From: Ajay Bhargav <ajay.bhargav@einfochips.com>
To: u-boot@lists.denx.de
Subject: [U-Boot] [PATCH v2 1/3] net: Adds Fast Ethernet Controller driver for Armada100
Date: Thu, 25 Aug 2011 10:51:25 +0530 (IST) [thread overview]
Message-ID: <316561839.535.1314249685865.JavaMail.root@ahm.einfochips.com> (raw)
In-Reply-To: <569776325.496.1314249408394.JavaMail.root@ahm.einfochips.com>
----- "Marek Vasut" <marek.vasut@gmail.com> wrote:
> On Wednesday, August 24, 2011 03:07:18 PM Ajay Bhargav wrote:
> > This patch adds support for Fast Ethernet Controller driver for
> > Armada100 series.
> >
> > Signed-off-by: Ajay Bhargav <ajay.bhargav@einfochips.com>
> > ---
>
> [...]
>
[...]
> > +
> > +static int armdfec_phy_timeout(u32 reg, u32 flag, int cond)
> > +{
> > + u32 timeout = PHY_WAIT_ITERATIONS;
> > + while (--timeout) {
> > + if (cond && (readl(reg) & flag))
> > + break;
> > + else if (!cond && !(readl(reg) & flag))
>
> You can read the register into some temporary variable so you don't
> need the
> readl() at two places.
>
readl will be called only once... do i really need a temp var?
> > + break;
> > + udelay(PHY_WAIT_MICRO_SECONDS);
> > + }
> > + return !timeout;
> > +}
> > +
> > +static int smi_reg_read(const char *devname, u8 phy_addr, u8
> phy_reg,
> > + u16 *value)
> > +{
> > + struct eth_device *dev = eth_get_dev_by_name(devname);
> > + struct armdfec_device *darmdfec = to_darmdfec(dev);
> > + struct armdfec_reg *regs = darmdfec->regs;
> > + u32 val, reg_data;
> > +
> > + if (phy_addr == PHY_ADR_REQ && phy_reg == PHY_ADR_REQ) {
> > + reg_data = readl(®s->phyadr);
> > + *value = (u16) (reg_data & 0x1f);
> > + return 0;
> > + }
> > +
> > + /* check parameters */
> > + if (phy_addr > PHY_MASK) {
> > + printf("Err..(%s) Invalid phy address: 0x%X\n",
> > + __func__, phy_addr);
> > + return -EINVAL;
> > + }
> > + if (phy_reg > PHY_MASK) {
> > + printf("Err..(%s) Invalid register offset: 0x%X\n",
> > + __func__, phy_reg);
> > + return -EINVAL;
> > + }
> > +
> > + /* wait for the SMI register to become available */
> > + if (armdfec_phy_timeout((u32)®s->smi, SMI_BUSY, FALSE)) {
>
> Adjust the function so you don't need the cast.
>
okay...
> > + printf("Error (%s) PHY busy timeout\n", __func__);
> > + return -1;
> > + }
> > +
> > + writel(phy_addr << 16 | phy_reg << 21 | SMI_OP_R, ®s->smi);
>
> Parentheses missing maybe ?
>
:) okay..
> > +
> > + /* now wait for the data to be valid */
> > + if (armdfec_phy_timeout((u32)®s->smi, SMI_R_VALID, TRUE)) {
> > + val = readl(®s->smi);
> > + printf("Err (%s) PHY Read timeout, val=0x%x\n", __func__, val);
> > + return -1;
> > + }
> > + val = readl(®s->smi);
> > + *value = val & 0xffff;
> > +
> > + return 0;
> > +}
> > +
> > +static int smi_reg_write(const char *devname,
> > + u8 phy_addr, u8 phy_reg, u16 value)
> > +{
> > + struct eth_device *dev = eth_get_dev_by_name(devname);
> > + struct armdfec_device *darmdfec = to_darmdfec(dev);
> > + struct armdfec_reg *regs = darmdfec->regs;
> > +
> > + if (phy_addr == PHY_ADR_REQ && phy_reg == PHY_ADR_REQ) {
> > + clrsetbits_le32(®s->phyadr, 0x1f, value & 0x1f);
> > + return 0;
> > + }
> > +
> > + /* check parameters */
> > + if (phy_addr > PHY_MASK) {
> > + printf("Err..(%s) Invalid phy address\n", __func__);
> > + return -EINVAL;
> > + }
> > + if (phy_reg > PHY_MASK) {
> > + printf("Err..(%s) Invalid register offset\n", __func__);
> > + return -EINVAL;
> > + }
> > +
> > + /* wait for the SMI register to become available */
> > + if (armdfec_phy_timeout((u32)®s->smi, SMI_BUSY, FALSE)) {
> > + printf("Error (%s) PHY busy timeout\n", __func__);
> > + return -1;
> > + }
> > +
> > + writel(phy_addr << 16 | phy_reg << 21 | SMI_OP_W | (value &
> 0xffff),
> > + ®s->smi);
> > + return 0;
> > +}
> > +
> > +/*
> > + * Abort any transmit and receive operations and put DMA
> > + * in idle state. AT and AR bits are cleared upon entering
> > + * in IDLE state. So poll those bits to verify operation.
> > + */
> > +static void abortdma(struct eth_device *dev)
> > +{
> > + struct armdfec_device *darmdfec = to_darmdfec(dev);
> > + struct armdfec_reg *regs = darmdfec->regs;
> > + int delay;
> > + int maxretries = 40;
> > +
> > + do {
> > + writel(SDMA_CMD_AR | SDMA_CMD_AT, ®s->sdma_cmd);
> > + udelay(100);
> > +
> > + delay = 10;
> > + while ((readl(®s->sdma_cmd) &
> > + (SDMA_CMD_AR | SDMA_CMD_AT))
> > + && delay-- > 0) {
> > + udelay(10);
> > + }
> > + } while (maxretries-- > 0 && delay <= 0);
>
> Didn't I comment on this one in V1?
>
Yes you did...
> > +
> > + if (maxretries <= 0)
> > + printf("%s : DMA Stuck\n", __func__);
> > +}
>
> [...]
>
> > +int armada100_fec_register()
> > +{
> > + struct armdfec_device *darmdfec;
> > + struct eth_device *dev;
> > + int phy_adr;
> > +
> > + darmdfec = malloc(sizeof(struct armdfec_device));
> > + if (!darmdfec)
> > + goto error;
> > +
> > + memset(darmdfec, 0, sizeof(struct armdfec_device));
> > +
> > + darmdfec->htpr = memalign(8, HASH_ADDR_TABLE_SIZE);
> > + if (!darmdfec->htpr)
> > + goto error;
> > +
> > + darmdfec->p_rxdesc = (struct rx_desc *) memalign(PKTALIGN,
> > + ARMDFEC_RXQ_DESC_ALIGNED_SIZE * RINGSZ + 1);
> > +
> > + if (!darmdfec->p_rxdesc)
> > + goto error;
> > +
> > + darmdfec->p_rxbuf = (u8 *) memalign(PKTALIGN,
> > + RINGSZ * PKTSIZE_ALIGN + 1);
>
> Is the cast needed ? You should review your casts, most of them can be
> solved by
> using proper type!
>
Thats a serious issue with me I guess :)
> > + if (!darmdfec->p_rxbuf)
> > + goto error;
> > +
> > + darmdfec->p_aligned_txbuf = memalign(8, PKTSIZE_ALIGN);
> > + if (!darmdfec->p_aligned_txbuf)
> > + goto error;
> > +
> > + darmdfec->p_txdesc = (struct tx_desc *)
> > + memalign(PKTALIGN, sizeof(struct tx_desc) + 1);
> > + if (!darmdfec->p_txdesc)
> > + goto error;
> > +
> > + dev = &darmdfec->dev;
> > + /* Assign ARMADA100 Fast Ethernet Controller Base Address */
> > + darmdfec->regs = (void *) ARMD1_FEC_BASE;
>
> Cheers!
>
Thanks,
Ajay Bhargav
next parent reply other threads:[~2011-08-25 5:21 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <569776325.496.1314249408394.JavaMail.root@ahm.einfochips.com>
2011-08-25 5:21 ` Ajay Bhargav [this message]
2011-08-25 12:15 ` [U-Boot] [PATCH v2 1/3] net: Adds Fast Ethernet Controller driver for Armada100 Marek Vasut
[not found] <1826454834.6634.1314336622428.JavaMail.root@ahm.einfochips.com>
2011-08-26 5:33 ` Ajay Bhargav
[not found] <583672539.2717.1314270415380.JavaMail.root@ahm.einfochips.com>
2011-08-25 11:07 ` Ajay Bhargav
2011-08-25 12:19 ` Marek Vasut
2011-08-25 12:12 ` Ajay Bhargav
2011-08-24 13:07 Ajay Bhargav
2011-08-24 15:42 ` Mike Frysinger
2011-08-24 16:38 ` Marek Vasut
2011-08-25 5:24 ` Ajay Bhargav
2011-08-25 12:14 ` Marek Vasut
2011-08-25 5:10 ` Ajay Bhargav
2011-08-25 14:13 ` Mike Frysinger
2011-08-24 16:36 ` Marek Vasut
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=316561839.535.1314249685865.JavaMail.root@ahm.einfochips.com \
--to=ajay.bhargav@einfochips.com \
--cc=u-boot@lists.denx.de \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.