| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <AANLkTimrAq_eBjt7h+PQbMuiDYypt7Pd618Xbq0R70HA@mail.gmail.com>
Date: Thu, 26 Aug 2010 13:26:22 -0600
From: Grant Likely <grant.likely@...retlab.ca>
To: Ville Sundell <ville.sundell@...ilent.ro>
Cc: davem@...emloft.net, john.linn@...inx.com, dkirjanov@...nel.org,
jpirko@...hat.com, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, devicetree-discuss@...ts.ozlabs.org,
Joshua Pederson <josh@...ilentinc.com>
Subject: Re: [PATCH] drivers/net/: ll_temac_main.c & ll_temac.h: Adding
ethtool interface
On Thu, Aug 26, 2010 at 12:13 PM, Ville Sundell
<ville.sundell@...ilent.ro> wrote:
> This patch adds an interface to the ethtool subsystem and
> implements read/write functionality for handling settings with the
> ethtool_cmd structure. "Speed" is the only setting implemented so far.
> Other settings (like reading the duplex value) can be easily implemented
> later on.
>
> One can read the speed of the link by reading /sys/class/net/eth0/speed.
> It seems that you can't write to /sys/class/net/eth0/speed, so the
> "ethtool" command line utility must be used to set the speed.
>
> So far, it is tested only in the DMA mode, since I don't have hardware
> with DCR available.
>
> Signed-off-by: Ville Sundell <ville.sundell@...ilent.ro>
The ll_temac driver uses phylib. The Ethtool ops should then also be
implemented using phylib and not by writing to the ll_temac registers
directly. Otherwise changes to the ll_temac speed will not be
reflected in the PHY.
g.
> ---
>
> When using a non-static PHY and 100Mbit connection in Temac, the speed
> is not changed and stays at the reset value (1000Mbit). This leads to a
> situation where data is not sent or received correctly since
> temac_adjust_link() is not called due to lack of static PHY. Therefore
> the link speed retains the reset value.
>
> There is also no possibility to read or set the link speed from the user
> space.
>
> The patch is done against the newest version of ll_temac_main.cc and
> ll_temac.h of the linux-next tree.
>
> Thanks to Michal Simek for his feedback on this patch!
>
> Regards,
> Ville Sundell
> Digilent Inc.
>
> drivers/net/ll_temac.h | 3 ++
> drivers/net/ll_temac_main.c | 76 +++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 79 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/net/ll_temac.h b/drivers/net/ll_temac.h
> index 522abe2..ddd00cc 100644
> --- a/drivers/net/ll_temac.h
> +++ b/drivers/net/ll_temac.h
> @@ -247,6 +247,9 @@ This option defaults to enabled (set) */
> #define XTE_EMCFG_LINKSPD_10 0x00000000 /* 10 Mbit LINKSPD_MASK */
> #define XTE_EMCFG_LINKSPD_100 (1 << 30) /* 100 Mbit LINKSPD_MASK */
> #define XTE_EMCFG_LINKSPD_1000 (1 << 31) /* 1000 Mbit LINKSPD_MASK */
> +#define XTE_EMCFG_LINKSPD_NA 0xC0000000 /* For undefined speed */
> +
> +#define TEMAC_SPEED_NA 0 /* To indicate XTE_EMCFG_LINKSPD_NA */
>
> #define XTE_GMIC_OFFSET 0x00000320 /* RGMII/SGMII config */
> #define XTE_MC_OFFSET 0x00000340 /* MDIO configuration */
> diff --git a/drivers/net/ll_temac_main.c b/drivers/net/ll_temac_main.c
> index bdf2149..ae08fd4 100644
> --- a/drivers/net/ll_temac_main.c
> +++ b/drivers/net/ll_temac_main.c
> @@ -951,6 +951,81 @@ static const struct attribute_group temac_attr_group = {
> .attrs = temac_device_attrs,
> };
>
> +/* Ethtool interface implementation
> + * SETTINGS
> + * With the following functions below, you can
> + * read and set following settings:
> + * Speed:
> + * Speed of the TEMAC controller is indicated by two most significant bits
> + * in the XTE_EMCFG_OFFSET register in the follwoing way:
> + * 00 = 10 Mbit connection
> + * 01 = 100 Mbit connection
> + * 10 = 1000 Mbit connection
> + * 11 = Undefined connection speed
> +*/
> +
> +static int temac_get_settings(struct net_device *ndev, struct ethtool_cmd *cmd)
> +{
> + struct temac_local *lp = netdev_priv(ndev);
> + u32 reg = temac_indirect_in32(lp, XTE_EMCFG_OFFSET);
> +
> + switch (reg & XTE_EMCFG_LINKSPD_MASK) {
> + case XTE_EMCFG_LINKSPD_10:
> + cmd->speed = SPEED_10;
> + break;
> + case XTE_EMCFG_LINKSPD_100:
> + cmd->speed = SPEED_100;
> + break;
> + case XTE_EMCFG_LINKSPD_1000:
> + cmd->speed = SPEED_1000;
> + break;
> + default:
> + cmd->speed = TEMAC_SPEED_NA; /* Undefined speed */
> + break;
> + }
> +
> + /* TODO: Insert more setting reading code here */
> +
> + return 0;
> +}
> +
> +static int temac_set_settings(struct net_device *ndev, struct ethtool_cmd *cmd)
> +{
> + struct temac_local *lp = netdev_priv(ndev);
> + u32 reg = temac_indirect_in32(lp, XTE_EMCFG_OFFSET);
> +
> + reg &= ~XTE_EMCFG_LINKSPD_MASK;
> + switch (cmd->speed) {
> + case SPEED_10:
> + reg |= XTE_EMCFG_LINKSPD_10;
> + break;
> + case SPEED_100:
> + reg |= XTE_EMCFG_LINKSPD_100;
> + break;
> + case SPEED_1000:
> + reg |= XTE_EMCFG_LINKSPD_1000;
> + break;
> + default:
> + reg |= XTE_EMCFG_LINKSPD_NA; /* Undefined speed */
> + break;
> + }
> +
> + /* TODO: Insert more setting writing code here */
> +
> + temac_indirect_out32(lp, XTE_EMCFG_OFFSET, reg);
> +
> + return 0;
> +}
> +
> +/*
> + * Defining the ethtool operations.
> + * This will be passed to the ethtool subsystem in the probe function
> + */
> +static const struct ethtool_ops temac_ethtool_ops = {
> + .get_settings = temac_get_settings,
> + .set_settings = temac_set_settings,
> +};
> +
> static int __init
> temac_of_probe(struct platform_device *op, const struct of_device_id *match)
> {
> @@ -970,6 +1045,7 @@ temac_of_probe(struct platform_device *op, const struct of_device_id *match)
> ether_setup(ndev);
> dev_set_drvdata(&op->dev, ndev);
> SET_NETDEV_DEV(ndev, &op->dev);
> + SET_ETHTOOL_OPS(ndev, &temac_ethtool_ops);
> ndev->flags &= ~IFF_MULTICAST; /* clear multicast */
> ndev->features = NETIF_F_SG | NETIF_F_FRAGLIST;
> ndev->netdev_ops = &temac_netdev_ops;
>
>
>
--
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists