netdev - Re: [PATCH 2/4] Ethtool: convert get_sg/set_sg calls to hw

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20101103225846.GA32330@rere.qmqm.pl>
Date:	Wed, 3 Nov 2010 23:58:46 +0100
From:	Michał Mirosław <mirq-linux@...e.qmqm.pl>
To:	Matt Carlson <mcarlson@...adcom.com>
Cc:	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"e1000-devel@...ts.sourceforge.net" 
	<e1000-devel@...ts.sourceforge.net>,
	Steve Glendinning <steve.glendinning@...c.com>,
	Greg Kroah-Hartman <gregkh@...e.de>,
	Rasesh Mody <rmody@...cade.com>,
	Debashis Dutt <ddutt@...cade.com>,
	Kristoffer Glembo <kristoffer@...sler.com>,
	"linux-driver@...gic.com" <linux-driver@...gic.com>,
	"linux-net-drivers@...arflare.com" <linux-net-drivers@...arflare.com>
Subject: Re: [PATCH 2/4] Ethtool: convert get_sg/set_sg calls to
	hw_features flag

On Wed, Nov 03, 2010 at 03:42:47PM -0700, Matt Carlson wrote:
> On Wed, Nov 03, 2010 at 03:29:10PM -0700, Micha?? Miros??aw wrote:
> > On Mon, Nov 01, 2010 at 07:24:38PM -0700, Matt Carlson wrote:
> > > On Fri, Oct 29, 2010 at 09:28:26PM -0700, Micha?? Miros??aw wrote:
> > > > diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c
> > > > index 30ccbb6..b07e2d1 100644
> > > > --- a/drivers/net/tg3.c
> > > > +++ b/drivers/net/tg3.c
> > > > @@ -11306,7 +11306,6 @@ 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_sg                 = ethtool_op_set_sg,

This is exchanged ...

> > > >         .set_tso                = tg3_set_tso,
> > > >         .self_test              = tg3_self_test,
> > > >         .get_strings            = tg3_get_strings,
> > > > @@ -14681,6 +14680,7 @@ static int __devinit tg3_init_one(struct pci_dev *pdev,
> > > >         tp->rx_pending = TG3_DEF_RX_RING_PENDING;
> > > >         tp->rx_jumbo_pending = TG3_DEF_RX_JUMBO_RING_PENDING;
> > > > 
> > > > +       dev->hw_features |= NETIF_F_SG;

... for this. (This introduces no functional changes, whatsoever.)

> > > Scatter-gather should not be enabled if TG3_FLAG_BROKEN_CHECKSUMS is set.  I
> > > would do the following instead:
> > > 
> > > 	if (!(tp->tg3_flags & TG3_FLAG_BROKEN_CHECKSUMS))
> > > 		dev->hw_features |= NETIF_F_SG;
> > > 
> > > TG3_FLAG_BROKEN_CHECKSUMS is set in tg3_get_invariants(), so this code
> > > would need to be placed later than that function call.
> > 
> > This bug is there now, so I'll queue this as all other hints of existent
> > bugs that this patch series "uncovers".
> 
> How so?  This patch would be introducing the bug.  From tg3_get_invariants:
> 
> 	if (tp->pci_chip_rev_id == CHIPREV_ID_5700_B0)
> 		tp->tg3_flags |= TG3_FLAG_BROKEN_CHECKSUMS;
> 	else {
> 		unsigned long features = NETIF_F_IP_CSUM | NETIF_F_SG | NETIF_F_GRO;
> 
> 		tp->tg3_flags |= TG3_FLAG_RX_CHECKSUMS;
> 		if (tp->tg3_flags3 & TG3_FLG3_5755_PLUS)
> 			features |= NETIF_F_IPV6_CSUM;
> 		tp->dev->features |= features;
> 		vlan_features_add(tp->dev, features);
> 	}

Actually this is hidden anyway, because currently scatter-gather depends
on checksumming offload. So the SG won't be enabled based on check left
in ethtool_set_sg().

Note, that hw_features flags enable toggling of the offloads but don't
enable them unless requested by a user later.

Best Regards,
Michał Mirosław
--
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