Re: Xilinx PowerPC

Re: Xilinx PowerPC

[David H. Lynch Jr.]

Thanks,

    I have alot of work to do on our stuff, I might as well see if I can
move to the powerpc tree at the same time.

    BTW is there even the beginings of a non-xilinx lltemac driver out
there ? There were hints on the list, but I have not seen anything.

    I would be happy to help advance the ball on anything anyone has
started.

   

Grant Likely wrote:
> On Sun, Apr 20, 2008 at 2:31 PM, David H. Lynch Jr. <dhlii@dlasys.net> wrote:
>   
>> Thanks.
>>
>>     I am running linus's 2.6.25-rc9, but I can pull the Xilinx tree or
>>  yours - when your server is up.
>>
>>     I can not find any Xilinx powerpc configs in arch/powerpc/config
>>     Do I just need to do a
>>     make ARCH=powerpc menuconfig and create one from scratch ?
>>     
>
> That's right; I haven't merged any defconfigs.  Roll your own.
>
>   
>>     Is simpleboot in your tree or the xilinx tree, if  can not find it
>>  in Linus's ?
>>     
>
> If you want to use the "simpleboot" wrapper; then you'll need to pull
> paulus' tree (Linus hasn't yet pulled his tree; but he probably will
> any moment now).
>
> Cheers,
> g.
>
>   


-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein

RE: Xilinx PowerPC

[Koss, Mike (Mission Systems)]

I have the start of a lltemac based on the new XPS_LL_TEMAC from EDK
9.2. I'm currently in the process of upgrading my design to 9.2. So I
will be very much interested in the driver working ;).

I'll see what I can get cleaned up and posted somewhere, if you'd like
to take a look at it.

-- Mike

(sorry for the double post, if it happens but it seems that my last
e-mail was blank per my records.)=20

-----Original Message-----
From: David H. Lynch Jr. [mailto:dhlii@dlasys.net]=20
Sent: Sunday, April 20, 2008 7:49 PM
To: Grant Likely
Cc: linuxppc-embedded
Subject: Re: Xilinx PowerPC

Thanks,

    I have alot of work to do on our stuff, I might as well see if I can
move to the powerpc tree at the same time.

    BTW is there even the beginings of a non-xilinx lltemac driver out
there ? There were hints on the list, but I have not seen anything.

    I would be happy to help advance the ball on anything anyone has
started.

  =20

Grant Likely wrote:
> On Sun, Apr 20, 2008 at 2:31 PM, David H. Lynch Jr. <dhlii@dlasys.net>
wrote:
>  =20
>> Thanks.
>>
>>     I am running linus's 2.6.25-rc9, but I can pull the Xilinx tree=20
>> or  yours - when your server is up.
>>
>>     I can not find any Xilinx powerpc configs in arch/powerpc/config
>>     Do I just need to do a
>>     make ARCH=3Dpowerpc menuconfig and create one from scratch ?
>>    =20
>
> That's right; I haven't merged any defconfigs.  Roll your own.
>
>  =20
>>     Is simpleboot in your tree or the xilinx tree, if  can not find=20
>> it  in Linus's ?
>>    =20
>
> If you want to use the "simpleboot" wrapper; then you'll need to pull=20
> paulus' tree (Linus hasn't yet pulled his tree; but he probably will=20
> any moment now).
>
> Cheers,
> g.
>
>  =20


--=20
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too
numerous to list.

"Any intelligent fool can make things bigger and more complex... It
takes a touch of genius - and a lot of courage to move in the opposite
direction."
Albert Einstein

Re: Xilinx PowerPC

[David H. Lynch Jr.]

 I would be ecstatic to look at/collaborate on what you have.
I do not care if you clean it up. It can not be harder to work with than
my own code or the Xilinx code.

In the end I need Linux, GreenHills, as well as a lightweight stripped
driver for a monitor and embedded mini-web server.
I have all the above for the older PLB TEMAC in FIFO mode, but we are
moving to the LL TEMAC.

I started through the xilinx code and rapidly remembered how long it
took to demacrofy the xilinx PLB TEMAC code down to a single source file
with only
the code actually needed, and got depressed.

This is critical to us now, so I can put time into it.

If I can figure out how to bring up a git server on my hosting service I
am willing to host work on this -
though one of the existing trees might be a better choice.


Koss, Mike (Mission Systems) wrote:
> I have the start of a lltemac based on the new XPS_LL_TEMAC from EDK
> 9.2. I'm currently in the process of upgrading my design to 9.2. So I
> will be very much interested in the driver working ;).
>
> I'll see what I can get cleaned up and posted somewhere, if you'd like
> to take a look at it.
>
> -- Mike
>
> (sorry for the double post, if it happens but it seems that my last
> e-mail was blank per my records.) 
>
> -----Original Message-----
> From: David H. Lynch Jr. [mailto:dhlii@dlasys.net] 
> Sent: Sunday, April 20, 2008 7:49 PM
> To: Grant Likely
> Cc: linuxppc-embedded
> Subject: Re: Xilinx PowerPC
>
> Thanks,
>
>     I have alot of work to do on our stuff, I might as well see if I can
> move to the powerpc tree at the same time.
>
>     BTW is there even the beginings of a non-xilinx lltemac driver out
> there ? There were hints on the list, but I have not seen anything.
>
>     I would be happy to help advance the ball on anything anyone has
> started.
>
>    
>
> Grant Likely wrote:
>   
>> On Sun, Apr 20, 2008 at 2:31 PM, David H. Lynch Jr. <dhlii@dlasys.net>
>>     
> wrote:
>   
>>   
>>     
>>> Thanks.
>>>
>>>     I am running linus's 2.6.25-rc9, but I can pull the Xilinx tree 
>>> or  yours - when your server is up.
>>>
>>>     I can not find any Xilinx powerpc configs in arch/powerpc/config
>>>     Do I just need to do a
>>>     make ARCH=powerpc menuconfig and create one from scratch ?
>>>     
>>>       
>> That's right; I haven't merged any defconfigs.  Roll your own.
>>
>>   
>>     
>>>     Is simpleboot in your tree or the xilinx tree, if  can not find 
>>> it  in Linus's ?
>>>     
>>>       
>> If you want to use the "simpleboot" wrapper; then you'll need to pull 
>> paulus' tree (Linus hasn't yet pulled his tree; but he probably will 
>> any moment now).
>>
>> Cheers,
>> g.
>>
>>   
>>     
>
>
>   


-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein

Re: Xilinx PowerPC

[Yoshio Kashiwagi]

Hi,

I am writing the Non-Xilinx XPS_LL_TEMAC driver.
Checksum offloading is incomplete although NAPI and KGDBOE are supported.
Basic operation is working on EDK9.2 and EDK10.1.

Furthermore, although the simple Non-Interrupt version for u-boot is
also written, it is not known where I should post.

Best Regards,

Yoshio Kashiwagi - Nissin Systems

/*
 *
 * Xilinx Gigabit System Referece Design Ethenet driver
 *
 * Driver for Xilinx Virtex-4FX Based Platform
 *
 * Author: Yoshio Kashiwagi
 *
 * Copyright (c) 2008 Nissin Systems Co.,Ltd.
 *
 * March 2008 created
 *
 * This program is free software; you can redistribute  it and/or modify 
it
 * under  the terms of  the GNU General  Public License as published by 
the
 * Free Software Foundation;  either version 2 of the  License, or (at 
your
 * option) any later version.
 *
*/

#include <linux/autoconf.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/init.h>
#include <linux/in.h>
#include <linux/moduleparam.h>
#include <linux/spinlock.h>
#include <linux/pci.h>
#include <asm/io.h>
#include <asm/delay.h>
#include <platforms/4xx/xparameters/xparameters.h>

#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <linux/ethtool.h>
#include <net/sock.h>
#include <net/checksum.h>
#include <linux/if_ether.h>
#include <linux/if_arp.h>
#include <linux/ip.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/percpu.h>

#define S_DMA_CTRL_BASEADDR XPAR_LLTEMAC_0_LLINK_CONNECTED_BASEADDR
#define XPS_LLTEMAC_BASEADDR    XPAR_LLTEMAC_0_BASEADDR

/* XPS_LL_TEMAC SDMA registers definition */

#define TX_NXTDESC_PTR      0x00
#define TX_CURBUF_ADDR      0x04
#define TX_CURBUF_LENGTH    0x08
#define TX_CURDESC_PTR      0x0c
#define TX_TAILDESC_PTR     0x10
#define TX_CHNL_CTRL        0x14
#define TX_IRQ_REG      0x18
#define TX_CHNL_STS     0x1c

#define RX_NXTDESC_PTR      0x20
#define RX_CURBUF_ADDR      0x24
#define RX_CURBUF_LENGTH    0x28
#define RX_CURDESC_PTR      0x2c
#define RX_TAILDESC_PTR     0x30
#define RX_CHNL_CTRL        0x34
#define RX_IRQ_REG      0x38
#define RX_CHNL_STS     0x3c

#define DMA_CONTROL_REG     0x40

/* XPS_LL_TEMAC direct registers definition */

#define TEMAC_RAF0      0x00
#define TEMAC_TPF0      0x04
#define TEMAC_IFGP0     0x08
#define TEMAC_IS0       0x0c
#define TEMAC_IP0       0x10
#define TEMAC_IE0       0x14

#define TEMAC_MSW0      0x20
#define TEMAC_LSW0      0x24
#define TEMAC_CTL0      0x28
#define TEMAC_RDY0      0x2c

#define XTE_RSE_MIIM_RR_MASK      0x0002
#define XTE_RSE_MIIM_WR_MASK      0x0004
#define XTE_RSE_CFG_RR_MASK       0x0020
#define XTE_RSE_CFG_WR_MASK       0x0040

/* XPS_LL_TEMAC indirect registers offset definition */

#define RCW0    0x200
#define RCW1    0x240
#define TC  0x280
#define FCC 0x2c0
#define EMMC    0x300
#define PHYC    0x320
#define MC  0x340
#define UAW0    0x380
#define UAW1    0x384
#define MAW0    0x388
#define MAW1    0x38c
#define AFM 0x390
#define TIS 0x3a0
#define TIE 0x3a4
#define MIIMWD  0x3b0
#define MIIMAI  0x3b4

#define CNTLREG_WRITE_ENABLE_MASK   0x8000
#define CNTLREG_EMAC1SEL_MASK       0x0400
#define CNTLREG_ADDRESSCODE_MASK    0x03ff

#define MDIO_ENABLE_MASK        0x40
#define MDIO_CLOCK_DIV_MASK 0x3F
#define MDIO_CLOCK_DIV_100MHz   0x28

#define ETHER_MTU       1500
#define EMAC_PHY_ID     7

/* CDMAC descriptor status bit definitions */

#define BDSTAT_ERROR_MASK       0x80000000
#define BDSTAT_INT_ON_END_MASK  0x40000000
#define BDSTAT_STOP_ON_END_MASK 0x20000000
#define BDSTAT_COMPLETED_MASK   0x10000000
#define BDSTAT_SOP_MASK         0x08000000
#define BDSTAT_EOP_MASK         0x04000000
#define BDSTAT_CHANBUSY_MASK    0x02000000
#define BDSTAT_CHANRESET_MASK   0x01000000

#define TEMAC_MAC_ADDR_SIZE 6
#define TEMAC_MTU       1500
#define TEMAC_JUMBO_MTU     9000
#define TEMAC_HDR_SIZE      14
#define TEMAC_HDR_VLAN_SIZE 18
#define TEMAC_TRL_SIZE      4
#define TEMAC_MAX_FRAME_SIZE        (TEMAC_MTU + TEMAC_HDR_SIZE + TEMAC_
TRL_SIZE)
#define TEMAC_MAX_VLAN_FRAME_SIZE   (TEMAC_MTU + TEMAC_HDR_VLAN_SIZE + 
TEMAC_TRL_SIZE)
#define TEMAC_MAX_JUMBO_FRAME_SIZE  (TEMAC_JUMBO_MTU + TEMAC_HDR_SIZE + 
TEMAC_TRL_SIZE)

#define TX_CONTROL_CALC_CSUM_MASK   1

#define SDMA_MASTER_IRQ     (1 << 7)
#define SDMA_COA_IRQ        (1)
#define SDMA_DLY_IRQ        (1 << 1)
#define SDMA_ERR_IRQ        (1 << 2)

#define ALIGNMENT       32
#define BUFFER_ALIGN(adr) ((ALIGNMENT - ((u32) adr)) % ALIGNMENT)

#define MULTICAST_CAM_TABLE_NUM 4

#define TX_BD_NUM   64
#define RX_BD_NUM   128

#define XEM_MAX_FRAME_SIZE  TEMAC_MAX_JUMBO_FRAME_SIZE

#define XILINX_GSRD3_NAPI

#define sdma_reg_write(dev, offset, value)  (*(unsigned int *)(dev->
sdma_reg_base + offset) = value)
#define sdma_reg_read(dev, offset)  (*(volatile unsigned int *)(dev->
sdma_reg_base + offset))
#define temac_reg_write(dev, offset, value) (*(unsigned int *)(dev->
temac_reg_base + offset) = value)
#define temac_reg_read(dev, offset) (*(volatile unsigned int *)(dev->
temac_reg_base + offset))

struct net_local {
    struct net_device_stats stats;
    struct net_device *next_dev;
    int index;
    unsigned int sdma_reg_base;
    unsigned int temac_reg_base;
    int tx_irq;
    int rx_irq;
    struct timer_list phy_timer;
};

typedef struct cdmac_bd_t {
    struct cdmac_bd_t *next_p;
    unsigned char *phys_buf_p;
    unsigned long buf_len;
    unsigned long app0;
    unsigned long app1;
    unsigned long app2;
    unsigned long app3;
    unsigned long app4;
} cdmac_bd ;

typedef struct cdmac_tx_bd_t {
    cdmac_bd tx_bd[TX_BD_NUM];
} cdmac_tx_bd ;

typedef struct cdmac_rx_bd_t {
    cdmac_bd rx_bd[RX_BD_NUM];
} cdmac_rx_bd ;

static cdmac_tx_bd *cdmac_tx_bd_virt_p;
static cdmac_rx_bd *cdmac_rx_bd_virt_p;
static cdmac_tx_bd *cdmac_tx_bd_phys_p;
static cdmac_rx_bd *cdmac_rx_bd_phys_p;

static struct sk_buff *rx_skb[RX_BD_NUM];

static volatile int cur_tx_bd =  0;
static volatile int next_tx_bd = 0;
static volatile int tail_tx_bd = 0;
static int cur_rx_bd = 0;

static struct net_device *dev_list = NULL;
static spinlock_t dev_lock;
static spinlock_t rcv_lock;
static spinlock_t xmt_lock;

static int xps_ll_temac_xmit(struct sk_buff *skb, struct net_device *dev)
;
static struct net_device_stats *xps_ll_temac_get_stats(struct net_device 
*dev);

static void xps_ll_temac_hostif_set(struct net_device *dev, int emac, 
int phy_addr, int reg_addr, int phy_data)
{
    struct net_local *lp = (struct net_local *)dev->priv;

    temac_reg_write(lp, TEMAC_LSW0, phy_data);
    temac_reg_write(lp, TEMAC_CTL0, CNTLREG_WRITE_ENABLE_MASK | MIIMWD);
    temac_reg_write(lp, TEMAC_LSW0, ((phy_addr << 5) | (reg_addr)));
    temac_reg_write(lp, TEMAC_CTL0, CNTLREG_WRITE_ENABLE_MASK | MIIMAI | 
(emac << 10));
    while(!(temac_reg_read(lp, TEMAC_RDY0) & XTE_RSE_MIIM_WR_MASK));
}

static unsigned int xps_ll_temac_hostif_get(struct net_device *dev, int 
emac, int phy_addr, int reg_addr)
{
    struct net_local *lp = (struct net_local *)dev->priv;

    temac_reg_write(lp, TEMAC_LSW0, ((phy_addr << 5) | (reg_addr)));
    temac_reg_write(lp, TEMAC_CTL0, MIIMAI | (emac << 10));

    while(!(temac_reg_read(lp, TEMAC_RDY0) & XTE_RSE_MIIM_RR_MASK));
    return temac_reg_read(lp, TEMAC_LSW0);
}

static void xps_ll_temac_indirect_set(struct net_device *dev, int emac, 
int reg_offset, int reg_data)
{
    struct net_local *lp = (struct net_local *)dev->priv;

    temac_reg_write(lp, TEMAC_LSW0, reg_data);
    temac_reg_write(lp, TEMAC_CTL0, (CNTLREG_WRITE_ENABLE_MASK | (emac <
< 10) | reg_offset));

    while(!(temac_reg_read(lp, TEMAC_RDY0) & XTE_RSE_CFG_WR_MASK));
}

static void xps_ll_temac_phy_ctrl(struct net_device *dev)
{
    unsigned int result;

    result = xps_ll_temac_hostif_get(dev, 0, EMAC_PHY_ID, 10);
    if((result & 0x0800) == 0x0800) {
        xps_ll_temac_indirect_set(dev, 0, EMMC, 0x80000000);
        printk("1000BASE-T/FD\n");
        return;
    }
    result = xps_ll_temac_hostif_get(dev, 0, EMAC_PHY_ID, 5);
    if((result & 0x0100) == 0x0100) {
        xps_ll_temac_indirect_set(dev, 0, EMMC, 0x40000000);
        printk("100BASE-T/FD\n");
    } else if((result & 0x0040) == 0x0040) {
        xps_ll_temac_indirect_set(dev, 0, EMMC, 0x00000000);
        printk("10BASE-T/FD\n");
    } else {
        printk("Half Duplex not supported\n");
    }
}

static int xps_ll_temac_set_address(struct net_device *dev, void *p)
{
    struct sockaddr *sa = p;

    memcpy(dev->dev_addr, sa->sa_data, ETH_ALEN);

    xps_ll_temac_indirect_set(dev, 0, UAW0, (dev->dev_addr[0] & 
0x000000ff) |
        ((dev->dev_addr[1] << 8) & 0x0000ff00) |
        ((dev->dev_addr[2] << 16)& 0x00ff0000) |
        ((dev->dev_addr[3] << 24) & 0xff000000));
    xps_ll_temac_indirect_set(dev, 0, UAW1, (dev->dev_addr[4] & 
0x000000ff) |
        ((dev->dev_addr[5] << 8) & 0x0000ff00));

    return 0;
}

static void __set_rx_mode (struct net_device *dev)
{
    unsigned long waddr_msw, waddr_lsw;
    int i;

    if(dev->flags & IFF_PROMISC) {
        printk(KERN_NOTICE "%s: Promiscuos mode enabled.\n", dev->name);

        xps_ll_temac_indirect_set(dev, 0, AFM, 0x80000000);
    }  else {
        struct dev_mc_list *mclist;
        for(i = 0, mclist = dev->mc_list; mclist && i < dev->mc_count;
            i++, mclist = mclist->next) {

            if(i >= MULTICAST_CAM_TABLE_NUM) break;
            waddr_msw = ((mclist->dmi_addr[3] << 24) | (mclist->dmi_
addr[2] << 16) |
                (mclist->dmi_addr[1] << 8) | mclist->dmi_addr[0]);
            xps_ll_temac_indirect_set(dev, 0, MAW0, waddr_msw);
            waddr_lsw = ((mclist->dmi_addr[5] << 8) | mclist->dmi_addr[4]
);
            waddr_lsw |= (i << 16);
            xps_ll_temac_indirect_set(dev, 0, MAW1, waddr_lsw);
        }
    }
}

static void xps_ll_temac_set_rx_mode (struct net_device *dev)
{
    spin_lock(&dev_lock);
    __set_rx_mode(dev);
    spin_unlock(&dev_lock);
}

static int xps_ll_temac_xmit_done(struct net_device *dev)
{
    struct net_local *lp = (struct net_local *)dev->priv;
    cdmac_bd *cur_p;
    unsigned int stat = 0;
    unsigned int flags;

    spin_lock_irqsave(&xmt_lock, flags);

    cur_p = &cdmac_tx_bd_virt_p->tx_bd[cur_tx_bd];
    stat = cur_p->app0;

    while(stat & BDSTAT_COMPLETED_MASK) {
        pci_unmap_single(NULL, (unsigned long)cur_p->phys_buf_p,
            cur_p->buf_len, PCI_DMA_TODEVICE);
        if (cur_p->app4)
            dev_kfree_skb_irq((struct sk_buff *)cur_p->app4);
        cur_p->app0 = 0;

        lp->stats.tx_packets++;
        lp->stats.tx_bytes += cur_p->buf_len;

        cur_tx_bd++;
        if (cur_tx_bd >= TX_BD_NUM) cur_tx_bd = 0;

        cur_p = &cdmac_tx_bd_virt_p->tx_bd[cur_tx_bd];
        stat = cur_p->app0;
    }

    spin_unlock_irqrestore(&xmt_lock, flags);

    if(netif_queue_stopped(dev)) {
        netif_wake_queue(dev);
    }
}

static int xps_ll_temac_xmit(struct sk_buff *skb, struct net_device *dev)
{
    struct net_local *lp = (struct net_local *)dev->priv;
    cdmac_bd *cur_p, *start_p, *tail_p;
    int i;
    unsigned long num_frag;
    skb_frag_t *frag;

    spin_lock(&xmt_lock);

    num_frag = skb_shinfo(skb)->nr_frags;
    frag = &skb_shinfo(skb)->frags[0];
    start_p = &cdmac_tx_bd_phys_p->tx_bd[tail_tx_bd];
    cur_p = &cdmac_tx_bd_virt_p->tx_bd[tail_tx_bd];

    if(cur_p->app0 & BDSTAT_COMPLETED_MASK) {
        if(!netif_queue_stopped(dev)) {
            netif_stop_queue(dev);
            spin_unlock(&xmt_lock);
            return NETDEV_TX_BUSY;
        }
        return NETDEV_TX_BUSY;
    }

    cur_p->app0 = 0;
    if(skb->ip_summed == CHECKSUM_PARTIAL) {
        const struct iphdr *ip = skb->nh.iph;
        int length, start, insert, headlen;

        switch(ip->protocol) {
        case IPPROTO_TCP:
            start = sizeof(struct iphdr) + ETH_HLEN;
            insert = sizeof(struct iphdr) + ETH_HLEN + 16;
            length = ip->tot_len - sizeof(struct iphdr);
            headlen = ETH_HLEN + sizeof(struct iphdr) + sizeof(struct 
tcphdr);
            break;
        case IPPROTO_UDP:
            start = sizeof(struct iphdr) + ETH_HLEN;
            insert = sizeof(struct iphdr) + ETH_HLEN + 6;
            length = ip->tot_len - sizeof(struct iphdr);
            headlen = ETH_HLEN + sizeof(struct iphdr) + sizeof(struct 
udphdr);
            break;
        default:
            break;
        }
        cur_p->app1 = ((start << 16) | insert);
        cur_p->app2 = csum_tcpudp_magic(ip->saddr, ip->daddr, 
            length, ip->protocol, 0);
        skb->data[insert] = 0;
        skb->data[insert + 1] = 0;
    }
    cur_p->app0 |= BDSTAT_SOP_MASK;
    cur_p->buf_len = skb_headlen(skb);
    cur_p->phys_buf_p = (unsigned char *)pci_map_single(NULL, skb->data, 
skb->len, PCI_DMA_TODEVICE);
    cur_p->app4 = (unsigned long)skb;

    for(i = 0;i < num_frag;i++) {
        tail_tx_bd++;
        if (tail_tx_bd >= TX_BD_NUM) tail_tx_bd = 0;

        cur_p = &cdmac_tx_bd_virt_p->tx_bd[tail_tx_bd];
        cur_p->phys_buf_p = (unsigned char *)pci_map_single(NULL,
            (void *)page_address(frag->page) + frag->page_offset, frag->
size, PCI_DMA_TODEVICE);
        cur_p->buf_len = frag->size;
        cur_p->app0 = 0;
        frag++;
    }
    cur_p->app0 |= BDSTAT_EOP_MASK;

    tail_p = &cdmac_tx_bd_phys_p->tx_bd[tail_tx_bd];
    tail_tx_bd++;
    if (tail_tx_bd >= TX_BD_NUM) tail_tx_bd = 0;

    if(!(sdma_reg_read(lp, TX_CHNL_STS) & 2)) {
        sdma_reg_write(lp, TX_CURDESC_PTR, start_p);
        sdma_reg_write(lp, TX_TAILDESC_PTR, tail_p);    // DMA start 
    } else {
    }

    spin_unlock(&xmt_lock);

    return 0;
}

static int xps_ll_temac_poll(struct net_device *dev, int *budget)
{
    struct net_local *lp = (struct net_local *)dev->priv;
    struct sk_buff *skb, *new_skb;
    unsigned int bdstat;
    unsigned long align;
    cdmac_bd *cur_p, *tail_p;
    int length;
    unsigned long skb_vaddr;
    unsigned int flags;

    tail_p = &cdmac_rx_bd_phys_p->rx_bd[cur_rx_bd];
    cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];

    bdstat = cur_p->app0;
    while((bdstat & BDSTAT_COMPLETED_MASK) && *budget > 0) {

        skb = rx_skb[cur_rx_bd];
        length = cur_p->app4;

        skb_vaddr = virt_to_bus(skb->data);
        pci_unmap_single(NULL, skb_vaddr, length, PCI_DMA_FROMDEVICE);

        skb_put(skb, length);
        skb->dev = dev;
        skb->protocol = eth_type_trans(skb, dev);
        skb->ip_summed = CHECKSUM_NONE;

        netif_receive_skb(skb);
        dev->last_rx = jiffies;

        lp->stats.rx_packets++;
        lp->stats.rx_bytes += length;

        new_skb = alloc_skb(XEM_MAX_FRAME_SIZE + ALIGNMENT, GFP_ATOMIC);
        if(new_skb == 0) {
            printk("no memory for new sk_buff\n");
            spin_unlock_irqrestore(&rcv_lock, flags);
            return -ENOMEM;
        }

        align = BUFFER_ALIGN(new_skb->data);
        if(align) skb_reserve(new_skb, align);

        cur_p->app0 = BDSTAT_INT_ON_END_MASK;
        cur_p->phys_buf_p = (unsigned char *)
                pci_map_single(NULL, new_skb->data,
                           XEM_MAX_FRAME_SIZE,
                           PCI_DMA_FROMDEVICE);
        cur_p->buf_len    = XEM_MAX_FRAME_SIZE;
        rx_skb[cur_rx_bd] = new_skb;

        cur_rx_bd++;
        if(cur_rx_bd >= RX_BD_NUM) cur_rx_bd = 0;

        cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];
        bdstat = cur_p->app0 ;
        (*budget)--;
    }

    netif_rx_complete(dev);
    sdma_reg_write(lp, RX_CHNL_CTRL,
        sdma_reg_read(lp, RX_CHNL_CTRL) | SDMA_MASTER_IRQ);
    sdma_reg_write(lp, RX_TAILDESC_PTR, tail_p);

    return 0;
}

static void xps_ll_temac_recv(struct net_device *dev)
{
    struct net_local *lp = (struct net_local *)dev->priv;
    struct sk_buff *skb, *new_skb;
    unsigned int bdstat;
    unsigned long align;
    cdmac_bd *cur_p, *tail_p;
    int length;
    unsigned long skb_vaddr;
    unsigned int flags;

    spin_lock_irqsave(&rcv_lock, flags);

    tail_p = &cdmac_rx_bd_phys_p->rx_bd[cur_rx_bd];
    cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];

    bdstat = cur_p->app0;
    while((bdstat & BDSTAT_COMPLETED_MASK)) {

        skb = rx_skb[cur_rx_bd];
        length = cur_p->app4;

        skb_vaddr = virt_to_bus(skb->data);
        pci_unmap_single(NULL, skb_vaddr, length, PCI_DMA_FROMDEVICE);

        skb_put(skb, length);
        skb->dev = dev;
        skb->protocol = eth_type_trans(skb, dev);
        skb->ip_summed = CHECKSUM_NONE;

        netif_rx(skb);

        lp->stats.rx_packets++;
        lp->stats.rx_bytes += length;

        new_skb = alloc_skb(XEM_MAX_FRAME_SIZE + ALIGNMENT, GFP_ATOMIC);
        if(new_skb == 0) {
            printk("no memory for new sk_buff\n");
            spin_unlock_irqrestore(&rcv_lock, flags);
            return;
        }

        align = BUFFER_ALIGN(new_skb->data);
        if(align) skb_reserve(new_skb, align);

        cur_p->app0 = BDSTAT_INT_ON_END_MASK;
        cur_p->phys_buf_p = (unsigned char *)
                    pci_map_single(NULL, new_skb->data,
                               XEM_MAX_FRAME_SIZE,
                               PCI_DMA_FROMDEVICE);
        cur_p->buf_len    = XEM_MAX_FRAME_SIZE;
        rx_skb[cur_rx_bd] = new_skb;

        cur_rx_bd++;
        if(cur_rx_bd >= RX_BD_NUM) cur_rx_bd = 0;

        cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];
        bdstat = cur_p->app0;
    }
    sdma_reg_write(lp, RX_TAILDESC_PTR, tail_p);

    spin_unlock_irqrestore(&rcv_lock, flags);
}

static irqreturn_t
xps_ll_temac_tx_int(int irq, void * dev_id, struct pt_regs *regs)
{
    unsigned int status;
    struct net_device *dev = (struct net_device *)dev_id;
    struct net_local *lp = (struct net_local *)dev->priv;

    status = sdma_reg_read(lp, TX_IRQ_REG);
    sdma_reg_write(lp, TX_IRQ_REG, status);

    if(status & 3) xps_ll_temac_xmit_done(dev);
    if(status & 0x080) printk("DMA error 0x%x\n", status);

    return IRQ_HANDLED;
}

static irqreturn_t
xps_ll_temac_rx_int(int irq, void * dev_id, struct pt_regs *regs)
{
    unsigned int status;
    struct net_device *dev = (struct net_device *)dev_id;
    struct net_local *lp = (struct net_local *)dev->priv;

    status = sdma_reg_read(lp, RX_IRQ_REG);
    sdma_reg_write(lp, RX_IRQ_REG, status);

#ifdef XPS_LL_TEMAC_NAPI
    if(status & 1) {
        if(likely(netif_rx_schedule_prep(dev))) {
            sdma_reg_write(lp, RX_CHNL_CTRL,
                sdma_reg_read(lp, RX_CHNL_CTRL) & ~SDMA_MASTER_IRQ);
            __netif_rx_schedule(dev);
        }
    }
#else
    if(status & 3) xps_ll_temac_recv(dev);
#endif

    return IRQ_HANDLED;
}

static void xps_ll_temac_netpoll(struct net_device *dev)
{
    //struct net_local *lp = (struct net_local *)dev->priv;

    disable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR);
    disable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR);

    xps_ll_temac_rx_int(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR, 
dev, 0);
    xps_ll_temac_tx_int(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR, 
dev, 0);

    enable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR);
    enable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR);
}

static struct net_device_stats *xps_ll_temac_get_stats(struct net_device 
*dev)
{
    return netdev_priv(dev);
}

static int xps_ll_temac_init_descriptor(void)
{
    struct sk_buff *skb;
    unsigned long align;
    int i;

    cdmac_tx_bd_virt_p = dma_alloc_coherent(NULL, sizeof(struct cdmac_
tx_bd_t),
        (dma_addr_t *)&cdmac_tx_bd_phys_p, GFP_KERNEL);
    cdmac_rx_bd_virt_p = dma_alloc_coherent(NULL, sizeof(struct cdmac_
rx_bd_t),
        (dma_addr_t *)&cdmac_rx_bd_phys_p, GFP_KERNEL);

    for(i = 0;i < TX_BD_NUM;i++) {
        memset((char *)&cdmac_tx_bd_virt_p->tx_bd[i], 0, sizeof(struct 
cdmac_bd_t));
        if(i == (TX_BD_NUM - 1)) {
            cdmac_tx_bd_virt_p->tx_bd[i].next_p =
                &cdmac_tx_bd_phys_p->tx_bd[0];
        } else {
            cdmac_tx_bd_virt_p->tx_bd[i].next_p =
                &cdmac_tx_bd_phys_p->tx_bd[i + 1];
        }
    }
    for(i = 0;i < RX_BD_NUM;i++) {
        memset((char *)&cdmac_rx_bd_virt_p->rx_bd[i], 0, sizeof(struct 
cdmac_bd_t));
        if(i == (RX_BD_NUM - 1)) {
            cdmac_rx_bd_virt_p->rx_bd[i].next_p =
                &cdmac_rx_bd_phys_p->rx_bd[0];
        } else {
            cdmac_rx_bd_virt_p->rx_bd[i].next_p =
                &cdmac_rx_bd_phys_p->rx_bd[i + 1];
        }
        skb = alloc_skb(XEM_MAX_FRAME_SIZE + ALIGNMENT, GFP_ATOMIC);
        if(skb == 0) {
            printk("alloc_skb error %d\n", i);
            return -1;
        }
        rx_skb[i] = skb;
        align = BUFFER_ALIGN(skb->data);
        if(align) skb_reserve(skb, align);

        cdmac_rx_bd_virt_p->rx_bd[i].phys_buf_p =
            (unsigned char *)pci_map_single(NULL,
                skb->data, XEM_MAX_FRAME_SIZE,
                PCI_DMA_FROMDEVICE);
        cdmac_rx_bd_virt_p->rx_bd[i].buf_len = XEM_MAX_FRAME_SIZE;
        cdmac_rx_bd_virt_p->rx_bd[i].app0 = BDSTAT_INT_ON_END_MASK;
    }

    return 0;
}

static int xps_ll_temac_changemtu(struct net_device *dev, int newmtu)
{
        printk("[xilinx_enet]new MTU %d\n", newmtu);
        dev->mtu = newmtu;

        return 0;
}

static int xps_ll_temac_open(struct net_device *dev)
{

    return 0;
}

static int xps_ll_temac_close(struct net_device *dev)
{

    return 0;
}

static struct net_device **xps_ll_temacs;

static int __init xps_ll_temac_init_one(int index)
{
    struct net_device *dev = alloc_etherdev(sizeof(struct net_local));
    struct net_local *lp;
    int err = 0;

    if (!dev) return -ENOMEM;

    sprintf(dev->name, "eth%d", index);

    /* Initialize the device structure. */
    dev->get_stats = xps_ll_temac_get_stats;
    dev->hard_start_xmit = xps_ll_temac_xmit;
    dev->open = xps_ll_temac_open;
    dev->stop = xps_ll_temac_close;
#ifdef XPS_LL_TEMAC_NAPI
    dev->poll = &xps_ll_temac_poll;
#endif
    dev->weight = 64;
    dev->set_multicast_list = &xps_ll_temac_set_rx_mode;
    dev->set_mac_address = xps_ll_temac_set_address;
    dev->poll_controller = xps_ll_temac_netpoll;

    SET_MODULE_OWNER(dev);
#if 1
//  dev->features = NETIF_F_SG | NETIF_F_FRAGLIST
//      | NETIF_F_IP_CSUM;
    dev->features = NETIF_F_SG | NETIF_F_FRAGLIST;
#else
    dev->features = NETIF_F_SG | NETIF_F_FRAGLIST
        | NETIF_F_TSO | NETIF_F_IP_CSUM;
#endif

    ether_setup(dev);

    lp = (struct net_local *)dev->priv;
    memset(lp, 0, sizeof(struct net_local));
    lp->index = index;

    lp->next_dev = dev_list;
    lp->sdma_reg_base = ioremap(S_DMA_CTRL_BASEADDR, 128);
    lp->temac_reg_base = ioremap(XPS_LLTEMAC_BASEADDR, 128);

    dev_list = dev;

    dev->tx_queue_len = 0;
    dev->change_mtu = xps_ll_temac_changemtu;

    sdma_reg_write(lp, DMA_CONTROL_REG, 1);

    printk(KERN_INFO "%s: Xilinx Embedded Tri-Mode Ethernet MAC\n", dev-
>name);
    xps_ll_temac_init_descriptor();

    dev->dev_addr[0] = 0x00;
    dev->dev_addr[1] = 0x80;
    dev->dev_addr[2] = 0x49;
    dev->dev_addr[3] = 0x00;
    dev->dev_addr[4] = 0x00;
    dev->dev_addr[5] = 0x00;

    request_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR,
        &xps_ll_temac_tx_int, 0, dev->name, dev);
    request_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR,
        &xps_ll_temac_rx_int, 0, dev->name, dev);

    xps_ll_temac_indirect_set(dev, 0, MC, MDIO_ENABLE_MASK | MDIO_CLOCK_
DIV_100MHz);

    xps_ll_temac_indirect_set(dev, 0, RCW1, 0x10000000);    // Enable 
Receiver
    xps_ll_temac_indirect_set(dev, 0, TC, 0x10000000);  // Enable 
Transmitter
    xps_ll_temac_indirect_set(dev, 0, EMMC, 0x84000000);

    xps_ll_temac_indirect_set(dev, 0, UAW0, (dev->dev_addr[0] & 
0x000000ff) |
        ((dev->dev_addr[1] << 8) & 0x0000ff00) |
        ((dev->dev_addr[2] << 16)& 0x00ff0000) |
        ((dev->dev_addr[3] << 24) & 0xff000000));
    xps_ll_temac_indirect_set(dev, 0, UAW1, (dev->dev_addr[4] & 
0x000000ff) |
        ((dev->dev_addr[5] << 8) & 0x0000ff00));

    xps_ll_temac_indirect_set(dev, 0, AFM, 0x00000000);

    xps_ll_temac_phy_ctrl(dev);

    sdma_reg_write(lp, TX_CHNL_CTRL, 0x10220483);
    //sdma_reg_write(lp, TX_CHNL_CTRL, 0x00100483);
    sdma_reg_write(lp, RX_CHNL_CTRL, 0xff010283);

    sdma_reg_write(lp, RX_CURDESC_PTR, (unsigned int)&cdmac_rx_bd_phys_
p->rx_bd[0]);
    sdma_reg_write(lp, RX_TAILDESC_PTR, (unsigned int)&cdmac_rx_bd_phys_
p->rx_bd[RX_BD_NUM - 1]);

    if ((err = register_netdev(dev))) {
        free_netdev(dev);
        dev = NULL;
    } else {
        xps_ll_temacs[index] = dev;
    }

    return 0;
}

static void xps_ll_temac_free_one(int index)
{
    unregister_netdev(xps_ll_temacs[index]);
    free_netdev(xps_ll_temacs[index]);
}

static int __init xps_ll_temac_init_module(void)
{
    int err = 0;

    xps_ll_temacs = kmalloc(sizeof(void *), GFP_KERNEL);

    if (!xps_ll_temacs) return -ENOMEM;

    spin_lock_init(&dev_lock);
    spin_lock_init(&rcv_lock);
    spin_lock_init(&xmt_lock);

    if((err = xps_ll_temac_init_one(0))) xps_ll_temac_free_one(0);

    return err;
}

static void __exit xps_ll_temac_cleanup_module(void)
{
    xps_ll_temac_free_one(0);
    kfree(xps_ll_temacs);
}

module_init(xps_ll_temac_init_module);
module_exit(xps_ll_temac_cleanup_module);
MODULE_LICENSE("GPL");


> Thanks,
> 
>     I have alot of work to do on our stuff, I might as well see if I 
can
> move to the powerpc tree at the same time.
> 
>     BTW is there even the beginings of a non-xilinx lltemac driver out
> there ? There were hints on the list, but I have not seen anything.
> 
>     I would be happy to help advance the ball on anything anyone has
> started.
> 
>    
> 
> Grant Likely wrote:
> > On Sun, Apr 20, 2008 at 2:31 PM, David H. Lynch Jr. <dhlii@dlasys.net
> wrote:
> >   
> >> Thanks.
> >>
> >>     I am running linus's 2.6.25-rc9, but I can pull the Xilinx tree 
or
> >>  yours - when your server is up.
> >>
> >>     I can not find any Xilinx powerpc configs in arch/powerpc/
config
> >>     Do I just need to do a
> >>     make ARCH=powerpc menuconfig and create one from scratch ?
> >>     
> >
> > That's right; I haven't merged any defconfigs.  Roll your own.
> >
> >   
> >>     Is simpleboot in your tree or the xilinx tree, if  can not find 
it
> >>  in Linus's ?
> >>     
> >
> > If you want to use the "simpleboot" wrapper; then you'll need to 
pull
> > paulus' tree (Linus hasn't yet pulled his tree; but he probably will
> > any moment now).
> >
> > Cheers,
> > g.
> >
> >   
> 
> 
> -- 
> Dave Lynch                            DLA Systems
> Software Development:                          Embedded Linux
> 717.627.3770         dhlii@dlasys.net       http://www.dlasys.net
> fax: 1.253.369.9244                      Cell: 1.717.587.7774
> Over 25 years' experience in platforms, languages, and technologies 
too numerous to list.
> 
> "Any intelligent fool can make things bigger and more complex... It 
takes a touch of genius - and a lot of courage to move in the opposite 
direction."
> Albert Einstein
> 
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> 
※ ４月から所属・部署が変更になりました
------------------------------------------------------------------
柏木良夫
株式会社日新システムズ 東日本営業部

本      社 〒600-8482 京都市下京区堀川通四条下ル東側 堀川四条ビル
TEL 075-344-7977 FAX 075-344-7887
東京事務所 〒101-0024 東京都千代田区神田和泉町1番地 神田和泉町ビル
TEL 03-5825-2081 FAX 03-5821-1259
E-Mail kashiwagi@co-nss.co.jp HTTP http://www.co-nss.co.jp/
------------------------------------------------------------------

Re: Xilinx PowerPC

[David H. Lynch Jr.]

 Thanks alot.
On quick review this looks great. I will see if I can get it working for
me this weekend.


Yoshio Kashiwagi wrote:
> Hi,
>
> I am writing the Non-Xilinx XPS_LL_TEMAC driver.
> Checksum offloading is incomplete although NAPI and KGDBOE are supported.
> Basic operation is working on EDK9.2 and EDK10.1.
>
> Furthermore, although the simple Non-Interrupt version for u-boot is
> also written, it is not known where I should post.
>
> Best Regards,
>
> Yoshio Kashiwagi - Nissin Systems

-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein

Re: Xilinx PowerPC

[David H. Lynch Jr.]

 I was able to incorporate Yoshio's driver and make use of it.
It is a single file, fairly short, easy to understand, mostly conforming
to linux kernel programming style. While it needs some work fo NAPI and
OF, it is usable as it is.

I would be happy to see it form the starting point for a distribution LL
TEMAC driver.

If Yoshio does not wish to put together a patch, for it I can, however,
I have made some name changes and restructing to compare to my older PLB
TEMAC.




Yoshio Kashiwagi wrote:
> Hi,
>
> I am writing the Non-Xilinx XPS_LL_TEMAC driver.
> Checksum offloading is incomplete although NAPI and KGDBOE are supported.
> Basic operation is working on EDK9.2 and EDK10.1.
>
> Furthermore, although the simple Non-Interrupt version for u-boot is
> also written, it is not known where I should post.
>
> Best Regards,
>
> Yoshio Kashiwagi - Nissin Systems
>
> /*
>  *
>  * Xilinx Gigabit System Referece Design Ethenet driver
>  *
>  * Driver for Xilinx Virtex-4FX Based Platform
>  *
>  * Author: Yoshio Kashiwagi
>  *
>  * Copyright (c) 2008 Nissin Systems Co.,Ltd.
>  *
>  * March 2008 created
>  *
>  * This program is free software; you can redistribute  it and/or modify 
> it
>  * under  the terms of  the GNU General  Public License as published by 
> the
>  * Free Software Foundation;  either version 2 of the  License, or (at 
> your
>  * option) any later version.
>  *
> */
>
> #include <linux/autoconf.h>
> #include <linux/module.h>
> #include <linux/kernel.h>
> #include <linux/netdevice.h>
> #include <linux/etherdevice.h>
> #include <linux/init.h>
> #include <linux/in.h>
> #include <linux/moduleparam.h>
> #include <linux/spinlock.h>
> #include <linux/pci.h>
> #include <asm/io.h>
> #include <asm/delay.h>
> #include <platforms/4xx/xparameters/xparameters.h>
>
> #include <linux/inet.h>
> #include <linux/netdevice.h>
> #include <linux/etherdevice.h>
> #include <linux/skbuff.h>
> #include <linux/ethtool.h>
> #include <net/sock.h>
> #include <net/checksum.h>
> #include <linux/if_ether.h>
> #include <linux/if_arp.h>
> #include <linux/ip.h>
> #include <linux/tcp.h>
> #include <linux/udp.h>
> #include <linux/percpu.h>
>
> #define S_DMA_CTRL_BASEADDR XPAR_LLTEMAC_0_LLINK_CONNECTED_BASEADDR
> #define XPS_LLTEMAC_BASEADDR    XPAR_LLTEMAC_0_BASEADDR
>
> /* XPS_LL_TEMAC SDMA registers definition */
>
> #define TX_NXTDESC_PTR      0x00
> #define TX_CURBUF_ADDR      0x04
> #define TX_CURBUF_LENGTH    0x08
> #define TX_CURDESC_PTR      0x0c
> #define TX_TAILDESC_PTR     0x10
> #define TX_CHNL_CTRL        0x14
> #define TX_IRQ_REG      0x18
> #define TX_CHNL_STS     0x1c
>
> #define RX_NXTDESC_PTR      0x20
> #define RX_CURBUF_ADDR      0x24
> #define RX_CURBUF_LENGTH    0x28
> #define RX_CURDESC_PTR      0x2c
> #define RX_TAILDESC_PTR     0x30
> #define RX_CHNL_CTRL        0x34
> #define RX_IRQ_REG      0x38
> #define RX_CHNL_STS     0x3c
>
> #define DMA_CONTROL_REG     0x40
>
> /* XPS_LL_TEMAC direct registers definition */
>
> #define TEMAC_RAF0      0x00
> #define TEMAC_TPF0      0x04
> #define TEMAC_IFGP0     0x08
> #define TEMAC_IS0       0x0c
> #define TEMAC_IP0       0x10
> #define TEMAC_IE0       0x14
>
> #define TEMAC_MSW0      0x20
> #define TEMAC_LSW0      0x24
> #define TEMAC_CTL0      0x28
> #define TEMAC_RDY0      0x2c
>
> #define XTE_RSE_MIIM_RR_MASK      0x0002
> #define XTE_RSE_MIIM_WR_MASK      0x0004
> #define XTE_RSE_CFG_RR_MASK       0x0020
> #define XTE_RSE_CFG_WR_MASK       0x0040
>
> /* XPS_LL_TEMAC indirect registers offset definition */
>
> #define RCW0    0x200
> #define RCW1    0x240
> #define TC  0x280
> #define FCC 0x2c0
> #define EMMC    0x300
> #define PHYC    0x320
> #define MC  0x340
> #define UAW0    0x380
> #define UAW1    0x384
> #define MAW0    0x388
> #define MAW1    0x38c
> #define AFM 0x390
> #define TIS 0x3a0
> #define TIE 0x3a4
> #define MIIMWD  0x3b0
> #define MIIMAI  0x3b4
>
> #define CNTLREG_WRITE_ENABLE_MASK   0x8000
> #define CNTLREG_EMAC1SEL_MASK       0x0400
> #define CNTLREG_ADDRESSCODE_MASK    0x03ff
>
> #define MDIO_ENABLE_MASK        0x40
> #define MDIO_CLOCK_DIV_MASK 0x3F
> #define MDIO_CLOCK_DIV_100MHz   0x28
>
> #define ETHER_MTU       1500
> #define EMAC_PHY_ID     7
>
> /* CDMAC descriptor status bit definitions */
>
> #define BDSTAT_ERROR_MASK       0x80000000
> #define BDSTAT_INT_ON_END_MASK  0x40000000
> #define BDSTAT_STOP_ON_END_MASK 0x20000000
> #define BDSTAT_COMPLETED_MASK   0x10000000
> #define BDSTAT_SOP_MASK         0x08000000
> #define BDSTAT_EOP_MASK         0x04000000
> #define BDSTAT_CHANBUSY_MASK    0x02000000
> #define BDSTAT_CHANRESET_MASK   0x01000000
>
> #define TEMAC_MAC_ADDR_SIZE 6
> #define TEMAC_MTU       1500
> #define TEMAC_JUMBO_MTU     9000
> #define TEMAC_HDR_SIZE      14
> #define TEMAC_HDR_VLAN_SIZE 18
> #define TEMAC_TRL_SIZE      4
> #define TEMAC_MAX_FRAME_SIZE        (TEMAC_MTU + TEMAC_HDR_SIZE + TEMAC_
> TRL_SIZE)
> #define TEMAC_MAX_VLAN_FRAME_SIZE   (TEMAC_MTU + TEMAC_HDR_VLAN_SIZE + 
> TEMAC_TRL_SIZE)
> #define TEMAC_MAX_JUMBO_FRAME_SIZE  (TEMAC_JUMBO_MTU + TEMAC_HDR_SIZE + 
> TEMAC_TRL_SIZE)
>
> #define TX_CONTROL_CALC_CSUM_MASK   1
>
> #define SDMA_MASTER_IRQ     (1 << 7)
> #define SDMA_COA_IRQ        (1)
> #define SDMA_DLY_IRQ        (1 << 1)
> #define SDMA_ERR_IRQ        (1 << 2)
>
> #define ALIGNMENT       32
> #define BUFFER_ALIGN(adr) ((ALIGNMENT - ((u32) adr)) % ALIGNMENT)
>
> #define MULTICAST_CAM_TABLE_NUM 4
>
> #define TX_BD_NUM   64
> #define RX_BD_NUM   128
>
> #define XEM_MAX_FRAME_SIZE  TEMAC_MAX_JUMBO_FRAME_SIZE
>
> #define XILINX_GSRD3_NAPI
>
> #define sdma_reg_write(dev, offset, value)  (*(unsigned int *)(dev->
> sdma_reg_base + offset) = value)
> #define sdma_reg_read(dev, offset)  (*(volatile unsigned int *)(dev->
> sdma_reg_base + offset))
> #define temac_reg_write(dev, offset, value) (*(unsigned int *)(dev->
> temac_reg_base + offset) = value)
> #define temac_reg_read(dev, offset) (*(volatile unsigned int *)(dev->
> temac_reg_base + offset))
>
> struct net_local {
>     struct net_device_stats stats;
>     struct net_device *next_dev;
>     int index;
>     unsigned int sdma_reg_base;
>     unsigned int temac_reg_base;
>     int tx_irq;
>     int rx_irq;
>     struct timer_list phy_timer;
> };
>
> typedef struct cdmac_bd_t {
>     struct cdmac_bd_t *next_p;
>     unsigned char *phys_buf_p;
>     unsigned long buf_len;
>     unsigned long app0;
>     unsigned long app1;
>     unsigned long app2;
>     unsigned long app3;
>     unsigned long app4;
> } cdmac_bd ;
>
> typedef struct cdmac_tx_bd_t {
>     cdmac_bd tx_bd[TX_BD_NUM];
> } cdmac_tx_bd ;
>
> typedef struct cdmac_rx_bd_t {
>     cdmac_bd rx_bd[RX_BD_NUM];
> } cdmac_rx_bd ;
>
> static cdmac_tx_bd *cdmac_tx_bd_virt_p;
> static cdmac_rx_bd *cdmac_rx_bd_virt_p;
> static cdmac_tx_bd *cdmac_tx_bd_phys_p;
> static cdmac_rx_bd *cdmac_rx_bd_phys_p;
>
> static struct sk_buff *rx_skb[RX_BD_NUM];
>
> static volatile int cur_tx_bd =  0;
> static volatile int next_tx_bd = 0;
> static volatile int tail_tx_bd = 0;
> static int cur_rx_bd = 0;
>
> static struct net_device *dev_list = NULL;
> static spinlock_t dev_lock;
> static spinlock_t rcv_lock;
> static spinlock_t xmt_lock;
>
> static int xps_ll_temac_xmit(struct sk_buff *skb, struct net_device *dev)
> ;
> static struct net_device_stats *xps_ll_temac_get_stats(struct net_device 
> *dev);
>
> static void xps_ll_temac_hostif_set(struct net_device *dev, int emac, 
> int phy_addr, int reg_addr, int phy_data)
> {
>     struct net_local *lp = (struct net_local *)dev->priv;
>
>     temac_reg_write(lp, TEMAC_LSW0, phy_data);
>     temac_reg_write(lp, TEMAC_CTL0, CNTLREG_WRITE_ENABLE_MASK | MIIMWD);
>     temac_reg_write(lp, TEMAC_LSW0, ((phy_addr << 5) | (reg_addr)));
>     temac_reg_write(lp, TEMAC_CTL0, CNTLREG_WRITE_ENABLE_MASK | MIIMAI | 
> (emac << 10));
>     while(!(temac_reg_read(lp, TEMAC_RDY0) & XTE_RSE_MIIM_WR_MASK));
> }
>
> static unsigned int xps_ll_temac_hostif_get(struct net_device *dev, int 
> emac, int phy_addr, int reg_addr)
> {
>     struct net_local *lp = (struct net_local *)dev->priv;
>
>     temac_reg_write(lp, TEMAC_LSW0, ((phy_addr << 5) | (reg_addr)));
>     temac_reg_write(lp, TEMAC_CTL0, MIIMAI | (emac << 10));
>
>     while(!(temac_reg_read(lp, TEMAC_RDY0) & XTE_RSE_MIIM_RR_MASK));
>     return temac_reg_read(lp, TEMAC_LSW0);
> }
>
> static void xps_ll_temac_indirect_set(struct net_device *dev, int emac, 
> int reg_offset, int reg_data)
> {
>     struct net_local *lp = (struct net_local *)dev->priv;
>
>     temac_reg_write(lp, TEMAC_LSW0, reg_data);
>     temac_reg_write(lp, TEMAC_CTL0, (CNTLREG_WRITE_ENABLE_MASK | (emac <
> < 10) | reg_offset));
>
>     while(!(temac_reg_read(lp, TEMAC_RDY0) & XTE_RSE_CFG_WR_MASK));
> }
>
> static void xps_ll_temac_phy_ctrl(struct net_device *dev)
> {
>     unsigned int result;
>
>     result = xps_ll_temac_hostif_get(dev, 0, EMAC_PHY_ID, 10);
>     if((result & 0x0800) == 0x0800) {
>         xps_ll_temac_indirect_set(dev, 0, EMMC, 0x80000000);
>         printk("1000BASE-T/FD\n");
>         return;
>     }
>     result = xps_ll_temac_hostif_get(dev, 0, EMAC_PHY_ID, 5);
>     if((result & 0x0100) == 0x0100) {
>         xps_ll_temac_indirect_set(dev, 0, EMMC, 0x40000000);
>         printk("100BASE-T/FD\n");
>     } else if((result & 0x0040) == 0x0040) {
>         xps_ll_temac_indirect_set(dev, 0, EMMC, 0x00000000);
>         printk("10BASE-T/FD\n");
>     } else {
>         printk("Half Duplex not supported\n");
>     }
> }
>
> static int xps_ll_temac_set_address(struct net_device *dev, void *p)
> {
>     struct sockaddr *sa = p;
>
>     memcpy(dev->dev_addr, sa->sa_data, ETH_ALEN);
>
>     xps_ll_temac_indirect_set(dev, 0, UAW0, (dev->dev_addr[0] & 
> 0x000000ff) |
>         ((dev->dev_addr[1] << 8) & 0x0000ff00) |
>         ((dev->dev_addr[2] << 16)& 0x00ff0000) |
>         ((dev->dev_addr[3] << 24) & 0xff000000));
>     xps_ll_temac_indirect_set(dev, 0, UAW1, (dev->dev_addr[4] & 
> 0x000000ff) |
>         ((dev->dev_addr[5] << 8) & 0x0000ff00));
>
>     return 0;
> }
>
> static void __set_rx_mode (struct net_device *dev)
> {
>     unsigned long waddr_msw, waddr_lsw;
>     int i;
>
>     if(dev->flags & IFF_PROMISC) {
>         printk(KERN_NOTICE "%s: Promiscuos mode enabled.\n", dev->name);
>
>         xps_ll_temac_indirect_set(dev, 0, AFM, 0x80000000);
>     }  else {
>         struct dev_mc_list *mclist;
>         for(i = 0, mclist = dev->mc_list; mclist && i < dev->mc_count;
>             i++, mclist = mclist->next) {
>
>             if(i >= MULTICAST_CAM_TABLE_NUM) break;
>             waddr_msw = ((mclist->dmi_addr[3] << 24) | (mclist->dmi_
> addr[2] << 16) |
>                 (mclist->dmi_addr[1] << 8) | mclist->dmi_addr[0]);
>             xps_ll_temac_indirect_set(dev, 0, MAW0, waddr_msw);
>             waddr_lsw = ((mclist->dmi_addr[5] << 8) | mclist->dmi_addr[4]
> );
>             waddr_lsw |= (i << 16);
>             xps_ll_temac_indirect_set(dev, 0, MAW1, waddr_lsw);
>         }
>     }
> }
>
> static void xps_ll_temac_set_rx_mode (struct net_device *dev)
> {
>     spin_lock(&dev_lock);
>     __set_rx_mode(dev);
>     spin_unlock(&dev_lock);
> }
>
> static int xps_ll_temac_xmit_done(struct net_device *dev)
> {
>     struct net_local *lp = (struct net_local *)dev->priv;
>     cdmac_bd *cur_p;
>     unsigned int stat = 0;
>     unsigned int flags;
>
>     spin_lock_irqsave(&xmt_lock, flags);
>
>     cur_p = &cdmac_tx_bd_virt_p->tx_bd[cur_tx_bd];
>     stat = cur_p->app0;
>
>     while(stat & BDSTAT_COMPLETED_MASK) {
>         pci_unmap_single(NULL, (unsigned long)cur_p->phys_buf_p,
>             cur_p->buf_len, PCI_DMA_TODEVICE);
>         if (cur_p->app4)
>             dev_kfree_skb_irq((struct sk_buff *)cur_p->app4);
>         cur_p->app0 = 0;
>
>         lp->stats.tx_packets++;
>         lp->stats.tx_bytes += cur_p->buf_len;
>
>         cur_tx_bd++;
>         if (cur_tx_bd >= TX_BD_NUM) cur_tx_bd = 0;
>
>         cur_p = &cdmac_tx_bd_virt_p->tx_bd[cur_tx_bd];
>         stat = cur_p->app0;
>     }
>
>     spin_unlock_irqrestore(&xmt_lock, flags);
>
>     if(netif_queue_stopped(dev)) {
>         netif_wake_queue(dev);
>     }
> }
>
> static int xps_ll_temac_xmit(struct sk_buff *skb, struct net_device *dev)
> {
>     struct net_local *lp = (struct net_local *)dev->priv;
>     cdmac_bd *cur_p, *start_p, *tail_p;
>     int i;
>     unsigned long num_frag;
>     skb_frag_t *frag;
>
>     spin_lock(&xmt_lock);
>
>     num_frag = skb_shinfo(skb)->nr_frags;
>     frag = &skb_shinfo(skb)->frags[0];
>     start_p = &cdmac_tx_bd_phys_p->tx_bd[tail_tx_bd];
>     cur_p = &cdmac_tx_bd_virt_p->tx_bd[tail_tx_bd];
>
>     if(cur_p->app0 & BDSTAT_COMPLETED_MASK) {
>         if(!netif_queue_stopped(dev)) {
>             netif_stop_queue(dev);
>             spin_unlock(&xmt_lock);
>             return NETDEV_TX_BUSY;
>         }
>         return NETDEV_TX_BUSY;
>     }
>
>     cur_p->app0 = 0;
>     if(skb->ip_summed == CHECKSUM_PARTIAL) {
>         const struct iphdr *ip = skb->nh.iph;
>         int length, start, insert, headlen;
>
>         switch(ip->protocol) {
>         case IPPROTO_TCP:
>             start = sizeof(struct iphdr) + ETH_HLEN;
>             insert = sizeof(struct iphdr) + ETH_HLEN + 16;
>             length = ip->tot_len - sizeof(struct iphdr);
>             headlen = ETH_HLEN + sizeof(struct iphdr) + sizeof(struct 
> tcphdr);
>             break;
>         case IPPROTO_UDP:
>             start = sizeof(struct iphdr) + ETH_HLEN;
>             insert = sizeof(struct iphdr) + ETH_HLEN + 6;
>             length = ip->tot_len - sizeof(struct iphdr);
>             headlen = ETH_HLEN + sizeof(struct iphdr) + sizeof(struct 
> udphdr);
>             break;
>         default:
>             break;
>         }
>         cur_p->app1 = ((start << 16) | insert);
>         cur_p->app2 = csum_tcpudp_magic(ip->saddr, ip->daddr, 
>             length, ip->protocol, 0);
>         skb->data[insert] = 0;
>         skb->data[insert + 1] = 0;
>     }
>     cur_p->app0 |= BDSTAT_SOP_MASK;
>     cur_p->buf_len = skb_headlen(skb);
>     cur_p->phys_buf_p = (unsigned char *)pci_map_single(NULL, skb->data, 
> skb->len, PCI_DMA_TODEVICE);
>     cur_p->app4 = (unsigned long)skb;
>
>     for(i = 0;i < num_frag;i++) {
>         tail_tx_bd++;
>         if (tail_tx_bd >= TX_BD_NUM) tail_tx_bd = 0;
>
>         cur_p = &cdmac_tx_bd_virt_p->tx_bd[tail_tx_bd];
>         cur_p->phys_buf_p = (unsigned char *)pci_map_single(NULL,
>             (void *)page_address(frag->page) + frag->page_offset, frag->
> size, PCI_DMA_TODEVICE);
>         cur_p->buf_len = frag->size;
>         cur_p->app0 = 0;
>         frag++;
>     }
>     cur_p->app0 |= BDSTAT_EOP_MASK;
>
>     tail_p = &cdmac_tx_bd_phys_p->tx_bd[tail_tx_bd];
>     tail_tx_bd++;
>     if (tail_tx_bd >= TX_BD_NUM) tail_tx_bd = 0;
>
>     if(!(sdma_reg_read(lp, TX_CHNL_STS) & 2)) {
>         sdma_reg_write(lp, TX_CURDESC_PTR, start_p);
>         sdma_reg_write(lp, TX_TAILDESC_PTR, tail_p);    // DMA start 
>     } else {
>     }
>
>     spin_unlock(&xmt_lock);
>
>     return 0;
> }
>
> static int xps_ll_temac_poll(struct net_device *dev, int *budget)
> {
>     struct net_local *lp = (struct net_local *)dev->priv;
>     struct sk_buff *skb, *new_skb;
>     unsigned int bdstat;
>     unsigned long align;
>     cdmac_bd *cur_p, *tail_p;
>     int length;
>     unsigned long skb_vaddr;
>     unsigned int flags;
>
>     tail_p = &cdmac_rx_bd_phys_p->rx_bd[cur_rx_bd];
>     cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];
>
>     bdstat = cur_p->app0;
>     while((bdstat & BDSTAT_COMPLETED_MASK) && *budget > 0) {
>
>         skb = rx_skb[cur_rx_bd];
>         length = cur_p->app4;
>
>         skb_vaddr = virt_to_bus(skb->data);
>         pci_unmap_single(NULL, skb_vaddr, length, PCI_DMA_FROMDEVICE);
>
>         skb_put(skb, length);
>         skb->dev = dev;
>         skb->protocol = eth_type_trans(skb, dev);
>         skb->ip_summed = CHECKSUM_NONE;
>
>         netif_receive_skb(skb);
>         dev->last_rx = jiffies;
>
>         lp->stats.rx_packets++;
>         lp->stats.rx_bytes += length;
>
>         new_skb = alloc_skb(XEM_MAX_FRAME_SIZE + ALIGNMENT, GFP_ATOMIC);
>         if(new_skb == 0) {
>             printk("no memory for new sk_buff\n");
>             spin_unlock_irqrestore(&rcv_lock, flags);
>             return -ENOMEM;
>         }
>
>         align = BUFFER_ALIGN(new_skb->data);
>         if(align) skb_reserve(new_skb, align);
>
>         cur_p->app0 = BDSTAT_INT_ON_END_MASK;
>         cur_p->phys_buf_p = (unsigned char *)
>                 pci_map_single(NULL, new_skb->data,
>                            XEM_MAX_FRAME_SIZE,
>                            PCI_DMA_FROMDEVICE);
>         cur_p->buf_len    = XEM_MAX_FRAME_SIZE;
>         rx_skb[cur_rx_bd] = new_skb;
>
>         cur_rx_bd++;
>         if(cur_rx_bd >= RX_BD_NUM) cur_rx_bd = 0;
>
>         cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];
>         bdstat = cur_p->app0 ;
>         (*budget)--;
>     }
>
>     netif_rx_complete(dev);
>     sdma_reg_write(lp, RX_CHNL_CTRL,
>         sdma_reg_read(lp, RX_CHNL_CTRL) | SDMA_MASTER_IRQ);
>     sdma_reg_write(lp, RX_TAILDESC_PTR, tail_p);
>
>     return 0;
> }
>
> static void xps_ll_temac_recv(struct net_device *dev)
> {
>     struct net_local *lp = (struct net_local *)dev->priv;
>     struct sk_buff *skb, *new_skb;
>     unsigned int bdstat;
>     unsigned long align;
>     cdmac_bd *cur_p, *tail_p;
>     int length;
>     unsigned long skb_vaddr;
>     unsigned int flags;
>
>     spin_lock_irqsave(&rcv_lock, flags);
>
>     tail_p = &cdmac_rx_bd_phys_p->rx_bd[cur_rx_bd];
>     cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];
>
>     bdstat = cur_p->app0;
>     while((bdstat & BDSTAT_COMPLETED_MASK)) {
>
>         skb = rx_skb[cur_rx_bd];
>         length = cur_p->app4;
>
>         skb_vaddr = virt_to_bus(skb->data);
>         pci_unmap_single(NULL, skb_vaddr, length, PCI_DMA_FROMDEVICE);
>
>         skb_put(skb, length);
>         skb->dev = dev;
>         skb->protocol = eth_type_trans(skb, dev);
>         skb->ip_summed = CHECKSUM_NONE;
>
>         netif_rx(skb);
>
>         lp->stats.rx_packets++;
>         lp->stats.rx_bytes += length;
>
>         new_skb = alloc_skb(XEM_MAX_FRAME_SIZE + ALIGNMENT, GFP_ATOMIC);
>         if(new_skb == 0) {
>             printk("no memory for new sk_buff\n");
>             spin_unlock_irqrestore(&rcv_lock, flags);
>             return;
>         }
>
>         align = BUFFER_ALIGN(new_skb->data);
>         if(align) skb_reserve(new_skb, align);
>
>         cur_p->app0 = BDSTAT_INT_ON_END_MASK;
>         cur_p->phys_buf_p = (unsigned char *)
>                     pci_map_single(NULL, new_skb->data,
>                                XEM_MAX_FRAME_SIZE,
>                                PCI_DMA_FROMDEVICE);
>         cur_p->buf_len    = XEM_MAX_FRAME_SIZE;
>         rx_skb[cur_rx_bd] = new_skb;
>
>         cur_rx_bd++;
>         if(cur_rx_bd >= RX_BD_NUM) cur_rx_bd = 0;
>
>         cur_p = &cdmac_rx_bd_virt_p->rx_bd[cur_rx_bd];
>         bdstat = cur_p->app0;
>     }
>     sdma_reg_write(lp, RX_TAILDESC_PTR, tail_p);
>
>     spin_unlock_irqrestore(&rcv_lock, flags);
> }
>
> static irqreturn_t
> xps_ll_temac_tx_int(int irq, void * dev_id, struct pt_regs *regs)
> {
>     unsigned int status;
>     struct net_device *dev = (struct net_device *)dev_id;
>     struct net_local *lp = (struct net_local *)dev->priv;
>
>     status = sdma_reg_read(lp, TX_IRQ_REG);
>     sdma_reg_write(lp, TX_IRQ_REG, status);
>
>     if(status & 3) xps_ll_temac_xmit_done(dev);
>     if(status & 0x080) printk("DMA error 0x%x\n", status);
>
>     return IRQ_HANDLED;
> }
>
> static irqreturn_t
> xps_ll_temac_rx_int(int irq, void * dev_id, struct pt_regs *regs)
> {
>     unsigned int status;
>     struct net_device *dev = (struct net_device *)dev_id;
>     struct net_local *lp = (struct net_local *)dev->priv;
>
>     status = sdma_reg_read(lp, RX_IRQ_REG);
>     sdma_reg_write(lp, RX_IRQ_REG, status);
>
> #ifdef XPS_LL_TEMAC_NAPI
>     if(status & 1) {
>         if(likely(netif_rx_schedule_prep(dev))) {
>             sdma_reg_write(lp, RX_CHNL_CTRL,
>                 sdma_reg_read(lp, RX_CHNL_CTRL) & ~SDMA_MASTER_IRQ);
>             __netif_rx_schedule(dev);
>         }
>     }
> #else
>     if(status & 3) xps_ll_temac_recv(dev);
> #endif
>
>     return IRQ_HANDLED;
> }
>
> static void xps_ll_temac_netpoll(struct net_device *dev)
> {
>     //struct net_local *lp = (struct net_local *)dev->priv;
>
>     disable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR);
>     disable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR);
>
>     xps_ll_temac_rx_int(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR, 
> dev, 0);
>     xps_ll_temac_tx_int(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR, 
> dev, 0);
>
>     enable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR);
>     enable_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR);
> }
>
> static struct net_device_stats *xps_ll_temac_get_stats(struct net_device 
> *dev)
> {
>     return netdev_priv(dev);
> }
>
> static int xps_ll_temac_init_descriptor(void)
> {
>     struct sk_buff *skb;
>     unsigned long align;
>     int i;
>
>     cdmac_tx_bd_virt_p = dma_alloc_coherent(NULL, sizeof(struct cdmac_
> tx_bd_t),
>         (dma_addr_t *)&cdmac_tx_bd_phys_p, GFP_KERNEL);
>     cdmac_rx_bd_virt_p = dma_alloc_coherent(NULL, sizeof(struct cdmac_
> rx_bd_t),
>         (dma_addr_t *)&cdmac_rx_bd_phys_p, GFP_KERNEL);
>
>     for(i = 0;i < TX_BD_NUM;i++) {
>         memset((char *)&cdmac_tx_bd_virt_p->tx_bd[i], 0, sizeof(struct 
> cdmac_bd_t));
>         if(i == (TX_BD_NUM - 1)) {
>             cdmac_tx_bd_virt_p->tx_bd[i].next_p =
>                 &cdmac_tx_bd_phys_p->tx_bd[0];
>         } else {
>             cdmac_tx_bd_virt_p->tx_bd[i].next_p =
>                 &cdmac_tx_bd_phys_p->tx_bd[i + 1];
>         }
>     }
>     for(i = 0;i < RX_BD_NUM;i++) {
>         memset((char *)&cdmac_rx_bd_virt_p->rx_bd[i], 0, sizeof(struct 
> cdmac_bd_t));
>         if(i == (RX_BD_NUM - 1)) {
>             cdmac_rx_bd_virt_p->rx_bd[i].next_p =
>                 &cdmac_rx_bd_phys_p->rx_bd[0];
>         } else {
>             cdmac_rx_bd_virt_p->rx_bd[i].next_p =
>                 &cdmac_rx_bd_phys_p->rx_bd[i + 1];
>         }
>         skb = alloc_skb(XEM_MAX_FRAME_SIZE + ALIGNMENT, GFP_ATOMIC);
>         if(skb == 0) {
>             printk("alloc_skb error %d\n", i);
>             return -1;
>         }
>         rx_skb[i] = skb;
>         align = BUFFER_ALIGN(skb->data);
>         if(align) skb_reserve(skb, align);
>
>         cdmac_rx_bd_virt_p->rx_bd[i].phys_buf_p =
>             (unsigned char *)pci_map_single(NULL,
>                 skb->data, XEM_MAX_FRAME_SIZE,
>                 PCI_DMA_FROMDEVICE);
>         cdmac_rx_bd_virt_p->rx_bd[i].buf_len = XEM_MAX_FRAME_SIZE;
>         cdmac_rx_bd_virt_p->rx_bd[i].app0 = BDSTAT_INT_ON_END_MASK;
>     }
>
>     return 0;
> }
>
> static int xps_ll_temac_changemtu(struct net_device *dev, int newmtu)
> {
>         printk("[xilinx_enet]new MTU %d\n", newmtu);
>         dev->mtu = newmtu;
>
>         return 0;
> }
>
> static int xps_ll_temac_open(struct net_device *dev)
> {
>
>     return 0;
> }
>
> static int xps_ll_temac_close(struct net_device *dev)
> {
>
>     return 0;
> }
>
> static struct net_device **xps_ll_temacs;
>
> static int __init xps_ll_temac_init_one(int index)
> {
>     struct net_device *dev = alloc_etherdev(sizeof(struct net_local));
>     struct net_local *lp;
>     int err = 0;
>
>     if (!dev) return -ENOMEM;
>
>     sprintf(dev->name, "eth%d", index);
>
>     /* Initialize the device structure. */
>     dev->get_stats = xps_ll_temac_get_stats;
>     dev->hard_start_xmit = xps_ll_temac_xmit;
>     dev->open = xps_ll_temac_open;
>     dev->stop = xps_ll_temac_close;
> #ifdef XPS_LL_TEMAC_NAPI
>     dev->poll = &xps_ll_temac_poll;
> #endif
>     dev->weight = 64;
>     dev->set_multicast_list = &xps_ll_temac_set_rx_mode;
>     dev->set_mac_address = xps_ll_temac_set_address;
>     dev->poll_controller = xps_ll_temac_netpoll;
>
>     SET_MODULE_OWNER(dev);
> #if 1
> //  dev->features = NETIF_F_SG | NETIF_F_FRAGLIST
> //      | NETIF_F_IP_CSUM;
>     dev->features = NETIF_F_SG | NETIF_F_FRAGLIST;
> #else
>     dev->features = NETIF_F_SG | NETIF_F_FRAGLIST
>         | NETIF_F_TSO | NETIF_F_IP_CSUM;
> #endif
>
>     ether_setup(dev);
>
>     lp = (struct net_local *)dev->priv;
>     memset(lp, 0, sizeof(struct net_local));
>     lp->index = index;
>
>     lp->next_dev = dev_list;
>     lp->sdma_reg_base = ioremap(S_DMA_CTRL_BASEADDR, 128);
>     lp->temac_reg_base = ioremap(XPS_LLTEMAC_BASEADDR, 128);
>
>     dev_list = dev;
>
>     dev->tx_queue_len = 0;
>     dev->change_mtu = xps_ll_temac_changemtu;
>
>     sdma_reg_write(lp, DMA_CONTROL_REG, 1);
>
>     printk(KERN_INFO "%s: Xilinx Embedded Tri-Mode Ethernet MAC\n", dev-
>   
>> name);
>>     
>     xps_ll_temac_init_descriptor();
>
>     dev->dev_addr[0] = 0x00;
>     dev->dev_addr[1] = 0x80;
>     dev->dev_addr[2] = 0x49;
>     dev->dev_addr[3] = 0x00;
>     dev->dev_addr[4] = 0x00;
>     dev->dev_addr[5] = 0x00;
>
>     request_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_TX_INTOUT_INTR,
>         &xps_ll_temac_tx_int, 0, dev->name, dev);
>     request_irq(XPAR_XPS_INTC_0_DDR_SDRAM_SDMA2_RX_INTOUT_INTR,
>         &xps_ll_temac_rx_int, 0, dev->name, dev);
>
>     xps_ll_temac_indirect_set(dev, 0, MC, MDIO_ENABLE_MASK | MDIO_CLOCK_
> DIV_100MHz);
>
>     xps_ll_temac_indirect_set(dev, 0, RCW1, 0x10000000);    // Enable 
> Receiver
>     xps_ll_temac_indirect_set(dev, 0, TC, 0x10000000);  // Enable 
> Transmitter
>     xps_ll_temac_indirect_set(dev, 0, EMMC, 0x84000000);
>
>     xps_ll_temac_indirect_set(dev, 0, UAW0, (dev->dev_addr[0] & 
> 0x000000ff) |
>         ((dev->dev_addr[1] << 8) & 0x0000ff00) |
>         ((dev->dev_addr[2] << 16)& 0x00ff0000) |
>         ((dev->dev_addr[3] << 24) & 0xff000000));
>     xps_ll_temac_indirect_set(dev, 0, UAW1, (dev->dev_addr[4] & 
> 0x000000ff) |
>         ((dev->dev_addr[5] << 8) & 0x0000ff00));
>
>     xps_ll_temac_indirect_set(dev, 0, AFM, 0x00000000);
>
>     xps_ll_temac_phy_ctrl(dev);
>
>     sdma_reg_write(lp, TX_CHNL_CTRL, 0x10220483);
>     //sdma_reg_write(lp, TX_CHNL_CTRL, 0x00100483);
>     sdma_reg_write(lp, RX_CHNL_CTRL, 0xff010283);
>
>     sdma_reg_write(lp, RX_CURDESC_PTR, (unsigned int)&cdmac_rx_bd_phys_
> p->rx_bd[0]);
>     sdma_reg_write(lp, RX_TAILDESC_PTR, (unsigned int)&cdmac_rx_bd_phys_
> p->rx_bd[RX_BD_NUM - 1]);
>
>     if ((err = register_netdev(dev))) {
>         free_netdev(dev);
>         dev = NULL;
>     } else {
>         xps_ll_temacs[index] = dev;
>     }
>
>     return 0;
> }
>
> static void xps_ll_temac_free_one(int index)
> {
>     unregister_netdev(xps_ll_temacs[index]);
>     free_netdev(xps_ll_temacs[index]);
> }
>
> static int __init xps_ll_temac_init_module(void)
> {
>     int err = 0;
>
>     xps_ll_temacs = kmalloc(sizeof(void *), GFP_KERNEL);
>
>     if (!xps_ll_temacs) return -ENOMEM;
>
>     spin_lock_init(&dev_lock);
>     spin_lock_init(&rcv_lock);
>     spin_lock_init(&xmt_lock);
>
>     if((err = xps_ll_temac_init_one(0))) xps_ll_temac_free_one(0);
>
>     return err;
> }
>
> static void __exit xps_ll_temac_cleanup_module(void)
> {
>     xps_ll_temac_free_one(0);
>     kfree(xps_ll_temacs);
> }
>
> module_init(xps_ll_temac_init_module);
> module_exit(xps_ll_temac_cleanup_module);
> MODULE_LICENSE("GPL");
>
>
>   
>> Thanks,
>>
>>     I have alot of work to do on our stuff, I might as well see if I 
>>     
> can
>   
>> move to the powerpc tree at the same time.
>>
>>     BTW is there even the beginings of a non-xilinx lltemac driver out
>> there ? There were hints on the list, but I have not seen anything.
>>
>>     I would be happy to help advance the ball on anything anyone has
>> started.
>>
>>    
>>
>> Grant Likely wrote:
>>     
>>> On Sun, Apr 20, 2008 at 2:31 PM, David H. Lynch Jr. <dhlii@dlasys.net
>>>       
>> wrote:
>>     
>>>   
>>>       
>>>> Thanks.
>>>>
>>>>     I am running linus's 2.6.25-rc9, but I can pull the Xilinx tree 
>>>>         
> or
>   
>>>>  yours - when your server is up.
>>>>
>>>>     I can not find any Xilinx powerpc configs in arch/powerpc/
>>>>         
> config
>   
>>>>     Do I just need to do a
>>>>     make ARCH=powerpc menuconfig and create one from scratch ?
>>>>     
>>>>         
>>> That's right; I haven't merged any defconfigs.  Roll your own.
>>>
>>>   
>>>       
>>>>     Is simpleboot in your tree or the xilinx tree, if  can not find 
>>>>         
> it
>   
>>>>  in Linus's ?
>>>>     
>>>>         
>>> If you want to use the "simpleboot" wrapper; then you'll need to 
>>>       
> pull
>   
>>> paulus' tree (Linus hasn't yet pulled his tree; but he probably will
>>> any moment now).
>>>
>>> Cheers,
>>> g.
>>>
>>>   
>>>       
>> -- 
>> Dave Lynch                            DLA Systems
>> Software Development:                          Embedded Linux
>> 717.627.3770         dhlii@dlasys.net       http://www.dlasys.net
>> fax: 1.253.369.9244                      Cell: 1.717.587.7774
>> Over 25 years' experience in platforms, languages, and technologies 
>>     
> too numerous to list.
>   
>> "Any intelligent fool can make things bigger and more complex... It 
>>     
> takes a touch of genius - and a lot of courage to move in the opposite 
> direction."
>   
>> Albert Einstein
>>
>> _______________________________________________
>> Linuxppc-embedded mailing list
>> Linuxppc-embedded@ozlabs.org
>> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>>
>>     
> ※ ４月から所属・部署が変更になりました
> ------------------------------------------------------------------
> 柏木良夫
> 株式会社日新システムズ 東日本営業部
>
> 本      社 〒600-8482 京都市下京区堀川通四条下ル東側 堀川四条ビル
> TEL 075-344-7977 FAX 075-344-7887
> 東京事務所 〒101-0024 東京都千代田区神田和泉町1番地 神田和泉町ビル
> TEL 03-5825-2081 FAX 03-5821-1259
> E-Mail kashiwagi@co-nss.co.jp HTTP http://www.co-nss.co.jp/
> ------------------------------------------------------------------
>
>   


-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein

Re: Xilinx PowerPC

[Yoshio Kashiwagi]

David-san,

This driver must add some corrections, for example, Checksum-offloading
is incomplete as you say it. As for me, it is very happy to obtain
your cooperation. Let me know if I can be of any work for it.

Moreover, although I wrote the xps_ll_temac driver for u-boot of a very 
simple single SDMA descriptor, I have not got used and put in the
contribution way.

Best Regards,
Yoshio Kashiwagi - Nissin Systems

>  I was able to incorporate Yoshio's driver and make use of it.
> It is a single file, fairly short, easy to understand, mostly conforming
> to linux kernel programming style. While it needs some work fo NAPI and
> OF, it is usable as it is.
> 
> I would be happy to see it form the starting point for a distribution LL
> TEMAC driver.
> 
> If Yoshio does not wish to put together a patch, for it I can, however,
> I have made some name changes and restructing to compare to my older PLB
> TEMAC.

Re: Xilinx PowerPC

[David H. Lynch Jr.]

I am preparing my own version of Yoshio's driver. Since most of my
changes are re-ordering, renaming, or other items of small consequence,
I will leave Yoshio as the author - unless he prefers otherwise.

I hope to have it ready to submit as a patch in a few days.
If Yoshio is unhappy with any of the changes I have made or chooses to
submit his original code or some permutation I will withdrawl mine.

While adding OF checksum offloading, ... would be nice,
All I am looking to do is get something that can start its way into the
kernel.
This driver is working for me - with very light testing so far.

If no one else is willing to Shepard something in, then I will try.
Hints, clues, etc would be very much appreciated.
I presume that if this driver is acceptable here, the next step would be
forwarding it to linux-netdev ?

Yoshio Kashiwagi wrote:
> David-san,
>
> This driver must add some corrections, for example, Checksum-offloading
> is incomplete as you say it. As for me, it is very happy to obtain
> your cooperation. Let me know if I can be of any work for it.
>
> Moreover, although I wrote the xps_ll_temac driver for u-boot of a very 
> simple single SDMA descriptor, I have not got used and put in the
> contribution way.
>
> Best Regards,
> Yoshio Kashiwagi - Nissin Systems

-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein

Re: Xilinx PowerPC

[Grant Likely]

On Sun, May 18, 2008 at 9:27 PM, David H. Lynch Jr. <dhlii@dlasys.net> wrote:
> I am preparing my own version of Yoshio's driver. Since most of my
> changes are re-ordering, renaming, or other items of small consequence,
> I will leave Yoshio as the author - unless he prefers otherwise.
>
> I hope to have it ready to submit as a patch in a few days.
> If Yoshio is unhappy with any of the changes I have made or chooses to
> submit his original code or some permutation I will withdrawl mine.
>
> While adding OF checksum offloading, ... would be nice,
> All I am looking to do is get something that can start its way into the
> kernel.
> This driver is working for me - with very light testing so far.
>
> If no one else is willing to Shepard something in, then I will try.
> Hints, clues, etc would be very much appreciated.
> I presume that if this driver is acceptable here, the next step would be
> forwarding it to linux-netdev ?

When you have a patch ready, post it both to here (linuxppc-dev) and
to netdev.  I'll provide review and help.  Most likely when it is
ready it will go in through jgarzik's tree, but if Jeff prefers, it
can go in via my Virtex tree.  Make sure you're able to generate and
send clean patches before you post it to the mailing lists (inline
instead of attachment, no line wrapping, applies cleanly to head of
tree).  If you need help, email me directly, and you can even send
test emails to me for me to check before posting to the list.

It doesn't have to be perfect before you submit, but you do need to be
prepared to respond to comments quickly and expect to respin the patch
several times.

Cheers,
g.

>
> Yoshio Kashiwagi wrote:
>> David-san,
>>
>> This driver must add some corrections, for example, Checksum-offloading
>> is incomplete as you say it. As for me, it is very happy to obtain
>> your cooperation. Let me know if I can be of any work for it.
>>
>> Moreover, although I wrote the xps_ll_temac driver for u-boot of a very
>> simple single SDMA descriptor, I have not got used and put in the
>> contribution way.
>>
>> Best Regards,
>> Yoshio Kashiwagi - Nissin Systems
>
> --
> Dave Lynch                                                  DLA Systems
> Software Development:                                    Embedded Linux
> 717.627.3770           dhlii@dlasys.net           http://www.dlasys.net
> fax: 1.253.369.9244                                Cell: 1.717.587.7774
> Over 25 years' experience in platforms, languages, and technologies too numerous to list.
>
> "Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
> Albert Einstein
>
>



-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.

Re: Xilinx PowerPC

[Yoshio Kashiwagi]

David-san,

I posted this source code so that usage of SDMA could be shared.
Probably this code will be able to be used also by Virtex-5 FXT.
Since this source code is a reference level and start point as you
say, it is happy that it is maintainable with people of this ML.
I'll leave it in your hands about author.

Best Regards,
Yoshio Kashiwagi - Nissin Systems

> I am preparing my own version of Yoshio's driver. Since most of my
> changes are re-ordering, renaming, or other items of small 
consequence,
> I will leave Yoshio as the author - unless he prefers otherwise.
> 
> I hope to have it ready to submit as a patch in a few days.
> If Yoshio is unhappy with any of the changes I have made or chooses to
> submit his original code or some permutation I will withdrawl mine.
> 
> While adding OF checksum offloading, ... would be nice,
> All I am looking to do is get something that can start its way into 
the
> kernel.
> This driver is working for me - with very light testing so far.
> 
> If no one else is willing to Shepard something in, then I will try.
> Hints, clues, etc would be very much appreciated.
> I presume that if this driver is acceptable here, the next step would 
be
> forwarding it to linux-netdev ?
> 
> Yoshio Kashiwagi wrote:
> > David-san,
> >
> > This driver must add some corrections, for example, Checksum-
offloading
> > is incomplete as you say it. As for me, it is very happy to obtain
> > your cooperation. Let me know if I can be of any work for it.
> >
> > Moreover, although I wrote the xps_ll_temac driver for u-boot of a 
very 
> > simple single SDMA descriptor, I have not got used and put in the
> > contribution way.
> >
> > Best Regards,
> > Yoshio Kashiwagi - Nissin Systems
> 
> -- 
> Dave Lynch                            DLA Systems
> Software Development:                          Embedded Linux
> 717.627.3770         dhlii@dlasys.net       http://www.dlasys.net
> fax: 1.253.369.9244                      Cell: 1.717.587.7774
> Over 25 years' experience in platforms, languages, and technologies 
too numerous to list.
> 
> "Any intelligent fool can make things bigger and more complex... It 
takes a touch of genius - and a lot of courage to move in the opposite 
direction."
> Albert Einstein

Re: Xilinx PowerPC

[Grant Likely]

On Sun, May 18, 2008 at 10:18 PM, Yoshio Kashiwagi
<kashiwagi@co-nss.co.jp> wrote:
> David-san,
>
> I posted this source code so that usage of SDMA could be shared.
> Probably this code will be able to be used also by Virtex-5 FXT.
> Since this source code is a reference level and start point as you
> say, it is happy that it is maintainable with people of this ML.
> I'll leave it in your hands about author.

It is probably appropriate for David to maintain your copyright notice
at the top of the file.  Also, you should provide your "signed-off-by"
line to confirm that there are no legal issues with you contributing
your code (see Documentation/SubmittingPatches for details).  Email it
to Dave, and he should preserve your signed-off-by line when he posts
his modified patch.

Cheers,
g.

-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.

Xilinx PowerPC

[David H. Lynch Jr.]

    So when you have Xilinx under powerpc working, do we pull it from your
    git tree or the xilinx one ?

    Will there be an announcement ?

    How about a one paragraph getting started guide to moving a xilinx
ppc bsp to
    xilinx powerpc.

    Like  ?

    step 1).
             Generate a dts - gen-mhs-devicetree ?
             And pass it to through your boot loader to Linux ?
   
    Step 2).
             the code in arch/powerpc/???? is the devicetree equvalent to
             arch/ppc/platforms/4xx/xilinx_ml410.c

    Step 3).
             device drivers need to change initialization/setup from
             ???? to ???
      
    Just something to give those of us with no clue, a small one to start.

     Thanks.

RE: Xilinx PowerPC

[Stephen Neuendorffer]

> -----Original Message-----
> From: linuxppc-embedded-bounces+stephen=3Dneuendorffer.name@ozlabs.org
[mailto:linuxppc-embedded-
> bounces+stephen=3Dneuendorffer.name@ozlabs.org] On Behalf Of David H.
Lynch Jr.
> Sent: Wednesday, February 20, 2008 10:59 PM
> To: Grant Likely; linuxppc-embedded
> Subject: Xilinx PowerPC
>=20
>=20
>     So when you have Xilinx under powerpc working, do we pull it from
your
>     git tree or the xilinx one ?
>=20
>     Will there be an announcement ?
>=20
>     How about a one paragraph getting started guide to moving a xilinx
> ppc bsp to
>     xilinx powerpc.
>=20
>     Like  ?
>=20
>     step 1).
>              Generate a dts - gen-mhs-devicetree ?
>              And pass it to through your boot loader to Linux ?

http://git.xilinx.com/gen-mhs-devtree.git contains two utilities for
generating device trees from EDK projects.  The older option is a python
script, originally written by Grant.  The newer (and probably more
mature at this point) option is an EDK BSP generator for u-boot,
originally written by Michal Simek.  Preferably this gets passed from a
uboot, although it's also possible to compile it into the kernel.
Currently, the uboot option is somewhat annoying because there's not a
good surefire way for getting uboot up and running on an EDK design.
(Unfortunately, the EDK BSP generator is incomplete, and the uboot
support for virtex needs some work to get it integrated as well).

>     Step 2).
>              the code in arch/powerpc/???? is the devicetree equvalent
to
>              arch/ppc/platforms/4xx/xilinx_ml410.c
http://git.xilinx.com/linux-2.6-xlnx.git contains a preliminary stab at
the bootcode for Virtex.  I've been using this for a while with
ARCH=3Dpowerpc.  There isn't really much need for board specific =
platform
code: This should (I think) be handled by the device tree.  All of this
is still pretty preliminary at the moment however.  The big problem with
the boot code in the Xilinx tree is that you need to be able to allocate
memory in order to parse the device tree in order to figure out how much
memory you have.  I just haven't rewritten the code to use libfdt, which
can query the device tree without memory allocation yet.  Grant has a
slightly different scheme in his tree, but it works as well.  Generally
I've been more focused on avoiding u-boot, whereas Grant has been using
u-boot exclusively, I think.

>     Step 3).
>              device drivers need to change initialization/setup from
>              ???? to ???
All the xilinx drivers in mainline and most of the ones in the
git.xilinx.com tree export both platform device and of_device
interfaces.  You should be able to look at any of them to figure out a
good way of adding of_device support to a driver.  the ll_temac is by
far the most complex instance of this, which has to do some non-trivial
traversal of the device tree to discover the information it needs about
it's interface to memory.

So, the short answer is, you might be able to get it to work today, but
the code isn't in shape for mainline yet.

Steve

Re: Xilinx PowerPC

[Grant Likely]

On Thu, Feb 21, 2008 at 10:50 AM, Stephen Neuendorffer
<stephen.neuendorffer@xilinx.com> wrote:
>  >     Step 2).
>  >              the code in arch/powerpc/???? is the devicetree equvalent
>  to
>  >              arch/ppc/platforms/4xx/xilinx_ml410.c
>  http://git.xilinx.com/linux-2.6-xlnx.git contains a preliminary stab at
>  the bootcode for Virtex.  I've been using this for a while with
>  ARCH=powerpc.  There isn't really much need for board specific platform
>  code: This should (I think) be handled by the device tree.  All of this
>  is still pretty preliminary at the moment however.  The big problem with
>  the boot code in the Xilinx tree is that you need to be able to allocate
>  memory in order to parse the device tree in order to figure out how much
>  memory you have.  I just haven't rewritten the code to use libfdt, which
>  can query the device tree without memory allocation yet.  Grant has a
>  slightly different scheme in his tree, but it works as well.  Generally
>  I've been more focused on avoiding u-boot, whereas Grant has been using
>  u-boot exclusively, I think.

Actually, I'm not using u-boot at all.  I've reworked the 'raw' image
target to use libfdt and properly initialize the cache and stuff.
I'll be merging it in the .26 merge window.

Cheers,
g.

-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.

Re: Xilinx PowerPC

[Alan Casey]

>-- Original Message --
>Date: Thu, 21 Feb 2008 12:02:21 -0700
>From: "Grant Likely" <grant.likely@secretlab.ca>
>To: "Stephen Neuendorffer" <stephen.neuendorffer@xilinx.com>
>Subject: Re: Xilinx PowerPC
>Cc: linuxppc-embedded <linuxppc-embedded@ozlabs.org>
>
>
>On Thu, Feb 21, 2008 at 10:50 AM, Stephen Neuendorffer
><stephen.neuendorffer@xilinx.com> wrote:
>>  >     Step 2).
>>  >              the code in arch/powerpc/???? is the devicetree equval=
ent
>>  to
>>  >              arch/ppc/platforms/4xx/xilinx_ml410.c
>>  http://git.xilinx.com/linux-2.6-xlnx.git contains a preliminary stab
at
>>  the bootcode for Virtex.  I've been using this for a while with
>>  ARCH=3Dpowerpc.  There isn't really much need for board specific plat=
form
>>  code: This should (I think) be handled by the device tree.  All of th=
is
>>  is still pretty preliminary at the moment however.  The big problem w=
ith
>>  the boot code in the Xilinx tree is that you need to be able to alloc=
ate
>>  memory in order to parse the device tree in order to figure out how m=
uch
>>  memory you have.  I just haven't rewritten the code to use libfdt, wh=
ich
>>  can query the device tree without memory allocation yet.  Grant has a=

>>  slightly different scheme in his tree, but it works as well.  General=
ly
>>  I've been more focused on avoiding u-boot, whereas Grant has been usi=
ng
>>  u-boot exclusively, I think.
>
>Actually, I'm not using u-boot at all.  I've reworked the 'raw' image
>target to use libfdt and properly initialize the cache and stuff.
>I'll be merging it in the .26 merge window.
>

Hi,

  Just wondering - is it possible to use U-Boot on the Xilinx XUPV2PRO
  board or any similar tools (such as libfdt?) to copy a Linux kernel and=


  ramdisk filesystem from a SystemACE card to memory, uncompress them and=


  boot the Linux kernel etc?? If so is there any notes/docs about how to
do
  this or similar online anywhere?

  Any info. appreciated,
  Regards&Thanks,
  Alan.
  

RE: Xilinx PowerPC

[Stephen Neuendorffer]

>=20
> Hi,
>=20
>   Just wondering - is it possible to use U-Boot on the Xilinx XUPV2PRO
>   board or any similar tools (such as libfdt?) to copy a Linux kernel
and
>=20
>   ramdisk filesystem from a SystemACE card to memory, uncompress them
and
>=20
>   boot the Linux kernel etc?? If so is there any notes/docs about how
to
> do
>   this or similar online anywhere?
>=20
>   Any info. appreciated,
>   Regards&Thanks,
>   Alan.

In theory, yes...  All the infrastructure to do this exists, but I don't
know of anyone who as actually tried it yet with ARCH=3Dpowerpc kernels.
There=20

Steve

RE: Xilinx PowerPC

[Alan Casey]

>-- Original Message --
>Subject: RE: Xilinx PowerPC
>Date: Fri, 22 Feb 2008 10:11:19 -0800
>From: "Stephen Neuendorffer" <stephen.neuendorffer@xilinx.com>
>To: <alan.casey5@mail.dcu.ie>,
>	"Grant Likely" <grant.likely@secretlab.ca>
>Cc: linuxppc-embedded <linuxppc-embedded@ozlabs.org>
>
>
>> 
>> Hi,
>> 
>>   Just wondering - is it possible to use U-Boot on the Xilinx XUPV2PRO=

>>   board or any similar tools (such as libfdt?) to copy a Linux kernel
>and
>> 
>>   ramdisk filesystem from a SystemACE card to memory, uncompress them
>and
>> 
>>   boot the Linux kernel etc?? If so is there any notes/docs about how
>to
>> do
>>   this or similar online anywhere?
>> 
>>   Any info. appreciated,
>>   Regards&Thanks,
>>   Alan.
>
>In theory, yes...  All the infrastructure to do this exists, but I don't=

>know of anyone who as actually tried it yet with ARCH=3Dpowerpc kernels.=

>There 
>
>Steve

 So i can try the latest official U-Boot releases with the ML300 configur=
ation
or is there a port or similar of U-Boot for the Xilinx XUPV2PRO board??

  Regards&Thanks,
  Alan.

RE: Xilinx PowerPC

[Stephen Neuendorffer]

> -----Original Message-----
> From: Alan Casey [mailto:alan.casey5@mail.dcu.ie]
> Sent: Friday, February 22, 2008 10:58 AM
> To: Stephen Neuendorffer; Grant Likely
> Cc: linuxppc-embedded
> Subject: RE: Xilinx PowerPC
>=20
>=20
> >-- Original Message --
> >Subject: RE: Xilinx PowerPC
> >Date: Fri, 22 Feb 2008 10:11:19 -0800
> >From: "Stephen Neuendorffer" <stephen.neuendorffer@xilinx.com>
> >To: <alan.casey5@mail.dcu.ie>,
> >	"Grant Likely" <grant.likely@secretlab.ca>
> >Cc: linuxppc-embedded <linuxppc-embedded@ozlabs.org>
> >
> >
> >>
> >> Hi,
> >>
> >>   Just wondering - is it possible to use U-Boot on the Xilinx
XUPV2PRO
> >>   board or any similar tools (such as libfdt?) to copy a Linux
kernel
> >and
> >>
> >>   ramdisk filesystem from a SystemACE card to memory, uncompress
them
> >and
> >>
> >>   boot the Linux kernel etc?? If so is there any notes/docs about
how
> >to
> >> do
> >>   this or similar online anywhere?
> >>
> >>   Any info. appreciated,
> >>   Regards&Thanks,
> >>   Alan.
> >
> >In theory, yes...  All the infrastructure to do this exists, but I
don't
> >know of anyone who as actually tried it yet with ARCH=3Dpowerpc
kernels.
> >There
> >
> >Steve
>=20
>  So i can try the latest official U-Boot releases with the ML300
configuration
> or is there a port or similar of U-Boot for the Xilinx XUPV2PRO
board??
>=20
>   Regards&Thanks,
>   Alan.
>=20
>=20

That's probably your best starting point....  There is a uboot
application note on the Xilinx website (XAPP 542), but it's a little
thin on porting information.

Steve

RE: Xilinx PowerPC

[Stephen Neuendorffer]

> -----Original Message-----
> From: David H. Lynch Jr. [mailto:dhlii@dlasys.lcl]
> Sent: Friday, February 22, 2008 1:51 PM
> To: Stephen Neuendorffer
> Cc: linuxppc-embedded
> Subject: Re: Xilinx PowerPC
>=20
> Stephen Neuendorffer wrote:
> >
> >> -----Original Message-----
> >> From:
linuxppc-embedded-bounces+stephen=3Dneuendorffer.name@ozlabs.org
> >>
> > [mailto:linuxppc-embedded-
> >
> >> bounces+stephen=3Dneuendorffer.name@ozlabs.org] On Behalf Of David =
H.
> >>
> > Lynch Jr.
> >
> >> Sent: Wednesday, February 20, 2008 10:59 PM
> >> To: Grant Likely; linuxppc-embedded
> >> Subject: Xilinx PowerPC
> >>
> >>
> >>     So when you have Xilinx under powerpc working, do we pull it
from
> >>
> > your
> >
> >>     git tree or the xilinx one ?
> >>
> >>     Will there be an announcement ?
> >>
> >>     How about a one paragraph getting started guide to moving a
xilinx
> >> ppc bsp to
> >>     xilinx powerpc.
> >>
> >>     Like  ?
> >>
> >>     step 1).
> >>              Generate a dts - gen-mhs-devicetree ?
> >>              And pass it to through your boot loader to Linux ?
> >>
> >
> > http://git.xilinx.com/gen-mhs-devtree.git contains two utilities for
> > generating device trees from EDK projects.  The older option is a
python
> > script, originally written by Grant.  The newer (and probably more
> > mature at this point) option is an EDK BSP generator for u-boot,
> > originally written by Michal Simek.  Preferably this gets passed
from a
> > uboot, although it's also possible to compile it into the kernel.
> > Currently, the uboot option is somewhat annoying because there's not
a
> > good surefire way for getting uboot up and running on an EDK design.
> > (Unfortunately, the EDK BSP generator is incomplete, and the uboot
> > support for virtex needs some work to get it integrated as well).
> >
>     Thanks, I found the python script. Please do not fixate on u-boot.
>      it is not an option for us. We have our own bootloader.
>     it handles .elf files, fits in 16k or BRAM and just starts when
the
>      FPGA is loaded. We will have to added  device tree support to it.
>     I had thought you were talking about appending the device tree to
> the bitfile.
>      That sounds like a really good option for us.
>=20
>     We also really don't have a great interest in the EDK BSP
generator.
>     We have our own BSP's for our cards for both GHS and Linux.
>     Except the rare instances - like right now while we are
> transitioning to the
>     newer LL_TEMAC,  we use normal linux drivers, either from the
> existing kernel
>     trees, or written inhouse.

The BSP generator is significantly different from the normal EDK Linux
BSP.  The key thing is that the device tree generated by the python
script is necessarily incomplete, since it does not have information
about default parameter values which appear in the core definitions.  I
would *highly recommend* using the uboot bsp to get a device tree.  You
can always ignore anything else that it does (or hack the code to your
hearts content to do what you want, such as incorporating it into your
own BSP).

> >>     Step 2).
> >>              the code in arch/powerpc/???? is the devicetree
equvalent
> >>
> > to
> >
> >>              arch/ppc/platforms/4xx/xilinx_ml410.c
> >>
> > http://git.xilinx.com/linux-2.6-xlnx.git contains a preliminary stab
at
> > the bootcode for Virtex.  I've been using this for a while with
> > ARCH=3Dpowerpc.  There isn't really much need for board specific
platform
> > code: This should (I think) be handled by the device tree.
>     So I have a device tree for my card somehow. Presumably I need the
> binary form.
>     My bootloader needs to pass a pointer to a copy in memory to
Linux.
>=20
>     How do I config the linux from your git tree ?
>     Do I just config for a generic Xilinx card and the device tree
> handles the rest ?

Essentially.  There is a generic virtex configuration option.

>     Next we have a hosted pseudo serial port. Devict Trees could cope
> with that.
>     But We also have early serial support for both it and Uartlite,
>     Also our Linux BSP and UartLite drivers will work without a PIC.
>     And would even do ethernet (badly) with the earlier LL_TEMAC
without
>     DMA and interrupts.
>     So do I modify the device tree to choose a different uartlite
driver ?
>     I had to put a fair amount of patches in to support early serial
with
>     non-8250 serial ports. Do I have to patch your generic xilinx BSP
to
>     deal with that ?

Early serial on uartlite is works fine.  You have to add an explicit
reference to the uartlite to get it to work.  Early serial on uart16550
doesn't work yet, because the device has an unusual spacing between the
registers.

>     Please do not develop a u-boot only solution.

I don't believe Linux will never *require* uboot.

>     Just to clarify, am I correct in presuming that say the uartlite
> driver in the
>     xilinx git tree will handle either the arch/ppc style platform
> device initialization
>     or the arch/powerpc devicetree initialization ?

Yes.  drivers are capable of probing both the platform bus and the of
bus.  It is expected that each device is advertised on only one.  In
general, ppc will advertise platform devices and powerpc will advertise
of devices.

>     Is there some means of doing a phased transition ?
>     Can I lift the device tree BSP setup code into my own BSP,
>     and then remove the platform device code peice by peice as things
work ?
probably, although changing the drivers is not challenging, as long as
all the information you need is in the device tree.

>     Is there something magical about being physically located in
> arch/ppc or arch/powerpc ?

Well, ppc doesn't support of/device trees.  All of the of bus stuff is
#ifdef'd out of the drivers.

Steve

Re: Xilinx PowerPC

[Michal Simek]

Hi All,

> http://git.xilinx.com/gen-mhs-devtree.git contains two utilities for
> generating device trees from EDK projects.  The older option is a python
> script, originally written by Grant.  The newer (and probably more
> mature at this point) option is an EDK BSP generator for u-boot,
> originally written by Michal Simek.  Preferably this gets passed from a
> uboot, although it's also possible to compile it into the kernel.
> Currently, the uboot option is somewhat annoying because there's not a
> good surefire way for getting uboot up and running on an EDK design.
> (Unfortunately, the EDK BSP generator is incomplete, and the uboot
> support for virtex needs some work to get it integrated as well).

U-BOOT part will be removed from EDK generator. The new name will be only
fdt_v1.00.a. I have prepared separated version for FDT generation and for U-BOOT
generation.

Steve: You can remove U-BOOT part from generator. This part is useful only for
Microblaze cpu.

Regards,
Michal Simek

RE: Xilinx PowerPC

[Stephen Neuendorffer]

Michel,

I've had it on my todo list to get the u-boot stuff working for powerpc
too...  However, I remember a brief comment from John Williams that
you'd separated the BSP generation process more.  Details (or better
yet, code and patches) would be appreciated, here, since I'm a bit
swamped at the moment.

Steve

> -----Original Message-----
> From: Michal Simek [mailto:monstr@seznam.cz]
> Sent: Sunday, March 02, 2008 10:25 AM
> To: Stephen Neuendorffer
> Cc: David H. Lynch Jr.; Grant Likely; linuxppc-embedded
> Subject: Re: Xilinx PowerPC
>=20
> Hi All,
>=20
> > http://git.xilinx.com/gen-mhs-devtree.git contains two utilities for
> > generating device trees from EDK projects.  The older option is a
python
> > script, originally written by Grant.  The newer (and probably more
> > mature at this point) option is an EDK BSP generator for u-boot,
> > originally written by Michal Simek.  Preferably this gets passed
from a
> > uboot, although it's also possible to compile it into the kernel.
> > Currently, the uboot option is somewhat annoying because there's not
a
> > good surefire way for getting uboot up and running on an EDK design.
> > (Unfortunately, the EDK BSP generator is incomplete, and the uboot
> > support for virtex needs some work to get it integrated as well).
>=20
> U-BOOT part will be removed from EDK generator. The new name will be
only
> fdt_v1.00.a. I have prepared separated version for FDT generation and
for U-BOOT
> generation.
>=20
> Steve: You can remove U-BOOT part from generator. This part is useful
only for
> Microblaze cpu.
>=20
> Regards,
> Michal Simek