From: "Matt Carlson" <mcarlson@broadcom.com>
To: "Micha?? Miros??aw" <mirq-linux@rere.qmqm.pl>
Cc: "netdev@vger.kernel.org" <netdev@vger.kernel.org>,
"e1000-devel@lists.sourceforge.net"
<e1000-devel@lists.sourceforge.net>,
"Steve Glendinning" <steve.glendinning@smsc.com>,
"Greg Kroah-Hartman" <gregkh@suse.de>,
"Rasesh Mody" <rmody@brocade.com>,
"Debashis Dutt" <ddutt@brocade.com>,
"Kristoffer Glembo" <kristoffer@gaisler.com>,
"linux-driver@qlogic.com" <linux-driver@qlogic.com>,
"linux-net-drivers@solarflare.com"
<linux-net-drivers@solarflare.com>
Subject: Re: [PATCH 3/4] Ethtool: convert get_tso/set_tso calls to hw_features flags
Date: Mon, 1 Nov 2010 19:49:57 -0700 [thread overview]
Message-ID: <20101102024957.GB4243@mcarlson.broadcom.com> (raw)
In-Reply-To: <db0a20ab6e4ab3f63d782bd2c4a9f7a873ea4a64.1288496404.git.mirq-linux@rere.qmqm.pl>
On Sat, Oct 30, 2010 at 01:44:17AM -0700, Micha?? Miros??aw wrote:
> diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c
> index b07e2d1..c08172d 100644
> --- a/drivers/net/tg3.c
> +++ b/drivers/net/tg3.c
> @@ -6142,7 +6142,7 @@ static inline void tg3_set_mtu(struct net_device *dev, struct tg3 *tp,
> if (new_mtu > ETH_DATA_LEN) {
> if (tp->tg3_flags2 & TG3_FLG2_5780_CLASS) {
> tp->tg3_flags2 &= ~TG3_FLG2_TSO_CAPABLE;
> - ethtool_op_set_tso(dev, 0);
> + dev->features &= ~NETIF_F_ALL_TSO;
> } else {
> tp->tg3_flags |= TG3_FLAG_JUMBO_RING_ENABLE;
> }
> @@ -9977,27 +9977,28 @@ static int tg3_set_tso(struct net_device *dev, u32 value)
> {
> struct tg3 *tp = netdev_priv(dev);
>
> - if (!(tp->tg3_flags2 & TG3_FLG2_TSO_CAPABLE)) {
> - if (value)
> - return -EINVAL;
> + if (!(tp->tg3_flags2 & TG3_FLG2_TSO_CAPABLE) && value)
> + return -EINVAL;
> +
> + if (!value)
> return 0;
> - }
> +
> + dev->features |= NETIF_F_TSO;
> +
> if ((dev->features & NETIF_F_IPV6_CSUM) &&
> ((tp->tg3_flags2 & TG3_FLG2_HW_TSO_2) ||
> (tp->tg3_flags2 & TG3_FLG2_HW_TSO_3))) {
> - if (value) {
> - dev->features |= NETIF_F_TSO6;
> - if ((tp->tg3_flags2 & TG3_FLG2_HW_TSO_3) ||
> - GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5761 ||
> - (GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5784 &&
> - GET_CHIP_REV(tp->pci_chip_rev_id) != CHIPREV_5784_AX) ||
> - GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5785 ||
> - GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_57780)
> - dev->features |= NETIF_F_TSO_ECN;
> - } else
> - dev->features &= ~(NETIF_F_TSO6 | NETIF_F_TSO_ECN);
> + dev->features |= NETIF_F_TSO6;
> + if ((tp->tg3_flags2 & TG3_FLG2_HW_TSO_3) ||
> + GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5761 ||
> + (GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5784 &&
> + GET_CHIP_REV(tp->pci_chip_rev_id) != CHIPREV_5784_AX) ||
> + GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5785 ||
> + GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_57780)
> + dev->features |= NETIF_F_TSO_ECN;
> }
> - return ethtool_op_set_tso(dev, value);
> +
> + return 1;
dev->hw_features looks to me like it should function as a set of flags
indicating what the hardware is currently capable of doing. It would
clean the above code a lot if we could do:
dev->features |= (dev->hw_features & NETIF_F_ALL_TSO);
In fact, the new member might serve as a replacement for the
TG3_FLG2_TSO_CAPABLE flag. Let's not do that right now though,
because it requires some careful attention in other code paths which would
be a distraction. I'll follow-up with a patch to do this.
> }
>
> static int tg3_nway_reset(struct net_device *dev)
> @@ -11306,7 +11307,7 @@ static const struct ethtool_ops tg3_ethtool_ops = {
> .get_rx_csum = tg3_get_rx_csum,
> .set_rx_csum = tg3_set_rx_csum,
> .set_tx_csum = tg3_set_tx_csum,
> - .set_tso = tg3_set_tso,
> + .hw_set_tso = tg3_set_tso,
> .self_test = tg3_self_test,
> .get_strings = tg3_get_strings,
> .phys_id = tg3_phys_id,
> @@ -14681,6 +14682,7 @@ static int __devinit tg3_init_one(struct pci_dev *pdev,
> tp->rx_jumbo_pending = TG3_DEF_RX_JUMBO_RING_PENDING;
>
> dev->hw_features |= NETIF_F_SG;
> + dev->hw_features |= NETIF_F_TSO|NETIF_F_TSO6|NETIF_F_TSO_ECN;
Not all devices are TSO capable either. There is code later in this
function that determines which TSO flags to set. I would probably reuse
that code.
> dev->ethtool_ops = &tg3_ethtool_ops;
> dev->watchdog_timeo = TG3_TX_TIMEOUT;
> dev->irq = pdev->irq;
next prev parent reply other threads:[~2010-11-02 2:50 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-10-31 3:40 [PATCH 0/4] Ethtool: cleanup strategy Michał Mirosław
2010-10-30 4:27 ` [PATCH 1/4] Ethtool: Introduce hw_features field in struct netdevice Michał Mirosław
2010-10-30 4:28 ` [PATCH 2/4] Ethtool: convert get_sg/set_sg calls to hw_features flag Michał Mirosław
2010-11-01 21:15 ` Ben Hutchings
2010-11-02 0:59 ` Michał Mirosław
2010-11-02 2:24 ` Matt Carlson
2010-11-03 22:29 ` Micha?? Miros??aw
2010-11-03 22:42 ` Matt Carlson
2010-11-03 22:58 ` Michał Mirosław
2010-10-30 8:44 ` [PATCH 3/4] Ethtool: convert get_tso/set_tso calls to hw_features flags Michał Mirosław
2010-11-01 21:25 ` Ben Hutchings
2010-11-02 1:14 ` Michał Mirosław
2010-11-02 2:49 ` Matt Carlson [this message]
2010-10-31 0:09 ` [PATCH 4/4] Ethtool: convert get_tx_csum/set_tx_csum " Michał Mirosław
2010-11-01 21:38 ` Ben Hutchings
2010-11-02 1:23 ` Michał Mirosław
2010-10-31 4:18 ` [PATCH 0/4] Ethtool: cleanup strategy David Miller
2010-10-31 4:30 ` Michał Mirosław
2010-11-01 21:05 ` Ben Hutchings
2010-11-02 1:02 ` Michał Mirosław
2010-11-02 1:14 ` Ben Hutchings
2010-11-02 1:30 ` Michał Mirosław
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=20101102024957.GB4243@mcarlson.broadcom.com \
--to=mcarlson@broadcom.com \
--cc=ddutt@brocade.com \
--cc=e1000-devel@lists.sourceforge.net \
--cc=gregkh@suse.de \
--cc=kristoffer@gaisler.com \
--cc=linux-driver@qlogic.com \
--cc=linux-net-drivers@solarflare.com \
--cc=mirq-linux@rere.qmqm.pl \
--cc=netdev@vger.kernel.org \
--cc=rmody@brocade.com \
--cc=steve.glendinning@smsc.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;
as well as URLs for NNTP newsgroup(s).