);
+
+	if (!priv->rxbd) {
+		dev_err(&pdev->dev, "failed to allocate data buffers\n");
+		err = -ENOMEM;
+		goto out;
+	}
+
+	priv->txbd = priv->rxbd + RX_BD_NUM;
+
+	priv->txbd_dma = priv->rxbd_dma + RX_RING_SZ;
+	dev_dbg(&pdev->dev, "EMAC Device addr: Rx Ring [0x%x], Tx Ring[%x]\n",
+		(unsigned int)priv->rxbd_dma, (unsigned int)priv->txbd_dma);
+
+	err = arc_mdio_probe(pdev, priv);
+	if (err) {
+		dev_err(&pdev->dev, "failed to probe MII bus\n");
+		goto out;
+	}
+
+	priv->phy_dev = of_phy_connect(ndev, phy_node, arc_emac_adjust_link, 0,
+				       PHY_INTERFACE_MODE_MII);
+	if (!priv->phy_dev) {
+		dev_err(&pdev->dev, "of_phy_connect() failed\n");
+		err = -ENODEV;
+		goto out;
+	}
+
+	dev_info(&pdev->dev, "connected to %s phy with id 0x%x\n",
+		 priv->phy_dev->drv->name, priv->phy_dev->phy_id);
+
+	netif_napi_add(ndev, &priv->napi, arc_emac_poll, ARC_EMAC_NAPI_WEIGHT);
+
+	err = register_netdev(ndev);
+	if (err) {
+		netif_napi_del(&priv->napi);
+		dev_err(&pdev->dev, "failed to register network device\n");
+		goto out;
+	}
+
+	return 0;
+
+out:
+	free_netdev(ndev);
+	return err;
+}
+
+static int arc_emac_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	phy_disconnect(priv->phy_dev);
+	priv->phy_dev = NULL;
+	arc_mdio_remove(priv);
+	unregister_netdev(ndev);
+	netif_napi_del(&priv->napi);
+	free_netdev(ndev);
+
+	return 0;
+}
+
+static const struct of_device_id arc_emac_dt_ids[] = {
+	{ .compatible = "snps,arc-emac" },
+	{ /* Sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, arc_emac_dt_ids);
+
+static struct platform_driver arc_emac_driver = {
+	.probe = arc_emac_probe,
+	.remove = arc_emac_remove,
+	.driver = {
+		.name = DRV_NAME,
+		.owner = THIS_MODULE,
+		.of_match_table  = arc_emac_dt_ids,
+		},
+};
+
+module_platform_driver(arc_emac_driver);
+
+MODULE_AUTHOR("Alexey Brodkin <abrodkin@synopsys.com>");
+MODULE_DESCRIPTION("ARC EMAC driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/arc/emac_mdio.c b/drivers/net/ethernet/arc/emac_mdio.c
new file mode 100644
index 0000000..26ba242
--- /dev/null
+++ b/drivers/net/ethernet/arc/emac_mdio.c
@@ -0,0 +1,152 @@ 
+/*
+ * Copyright (C) 2004-2013 Synopsys, Inc. (www.synopsys.com)
+ *
+ * MDIO implementation for ARC EMAC
+ */
+
+#include <linux/delay.h>
+#include <linux/of_mdio.h>
+#include <linux/platform_device.h>
+
+#include "emac.h"
+
+/* Number of seconds we wait for "MDIO complete" flag to appear */
+#define ARC_MDIO_COMPLETE_POLL_COUNT	1
+
+/**
+ * arc_mdio_complete_wait - Waits until MDIO transaction is completed.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ *
+ * returns:	0 on success, -ETIMEDOUT on a timeout.
+ */
+static int arc_mdio_complete_wait(struct arc_emac_priv *priv)
+{
+	unsigned int i;
+
+	for (i = 0; i < ARC_MDIO_COMPLETE_POLL_COUNT * 40; i++) {
+		unsigned int status = arc_reg_get(priv, R_STATUS);
+
+		status &= MDIO_MASK;
+
+		if (status) {
+			/* Reset "MDIO complete" flag */
+			arc_reg_set(priv, R_STATUS, status);
+			return 0;
+		}
+
+		msleep(25);
+	}
+
+	return -ETIMEDOUT;
+}
+
+/**
+ * arc_mdio_read - MDIO interface read function.
+ * @bus:	Pointer to MII bus structure.
+ * @phy_addr:	Address of the PHY device.
+ * @reg_num:	PHY register to read.
+ *
+ * returns:	The register contents on success, -ETIMEDOUT on a timeout.
+ *
+ * Reads the contents of the requested register from the requested PHY
+ * address.
+ */
+static int arc_mdio_read(struct mii_bus *bus, int phy_addr, int reg_num)
+{
+	struct arc_emac_priv *priv = bus->priv;
+	unsigned int value;
+	int error;
+
+	arc_reg_set(priv, R_MDIO,
+		    0x60020000 | (phy_addr << 23) | (reg_num << 18));
+
+	error = arc_mdio_complete_wait(priv);
+	if (error < 0)
+		return error;
+
+	value = arc_reg_get(priv, R_MDIO) & 0xffff;
+
+	dev_dbg(priv->dev, "arc_mdio_read(phy_addr=%i, reg_num=%x) = %x\n",
+		phy_addr, reg_num, value);
+
+	return value;
+}
+
+/**
+ * arc_mdio_write - MDIO interface write function.
+ * @bus:	Pointer to MII bus structure.
+ * @phy_addr:	Address of the PHY device.
+ * @reg_num:	PHY register to write to.
+ * @value:	Value to be written into the register.
+ *
+ * returns:	0 on success, -ETIMEDOUT on a timeout.
+ *
+ * Writes the value to the requested register.
+ */
+static int arc_mdio_write(struct mii_bus *bus, int phy_addr,
+			  int reg_num, u16 value)
+{
+	struct arc_emac_priv *priv = bus->priv;
+
+	dev_dbg(priv->dev,
+		"arc_mdio_write(phy_addr=%i, reg_num=%x, value=%x)\n",
+		phy_addr, reg_num, value);
+
+	arc_reg_set(priv, R_MDIO,
+		     0x50020000 | (phy_addr << 23) | (reg_num << 18) | value);
+
+	return arc_mdio_complete_wait(priv);
+}
+
+/**
+ * arc_mdio_probe - MDIO probe function.
+ * @pdev:	Pointer to platform device.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ *
+ * returns:	0 on success, -ENOMEM when mdiobus_alloc
+ * (to allocate memory for MII bus structure) fails.
+ *
+ * Sets up and registers the MDIO interface.
+ */
+int arc_mdio_probe(struct platform_device *pdev, struct arc_emac_priv *priv)
+{
+	struct mii_bus *bus;
+	int error;
+
+	bus = mdiobus_alloc();
+	if (!bus)
+		return -ENOMEM;
+
+	priv->bus = bus;
+	bus->priv = priv;
+	bus->parent = priv->dev;
+	bus->name = "Synopsys MII Bus",
+	bus->read = &arc_mdio_read;
+	bus->write = &arc_mdio_write;
+
+	snprintf(bus->id, MII_BUS_ID_SIZE, "%s", pdev->name);
+
+	error = of_mdiobus_register(bus, pdev->dev.of_node);
+	if (error) {
+		dev_err(priv->dev, "cannot register MDIO bus %s\n", bus->name);
+		mdiobus_free(bus);
+		return error;
+	}
+
+	return 0;
+}
+
+/**
+ * arc_mdio_remove - MDIO remove function.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ *
+ * Unregisters the MDIO and frees any associate memory for MII bus.
+ */
+int arc_mdio_remove(struct arc_emac_priv *priv)
+{
+	mdiobus_unregister(priv->bus);
+	mdiobus_free(priv->bus);
+	priv->bus = NULL;
+
+	return 0;
+}