From: Franck JULLIEN <franck.jullien@gmail.com>
To: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Cc: barebox <barebox@lists.infradead.org>
Subject: Re: [PATCH] [v2] Nios2: Add Altera TSE MAC driver
Date: Sun, 10 Apr 2011 09:13:04 +0200 [thread overview]
Message-ID: <BANLkTi=JnLBhDeYY_eXRtxb9b_3Z=2rS3A@mail.gmail.com> (raw)
In-Reply-To: <20110410040649.GA8392@game.jcrosoft.org>
[-- Attachment #1.1: Type: text/plain, Size: 9262 bytes --]
Hello,
2011/4/10 Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
> On 16:08 Sat 09 Apr , franck.jullien@gmail.com wrote:
> > From: Franck JULLIEN <franck.jullien@gmail.com>
> >
> > Add Altera Triple Speed Ethernet driver
> >
> > Signed-off-by: Franck JULLIEN <franck.jullien@gmail.com>
> > ---
> > drivers/net/Kconfig | 16 ++
> > drivers/net/Makefile | 5 +-
> > drivers/net/altera_tse.c | 620
> ++++++++++++++++++++++++++++++++++++++++++++++
> > drivers/net/altera_tse.h | 313 +++++++++++++++++++++++
> > 4 files changed, 952 insertions(+), 2 deletions(-)
> > create mode 100644 drivers/net/altera_tse.c
> > create mode 100644 drivers/net/altera_tse.h
> >
> > diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
> > index 6479417..19e35db 100644
> > --- a/drivers/net/Kconfig
> > +++ b/drivers/net/Kconfig
> > @@ -84,6 +84,22 @@ config DRIVER_NET_TAP
> > bool "tap Ethernet driver"
> > depends on LINUX
> >
> > +config DRIVER_NET_TSE
> > + depends on NIOS2
> > + bool "Altera TSE ethernet driver"
> > + select MIIDEV
> > + help
> > + This option enables support for the Altera TSE MAC.
> > +
> > +config TSE_USE_DEDICATED_DESC_MEM
> > + depends on DRIVER_NET_TSE
> > + bool "Altera TSE uses dedicated descriptor memory"
> > + help
> > + This option tells the TSE driver to use an onchip memory
> > + to store SGDMA descriptors. Descriptor memory is not
> > + reserved with a malloc but directly mapped to the memory
> > + address (defined in config.h)
> > +
> > source "drivers/net/usb/Kconfig"
> >
> > endmenu
> > diff --git a/drivers/net/Makefile b/drivers/net/Makefile
> > index 96d3d32..f02618b 100644
> > --- a/drivers/net/Makefile
> > +++ b/drivers/net/Makefile
> > @@ -1,7 +1,7 @@
> > -obj-$(CONFIG_DRIVER_NET_CS8900) += cs8900.o
> > +obj-$(CONFIG_DRIVER_NET_CS8900) += cs8900.o
> > obj-$(CONFIG_DRIVER_NET_SMC911X) += smc911x.o
> > obj-$(CONFIG_DRIVER_NET_SMC91111) += smc91111.o
> > -obj-$(CONFIG_DRIVER_NET_DM9000) += dm9000.o
> > +obj-$(CONFIG_DRIVER_NET_DM9000) += dm9000.o
> why do you touch this two?
>
> if you want to fix the whitespace please do it in an other commit
>
OK, I'll correct this.
> > obj-$(CONFIG_DRIVER_NET_NETX) += netx_eth.o
> > obj-$(CONFIG_DRIVER_NET_AT91_ETHER) += at91_ether.o
> > obj-$(CONFIG_DRIVER_NET_MPC5200) += fec_mpc5200.o
> > @@ -11,3 +11,4 @@ obj-$(CONFIG_DRIVER_NET_MACB) += macb.o
> > obj-$(CONFIG_DRIVER_NET_TAP) += tap.o
> > obj-$(CONFIG_MIIDEV) += miidev.o
> > obj-$(CONFIG_NET_USB) += usb/
> > +obj-$(CONFIG_DRIVER_NET_TSE) += altera_tse.o
> > diff --git a/drivers/net/altera_tse.c b/drivers/net/altera_tse.c
> > new file mode 100644
> > index 0000000..2687377
> > --- /dev/null
> > +
> > +static int tse_get_ethaddr(struct eth_device *edev, unsigned char *m)
> > +{
> > + /* There is no eeprom */
> so return the content of the register no?
>
Well, the register is reseted to 0 when the MAC starts so there is no
Ethernet address
to get.
> > + return -1;
> > +}
> > +
>
> > +static int tse_eth_send(struct eth_device *edev, void *packet, int
> length)
> > +{
> > +
> > + struct altera_tse_priv *priv = edev->priv;
> > + struct alt_sgdma_registers *tx_sgdma = priv->sgdma_tx;
> > + struct alt_sgdma_descriptor *tx_desc = (struct alt_sgdma_descriptor
> *)priv->tx_desc;
> > +
> > + struct alt_sgdma_descriptor *tx_desc_cur = (struct
> alt_sgdma_descriptor *)&tx_desc[0];
> > +
> > + flush_dcache_range((uint32_t)packet, (uint32_t)packet + length);
> > + alt_sgdma_construct_descriptor_burst(
> > + (struct alt_sgdma_descriptor *)&tx_desc[0],
> > + (struct alt_sgdma_descriptor *)&tx_desc[1],
> > + (uint32_t *)packet, /* read addr */
> > + (uint32_t *)0, /* */
> > + length, /* length or EOP ,will change for each
> tx */
> > + 0x1, /* gen eop */
> > + 0x0, /* read fixed */
> > + 0x1, /* write fixed or sop */
> > + 0x0, /* read burst */
> > + 0x0, /* write burst */
> > + 0x0 /* channel */
> please use tab for indent I see other in the patch please check
>
I use tab for indent, spaces for alignment. I checked the patch with
checkpatch and it
didn't find errors.....
> > + );
> > +
> > + alt_sgdma_do_sync_transfer(tx_sgdma, tx_desc_cur);
> > +
> > + return 0;;
> > +}
> > +
> > +static void tse_eth_halt(struct eth_device *edev)
> > +{
> > + struct altera_tse_priv *priv = edev->priv;
> > + struct alt_sgdma_registers *rx_sgdma = priv->sgdma_rx;
> > + struct alt_sgdma_registers *tx_sgdma = priv->sgdma_tx;
> > +
> > + writel(0, &rx_sgdma->control); /* Stop the controller and reset
> settings */
> > + writel(0, &tx_sgdma->control); /* Stop the controller and reset
> settings */
> > +}
> > +
> > +static int tse_eth_rx(struct eth_device *edev)
> > +{
> > + uint16_t packet_length = 0;
> > +
> > + struct altera_tse_priv *priv = edev->priv;
> > + struct alt_sgdma_descriptor *rx_desc = (struct alt_sgdma_descriptor
> *)priv->rx_desc;
> > + struct alt_sgdma_descriptor *rx_desc_cur = &rx_desc[0];
> > + struct alt_sgdma_registers *rx_sgdma = priv->sgdma_rx;
> > +
> > + if (rx_desc_cur->descriptor_status &
> > + ALT_SGDMA_DESCRIPTOR_STATUS_TERMINATED_BY_EOP_MSK) {
> > +
> > + packet_length = rx_desc->actual_bytes_transferred;
> > + net_receive(NetRxPackets[0], packet_length);
> > +
> > + /* Clear Run */
> > + rx_sgdma->control = (rx_sgdma->control &
> (~ALT_SGDMA_CONTROL_RUN_MSK));
> > +
> > + /* start descriptor again */
> > + flush_dcache_range((uint32_t)(NetRxPackets[0]),
> (uint32_t)(NetRxPackets[0]) + PKTSIZE);
> > + alt_sgdma_construct_descriptor_burst(
> > + (struct alt_sgdma_descriptor *)&rx_desc[0],
> > + (struct alt_sgdma_descriptor *)&rx_desc[1],
> > + (uint32_t)0x0, /* read addr */
> > + (uint32_t *)NetRxPackets[0], /* */
> > + 0x0, /* length or EOP */
> > + 0x0, /* gen eop */
> > + 0x0, /* read fixed */
> > + 0x0, /* write fixed or sop
> */
> > + 0x0, /* read burst */
> > + 0x0, /* write burst */
> > + 0x0 /* channel */
> please use tab for indent
> > + );
> > +
> > + /* setup the sgdma */
> > + alt_sgdma_do_async_transfer(priv->sgdma_rx, &rx_desc[0]);
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static char *get_phy_name(uint32_t phy_ID)
> > +{
> > + uint8_t i = 0;
> > +
> > + while (phy_name_table[i].phy_ID != 0) {
> > + if (phy_name_table[i].phy_ID == phy_ID)
> > + return phy_name_table[i].name;
> > + i++;
> > + }
> > +
> > + return "unknown";
> > +}
> why this?
> > +
> > +static int check_phy_address(struct eth_device *edev)
> > +{
> > + struct altera_tse_priv *priv = edev->priv;
> > + uint8_t mii_bus_scan = 0;
> > + uint8_t current_mii_address = 1;
> > + uint32_t phy_reg;
> > + uint32_t phy_ID;
> > +
> > + if ((priv->miidev)->address < 0)
> > + mii_bus_scan = 1;
> > + else
> > + current_mii_address = (priv->miidev)->address;
> > +
> > + do {
> > + /* Grab the bits from PHYIR1, and put them in the upper
> half */
> > + phy_reg = tse_phy_read(priv->miidev, current_mii_address,
> MII_PHYSID1);
> > + phy_ID = (phy_reg & 0xffff) << 16;
> > +
> > + /* Grab the bits from PHYIR2, and put them in the lower
> half */
> > + phy_reg = tse_phy_read(priv->miidev, current_mii_address,
> MII_PHYSID2);
> > + phy_ID |= (phy_reg & 0xffff);
> > +
> > + if (phy_ID != 0xffffffff) {
> > + printf("PHY (%s) found at address %d\n",
> get_phy_name(phy_ID), current_mii_address);
> > + (priv->miidev)->address = current_mii_address;
> > + return 0;
> > + }
> > +
> > + current_mii_address++;
> > +
> > + } while (current_mii_address < 32 && mii_bus_scan);
> > +
> > + return -1;
> > +}
> > +
> why this?
>
Because sometimes (often ?), hardware doesn't run first time. So I like to
have some information on MII bus running correctly....One day I had a PHY
getting the address 1 and sometimes this f***ing PHY got 31 and this
kind of auto scan helped me to immediately find the problem......
I could remove those function if it is a problem....
>
> Best Regards,
> J.
>
Regards,
Franck.
[-- Attachment #1.2: Type: text/html, Size: 12005 bytes --]
[-- Attachment #2: Type: text/plain, Size: 149 bytes --]
_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox
next prev parent reply other threads:[~2011-04-10 7:13 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-04-09 14:08 franck.jullien
2011-04-09 14:08 ` franck.jullien
2011-04-10 4:06 ` Jean-Christophe PLAGNIOL-VILLARD
2011-04-10 7:13 ` Franck JULLIEN [this message]
2011-04-10 10:51 ` Jean-Christophe PLAGNIOL-VILLARD
2011-04-10 18:21 ` Franck JULLIEN
2011-04-11 2:37 ` Jean-Christophe PLAGNIOL-VILLARD
2011-04-11 7:23 ` Franck JULLIEN
2011-04-11 7:47 ` Sascha Hauer
2011-04-11 7:52 ` Franck JULLIEN
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='BANLkTi=JnLBhDeYY_eXRtxb9b_3Z=2rS3A@mail.gmail.com' \
--to=franck.jullien@gmail.com \
--cc=barebox@lists.infradead.org \
--cc=plagnioj@jcrosoft.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox