* Re: [RFC] iproute2: add br command
From: Michał Mirosław @ 2011-10-04 22:32 UTC (permalink / raw)
To: Sridhar Samudrala; +Cc: Stephen Hemminger, Andi Kleen, Kevin Wilson, netdev
In-Reply-To: <1317766638.21658.11.camel@w-sridhar.beaverton.ibm.com>
2011/10/5 Sridhar Samudrala <sri@us.ibm.com>:
> On Tue, 2011-10-04 at 14:07 -0700, Stephen Hemminger wrote:
>> On Tue, 04 Oct 2011 09:58:55 -0700
>> Andi Kleen <andi@firstfloor.org> wrote:
>> > Stephen Hemminger <shemminger@vyatta.com> writes:
>> > > This adds a new 'br' command which is the bridging equivalent of
>> > > the ip command. More of a demo of how to use netlink and bridging
>> > > at this point.
>> > Please name it "bridge", not "br"
>> Ok, but it breaks the existing pattern.
> Is this supposed to replace brctl utility?
>
> Can we add/delete a bridge and add/delete interfaces to a bridge using
> this command?
>
> If so, why not make it
> ip bridge add/del <brname>
> ip bridge addif/delif <brname> <ifname>
I'll add one more idea:
ip link add/del <brname> type bridge
ip bridge add/del <brname> <ifname>
ip bridge fdb ...
Best Regards,
Michał Mirosław
^ permalink raw reply
* Re: big picture UDP/IP performance question re 2.6.18 -> 2.6.32
From: Con Kolivas @ 2011-10-04 22:32 UTC (permalink / raw)
To: Serge Belyshev
Cc: Christoph Lameter, Joe Perches, starlight, Eric Dumazet,
linux-kernel, netdev, Willy Tarreau, Peter Zijlstra
In-Reply-To: <87zkhg1ppl.fsf@depni.sinp.msu.ru>
On Wed, 5 Oct 2011 07:12:06 Serge Belyshev wrote:
> BFS FAQ writes:
> > NUMA aware?
> >
> > It is NOT NUMA aware in the sense that it does any fancy shit on NUMA,
> > but it will work on NUMA hardware just fine. Only the really big NUMA
> > hardware is likely to suffer in performance, and this is theoretically
> > only, since no one has that sort of hardware to prove it to me, but it
> > seems almost certain.
>
> This part of FAQ entry was written before 0.300, (i.e. more than 2 years
> ago)
>
> > v0.300 onwards have NUMA enhancements.
>
> And this was added after. =)
>
> As of now, BFS scheduling is NUMA-aware. (see resched_best_mask())
Admittedly the wording of this got confused as BFS developed. It originally
had no NUMA enhancements, but now it does. The code to benefit NUMA is quite
simple though, but then, so is everything in BFS.
Regards,
Con
--
-ck
^ permalink raw reply
* Re: [RFC] iproute2: add br command
From: Stephen Hemminger @ 2011-10-04 22:27 UTC (permalink / raw)
To: Sridhar Samudrala; +Cc: Andi Kleen, Kevin Wilson, netdev
In-Reply-To: <1317766638.21658.11.camel@w-sridhar.beaverton.ibm.com>
On Tue, 04 Oct 2011 15:17:18 -0700
Sridhar Samudrala <sri@us.ibm.com> wrote:
> On Tue, 2011-10-04 at 14:07 -0700, Stephen Hemminger wrote:
> > On Tue, 04 Oct 2011 09:58:55 -0700
> > Andi Kleen <andi@firstfloor.org> wrote:
> >
> > > Stephen Hemminger <shemminger@vyatta.com> writes:
> > >
> > > > This adds a new 'br' command which is the bridging equivalent of
> > > > the ip command. More of a demo of how to use netlink and bridging
> > > > at this point.
> > >
> > > Please name it "bridge", not "br"
> > >
> > > -Andi
> >
> > Ok, but it breaks the existing pattern.
>
> Is this supposed to replace brctl utility?
>
> Can we add/delete a bridge and add/delete interfaces to a bridge using
> this command?
>
> If so, why not make it
> ip bridge add/del <brname>
> ip bridge addif/delif <brname> <ifname>
>
You can already use 'ip link' to add/delete bridge, and add interface
to bridge by setting master.
^ permalink raw reply
* Re: [PATCH ethtool 3/3] ethtool: add ETHTOOL_{G,S}CHANNEL support.
From: Ben Hutchings @ 2011-10-04 22:21 UTC (permalink / raw)
To: Sucheta Chakraborty; +Cc: netdev, Dept_NX_Linux_NIC_Driver
In-Reply-To: <1316514695-17157-4-git-send-email-sucheta.chakraborty@qlogic.com>
On Tue, 2011-09-20 at 03:31 -0700, Sucheta Chakraborty wrote:
> Used to configure number of rx and tx rings.
> Reqd. man page changes are included.
[...]
> @@ -754,6 +764,24 @@ lB l.
> Specify the location/ID to insert the rule. This will overwrite
> any rule present in that location and will not go through any
> of the rule ordering process.
> +.TP
> +.B \-l \-\-show\-channels
> +Queries the specified network device for channel parameter information.
> +.TP
> +.B \-L \-\-set\-channels
> +Changes the channel parameters of the specified network device.
I think the manual page needs to explain briefly what is meant by a
channel. (So should ethtool.h, really!)
[...]
> @@ -495,6 +512,13 @@ static struct cmdline_info cmdline_ring[] = {
> { "tx", CMDL_S32, &ring_tx_wanted, &ering.tx_pending },
> };
>
> +static struct cmdline_info cmdline_channels[] = {
> + { "rx_count", CMDL_S32, &channels_rx_wanted, &echannels.rx_count },
> + { "tx_count", CMDL_S32, &channels_tx_wanted, &echannels.tx_count },
> + { "other_count", CMDL_S32, &channels_other_wanted, &echannels.other_count },
> + { "combined_count", CMDL_S32, &channels_combined_wanted, &echannels.combined_count },
> +};
I don't think it's necessary to include '_count' in these keywords.
[...]
> @@ -1751,6 +1785,32 @@ static int dump_ring(void)
> return 0;
> }
>
> +static int dump_channels(void)
> +{
> + fprintf(stdout,
> + "Re-set maximums:\n"
Should 'Re-set' be 'Pre-set'?
> + "Max RX: %u\n"
> + "Max TX: %u\n"
> + "Max Other: %u\n"
> + "Max Combined: %u\n",
> + echannels.max_rx, echannels.max_tx,
> + echannels.max_other,
> + echannels.max_combined);
The heading says they are maximums, so I don't think it's necessary to
include 'Max' on each line.
> + fprintf(stdout,
> + "Current hardware settings:\n"
> + "RX Count: %u\n"
> + "TX Count: %u\n"
> + "Other Count: %u\n"
> + "Combined Count: %u\n",
I don't think it's necessary to include 'Count' on each line here,
either.
[...]
> @@ -2114,6 +2178,58 @@ static int do_gring(int fd, struct ifreq *ifr)
> return 0;
> }
>
> +static int do_schannels(int fd, struct ifreq *ifr)
> +{
> + int err, changed = 0;
> +
> + echannels.cmd = ETHTOOL_GCHANNELS;
> + ifr->ifr_data = (caddr_t)&echannels;
> + err = send_ioctl(fd, ifr);
> + if (err) {
> + perror("Cannot get device channels settings");
> + return 1;
> + }
> +
> + do_generic_set(cmdline_channels, ARRAY_SIZE(cmdline_channels),
> + &changed);
> +
> + if (!changed) {
> + fprintf(stderr, "no channels parameters changed, aborting\n");
> + return 1;
> + }
[...]
These messages (and others below) are not grammatical. They should say
'channel settings' or 'channel parameters'.
Actually, you could be more specific about what the parameters are, and
just write 'numbers of channels'.
Ben.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [RFC] iproute2: add br command
From: Sridhar Samudrala @ 2011-10-04 22:17 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: Andi Kleen, Kevin Wilson, netdev
In-Reply-To: <20111004140754.0e677941@nehalam.linuxnetplumber.net>
On Tue, 2011-10-04 at 14:07 -0700, Stephen Hemminger wrote:
> On Tue, 04 Oct 2011 09:58:55 -0700
> Andi Kleen <andi@firstfloor.org> wrote:
>
> > Stephen Hemminger <shemminger@vyatta.com> writes:
> >
> > > This adds a new 'br' command which is the bridging equivalent of
> > > the ip command. More of a demo of how to use netlink and bridging
> > > at this point.
> >
> > Please name it "bridge", not "br"
> >
> > -Andi
>
> Ok, but it breaks the existing pattern.
Is this supposed to replace brctl utility?
Can we add/delete a bridge and add/delete interfaces to a bridge using
this command?
If so, why not make it
ip bridge add/del <brname>
ip bridge addif/delif <brname> <ifname>
Thanks
Sridhar
^ permalink raw reply
* Re: [PATCH ethtool 2/3] ethtool: add support for external loopback.
From: Ben Hutchings @ 2011-10-04 22:12 UTC (permalink / raw)
To: Sucheta Chakraborty; +Cc: netdev, Dept_NX_Linux_NIC_Driver
In-Reply-To: <1316514695-17157-3-git-send-email-sucheta.chakraborty@qlogic.com>
On Tue, 2011-09-20 at 03:31 -0700, Sucheta Chakraborty wrote:
> External loopback will be performed in addition to other offline tests.
> User need to pass new parameter "external_lb" for the same.
>
> Reqd. man page changes included.
Applied, but:
[...]
> @@ -1689,6 +1692,9 @@ static int dump_test(struct ethtool_drvinfo *info, struct ethtool_test *test,
>
> rc = test->flags & ETH_TEST_FL_FAILED;
> fprintf(stdout, "The test result is %s\n", rc ? "FAIL" : "PASS");
> + fprintf(stdout, "External loopback test is %s\n",
> + test->flags & ETH_TEST_FL_EXTERNAL_LB_DONE ? "executed" :
> + "not executed");
>
> if (info->testinfo_len)
> fprintf(stdout, "The test extra info:\n");
1. This message should say 'was', not 'is'.
2. There are undoubtedly scripts that parse the output of ethtool -t.
We might break them if we add a line to the output of existing commands.
I've made this conditional on test_type == EXTERNAL_LB.
Ben.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH ethtool 1/3] ethtool: file ethtool-copy.h update.
From: Ben Hutchings @ 2011-10-04 22:09 UTC (permalink / raw)
To: Sucheta Chakraborty; +Cc: netdev, Dept_NX_Linux_NIC_Driver
In-Reply-To: <1316514695-17157-2-git-send-email-sucheta.chakraborty@qlogic.com>
On Tue, 2011-09-20 at 03:31 -0700, Sucheta Chakraborty wrote:
> Used linux-3.0.1-net-next tree for the update.
[...]
I'll update ethtool-copy.h using the current net-next.
Ben.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: Application brokne due to XFRMA_ALG_AUTH_TRUNC
From: Ben Hutchings @ 2011-10-04 21:35 UTC (permalink / raw)
To: Eduardo Panisset; +Cc: netdev
In-Reply-To: <CADsjWXR66C2SdfmLP0wqRvbcDJkXbYBN=zzTZLP5_+pE0fenqQ@mail.gmail.com>
On Tue, 2011-10-04 at 18:08 -0300, Eduardo Panisset wrote:
> Hi All,
>
> I have an application compiled against kernel 2.6.32 and which is
> using XFRMA_MARK (value 20).
It was not defined in 2.6.32. Perhaps you are using a distribution that
backported this feature incorrectly?
> However the same aplication does not work anymore on kernel 2.6.38.8
> due to XFRMA_MARK has now the value 21 because of a new attribute
> XFRMA_ALG_AUTH_TRUNC inserted before it in the file linux/xfrm.h.
> Now the only solution would be recompiling this application. Is
> correct introducing a new parameter in the middle of an existing enum
> in the kernel headers ?
So far as I can see, in mainline Linux versions the value has always
been 21.
Ben.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [PATCH 2/2] IPVS netns shutdown/startup dead-lock (Take III)
From: Simon Horman @ 2011-10-04 21:15 UTC (permalink / raw)
To: Julian Anastasov
Cc: Hans Schillstrom, wensong, lvs-devel, netdev, netfilter-devel,
hans
In-Reply-To: <alpine.LFD.2.00.1110041032520.1547@ja.ssi.bg>
On Tue, Oct 04, 2011 at 10:40:29AM +0300, Julian Anastasov wrote:
>
> Hello,
>
> On Tue, 4 Oct 2011, Hans Schillstrom wrote:
>
> > From: Hans Schillstrom <hans@schillstrom.com>
> >
> > ip_vs_mutext is used by both netns shutdown code and startup
> > and both implicit uses sk_lock-AF_INET mutex.
> >
> > cleanup CPU-1 startup CPU-2
> > ip_vs_dst_event() ip_vs_genl_set_cmd()
> > sk_lock-AF_INET __ip_vs_mutex
> > sk_lock-AF_INET
> > __ip_vs_mutex
> > * DEAD LOCK *
> >
> > A new mutex placed in ip_vs netns struct called sync_mutex is added.
> >
> > Comments from Julian and Simon added.
> > This patch has been running for more than 3 month now and it seems to work.
> >
> > Ver. 3
> > IP_VS_SO_GET_DAEMON in do_ip_vs_get_ctl protected by sync_mutex
> > instead of __ip_vs_mutex as sugested by Julian.
> >
> > Signed-off-by: Hans Schillstrom <hans@schillstrom.com>
>
> Looks good to me, with one extra empty line in
> start_sync_thread
>
> Acked-by: Julian Anastasov <ja@ssi.bg>
Thanks, I will apply this change.
^ permalink raw reply
* Re: [PATCH 1/2] fix lockdep warning
From: Simon Horman @ 2011-10-04 21:14 UTC (permalink / raw)
To: Hans Schillstrom; +Cc: ja, wensong, lvs-devel, netdev, netfilter-devel, hans
In-Reply-To: <1317639399-4522-1-git-send-email-hans.schillstrom@ericsson.com>
On Mon, Oct 03, 2011 at 12:56:38PM +0200, Hans Schillstrom wrote:
> From: Hans Schillstrom <hans@schillstrom.com>
>
> rs_lock needs a key to make lock dep happy.
>
> Signed-off-by: Hans Schillstrom <hans.schillstrom@ericsson.com>
Hi Julian,
can I get an Ack or otherwise from you on this patch?
I already have one for
[PATCH 2/2] IPVS netns shutdown/startup dead-lock (Take III)
>
> diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
> index 2b771dc..a1af72f 100644
> --- a/net/netfilter/ipvs/ip_vs_ctl.c
> +++ b/net/netfilter/ipvs/ip_vs_ctl.c
> @@ -57,6 +57,7 @@ static DEFINE_MUTEX(__ip_vs_mutex);
>
> /* lock for service table */
> static DEFINE_RWLOCK(__ip_vs_svc_lock);
> +static struct lock_class_key ip_vs_rs_key;
>
> /* sysctl variables */
>
> @@ -3680,6 +3681,7 @@ int __net_init ip_vs_control_net_init(struct net *net)
> struct netns_ipvs *ipvs = net_ipvs(net);
>
> ipvs->rs_lock = __RW_LOCK_UNLOCKED(ipvs->rs_lock);
> + __rwlock_init(&ipvs->rs_lock, "ipvs->rs_lock", &ip_vs_rs_key);
>
> /* Initialize rs_table */
> for (idx = 0; idx < IP_VS_RTAB_SIZE; idx++)
> --
> 1.7.4.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe lvs-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply
* Application brokne due to XFRMA_ALG_AUTH_TRUNC
From: Eduardo Panisset @ 2011-10-04 21:08 UTC (permalink / raw)
To: netdev
Hi All,
I have an application compiled against kernel 2.6.32 and which is
using XFRMA_MARK (value 20).
However the same aplication does not work anymore on kernel 2.6.38.8
due to XFRMA_MARK has now the value 21 because of a new attribute
XFRMA_ALG_AUTH_TRUNC inserted before it in the file linux/xfrm.h.
Now the only solution would be recompiling this application. Is
correct introducing a new parameter in the middle of an existing enum
in the kernel headers ?
Rgds,
Eduardo Panisset.
^ permalink raw reply
* Re: [RFC] iproute2: add br command
From: Stephen Hemminger @ 2011-10-04 21:08 UTC (permalink / raw)
To: Ben Hutchings; +Cc: Kevin Wilson, netdev
In-Reply-To: <1317747490.2751.4.camel@bwh-desktop>
On Tue, 04 Oct 2011 17:58:10 +0100
Ben Hutchings <bhutchings@solarflare.com> wrote:
> On Tue, 2011-10-04 at 09:34 -0700, Stephen Hemminger wrote:
> > This adds a new 'br' command which is the bridging equivalent of
> > the ip command. More of a demo of how to use netlink and bridging
> > at this point.
> [...]
>
> This command name is already in use:
> http://packages.debian.org/sid/bottlerocket
Damn name squatters.
^ permalink raw reply
* Re: [RFC] iproute2: add br command
From: Stephen Hemminger @ 2011-10-04 21:07 UTC (permalink / raw)
To: Andi Kleen; +Cc: Kevin Wilson, netdev
In-Reply-To: <m27h4klmls.fsf@firstfloor.org>
On Tue, 04 Oct 2011 09:58:55 -0700
Andi Kleen <andi@firstfloor.org> wrote:
> Stephen Hemminger <shemminger@vyatta.com> writes:
>
> > This adds a new 'br' command which is the bridging equivalent of
> > the ip command. More of a demo of how to use netlink and bridging
> > at this point.
>
> Please name it "bridge", not "br"
>
> -Andi
Ok, but it breaks the existing pattern.
^ permalink raw reply
* Re: [PATCH] mlx4_en: fix transmit of packages when blue frame is enabled
From: Thadeu Lima de Souza Cascardo @ 2011-10-04 20:26 UTC (permalink / raw)
To: Benjamin Herrenschmidt
Cc: Yevgeny Petrilin, netdev@vger.kernel.org, Eli Cohen,
eli@dev.mellanox.co.il, linuxppc-dev
In-Reply-To: <1317708132.29415.213.camel@pasglop>
On Tue, Oct 04, 2011 at 08:02:12AM +0200, Benjamin Herrenschmidt wrote:
> On Mon, 2011-10-03 at 17:53 -0300, Thadeu Lima de Souza Cascardo wrote:
>
> .../...
>
> > > Can you also send me the output of ethtool -i?
> > > It seems that there is a problem with write combining on Power processors, we will check this issue.
> > >
> > > Yevgeny
> >
> > Hello, Yevgeny.
> >
> > You will find the output of ethtool -i below.
> >
> > I am copying Ben and powerpc list, in case this is an issue with Power
> > processors. They can provide us some more insight into this.
>
> May I get some background please ? :-)
>
> I'm not aware of a specific issue with write combining but I'd need to
> know more about what you are doing and the code to do it to comment on
> whether it should work or not.
>
> Cheers,
> Ben.
>
>
Hello, Ben.
Sorry for that. I am testing mlx4_en driver on a POWER. Yevgeny has
added blue frame support, that does not require writing to the device
memory to indicate a new packet (the doorbell register as it is called).
Well, the ring is getting full with no interrupt or packets transmitted.
I simply added a write to the doorbell register and it works for me.
Yevgeny says this is not the right fix, claiming there is a problem with
write combining on POWER. The code uses memory barriers, so I don't know
why there is any problem.
I am posting the code here to show better what the situation is.
Yevgeny can tell more about the device and the driver.
The code below is the driver as of now, including a diff with what I
changed and had resulted OK for me. Before the blue frame support, the
only code executed was the else part. I can't tell much what the device
should be seeing and doing after the blue frame part of the code is
executed. But it does send the packet if I write to the doorbell
register.
Yevgeny, can you tell us what the device should be doing and why you
think this is a problem on POWER? Is it possible that this is simply a
problem with the firmware version?
Thanks,
Cascardo.
---
if (ring->bf_enabled && desc_size <= MAX_BF && !bounce &&
!vlan_tag) {
*(u32 *) (&tx_desc->ctrl.vlan_tag) |=
ring->doorbell_qpn;
op_own |= htonl((bf_index & 0xffff) << 8);
/* Ensure new descirptor hits memory
* before setting ownership of this descriptor to HW */
wmb();
tx_desc->ctrl.owner_opcode = op_own;
wmb();
mlx4_bf_copy(ring->bf.reg + ring->bf.offset, (unsigned
long *) &tx_desc->ctrl,
desc_size);
wmb();
ring->bf.offset ^= ring->bf.buf_size;
} else {
/* Ensure new descirptor hits memory
* before setting ownership of this descriptor to HW */
wmb();
tx_desc->ctrl.owner_opcode = op_own;
- wmb();
- writel(ring->doorbell_qpn, ring->bf.uar->map +
MLX4_SEND_DOORBELL);
}
+ wmb();
+ writel(ring->doorbell_qpn, ring->bf.uar->map +
MLX4_SEND_DOORBELL);
+
---
^ permalink raw reply
* Re: [PATCH 3/3] NET: white space/coding style cleanup of asix driver
From: David Miller @ 2011-10-04 20:26 UTC (permalink / raw)
To: grundler; +Cc: netdev, linux-kernel, allan
In-Reply-To: <1317758118-24052-3-git-send-email-grundler@chromium.org>
From: Grant Grundler <grundler@chromium.org>
Date: Tue, 4 Oct 2011 12:55:18 -0700
> check patch was complaining...mostly replaced:
> if ((ret = asix_foo(xx)) < 0) ...
> with
> ret = asix_foo(xx);
> if (ret < 0) ...
>
> Signed-off-by: Grant Grundler <grundler@chromium.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 2/3] NET: fix phy init for Asix AX88178 USB (GigE)
From: David Miller @ 2011-10-04 20:26 UTC (permalink / raw)
To: grundler; +Cc: netdev, linux-kernel, allan, freddy
In-Reply-To: <1317758118-24052-2-git-send-email-grundler@chromium.org>
From: Grant Grundler <grundler@chromium.org>
Date: Tue, 4 Oct 2011 12:55:17 -0700
> Asix provided this patch and I've confirmed "Plugable USB2-E1000" and
> "Shenzhen Winstars NWU220G" USB dongles can get a link and TX/RX data.
>
> Signed-off-by: "Freddy Xin" <freddy@asix.com.tw>
> Signed-off-by: Grant Grundler <grundler@chromium.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 1/3] NET: fix phy init for AX88772 USB ethernet
From: David Miller @ 2011-10-04 20:25 UTC (permalink / raw)
To: grundler; +Cc: netdev, linux-kernel, allan, freddy
In-Reply-To: <1317758118-24052-1-git-send-email-grundler@chromium.org>
From: Grant Grundler <grundler@chromium.org>
Date: Tue, 4 Oct 2011 12:55:16 -0700
> Fix phy initialization for AX88772 (USB 2.0 100BT). Failure was
> occasionally DHCP wouldn't work after reboot or suspend/resume cycle.
>
> Remove MONITOR_MODE. In this mode, Received packets are not buffered when
> the remote wakeup is enabled.
>
> Signed-off-by: "Freddy Xin" <freddy@asix.com.tw>
> Signed-off-by: Grant Grundler <grundler@chromium.org>
> Acked-by: Olof Johansson <olofj@chromium.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: big picture UDP/IP performance question re 2.6.18 -> 2.6.32
From: Serge Belyshev @ 2011-10-04 20:12 UTC (permalink / raw)
To: Christoph Lameter
Cc: Joe Perches, starlight, Eric Dumazet, linux-kernel, netdev,
Willy Tarreau, Peter Zijlstra, Stephen Hemminger
In-Reply-To: <alpine.DEB.2.00.1110041501130.16183@router.home>
BFS FAQ writes:
> NUMA aware?
>
> It is NOT NUMA aware in the sense that it does any fancy shit on NUMA, but
> it will work on NUMA hardware just fine. Only the really big NUMA hardware
> is likely to suffer in performance, and this is theoretically only, since
> no one has that sort of hardware to prove it to me, but it seems almost
> certain.
This part of FAQ entry was written before 0.300, (i.e. more than 2 years ago)
> v0.300 onwards have NUMA enhancements.
And this was added after. =)
As of now, BFS scheduling is NUMA-aware. (see resched_best_mask())
^ permalink raw reply
* Re: big picture UDP/IP performance question re 2.6.18 -> 2.6.32
From: Christoph Lameter @ 2011-10-04 20:03 UTC (permalink / raw)
To: Serge Belyshev
Cc: Joe Perches, starlight, Eric Dumazet, linux-kernel, netdev,
Willy Tarreau, Peter Zijlstra, Stephen Hemminger
In-Reply-To: <874nzo35bu.fsf@depni.sinp.msu.ru>
On Tue, 4 Oct 2011, Serge Belyshev wrote:
>
> Christoph Lameter <cl@gentwo.org> writes:
> > No. BFS has no NUMA support.
>
> BFS has NUMA support since v0.300, 1st oct. 2009.
>From the BFS FAQ (lack of NUMA support causes additional memory latencies,
the high end business server these days have NUMA):
NUMA aware?
It is NOT NUMA aware in the sense that it does any fancy shit on NUMA, but
it will work on NUMA hardware just fine. Only the really big NUMA hardware
is likely to suffer in performance, and this is theoretically only, since
no one has that sort of hardware to prove it to me, but it seems almost
certain. v0.300 onwards have NUMA enhancements.
^ permalink raw reply
* [PATCH 3/3] NET: white space/coding style cleanup of asix driver
From: Grant Grundler @ 2011-10-04 19:55 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-kernel, allan, Grant Grundler
In-Reply-To: <1317758118-24052-1-git-send-email-grundler@chromium.org>
check patch was complaining...mostly replaced:
if ((ret = asix_foo(xx)) < 0) ...
with
ret = asix_foo(xx);
if (ret < 0) ...
Signed-off-by: Grant Grundler <grundler@chromium.org>
---
drivers/net/usb/asix.c | 153 ++++++++++++++++++++++++++----------------------
1 files changed, 82 insertions(+), 71 deletions(-)
diff --git a/drivers/net/usb/asix.c b/drivers/net/usb/asix.c
index 23ee22d..66e5ccb 100644
--- a/drivers/net/usb/asix.c
+++ b/drivers/net/usb/asix.c
@@ -37,7 +37,7 @@
#include <linux/slab.h>
#define DRIVER_VERSION "26-Sep-2011"
-static const char driver_name [] = "asix";
+#define DRIVER_NAME "asix"
/* ASIX AX8817X based USB 2.0 Ethernet Devices */
@@ -115,28 +115,27 @@ static const char driver_name [] = "asix";
#define AX88178_MEDIUM_DEFAULT \
(AX_MEDIUM_PS | AX_MEDIUM_FD | AX_MEDIUM_AC | \
AX_MEDIUM_RFC | AX_MEDIUM_TFC | AX_MEDIUM_JFE | \
- AX_MEDIUM_RE )
+ AX_MEDIUM_RE)
#define AX88772_MEDIUM_DEFAULT \
(AX_MEDIUM_FD | AX_MEDIUM_RFC | \
AX_MEDIUM_TFC | AX_MEDIUM_PS | \
- AX_MEDIUM_AC | AX_MEDIUM_RE )
+ AX_MEDIUM_AC | AX_MEDIUM_RE)
/* AX88772 & AX88178 RX_CTL values */
-#define AX_RX_CTL_SO 0x0080
-#define AX_RX_CTL_AP 0x0020
-#define AX_RX_CTL_AM 0x0010
-#define AX_RX_CTL_AB 0x0008
-#define AX_RX_CTL_SEP 0x0004
-#define AX_RX_CTL_AMALL 0x0002
-#define AX_RX_CTL_PRO 0x0001
-#define AX_RX_CTL_MFB_2048 0x0000
-#define AX_RX_CTL_MFB_4096 0x0100
-#define AX_RX_CTL_MFB_8192 0x0200
-#define AX_RX_CTL_MFB_16384 0x0300
-
-#define AX_DEFAULT_RX_CTL \
- (AX_RX_CTL_SO | AX_RX_CTL_AB )
+#define AX_RX_CTL_SO 0x0080
+#define AX_RX_CTL_AP 0x0020
+#define AX_RX_CTL_AM 0x0010
+#define AX_RX_CTL_AB 0x0008
+#define AX_RX_CTL_SEP 0x0004
+#define AX_RX_CTL_AMALL 0x0002
+#define AX_RX_CTL_PRO 0x0001
+#define AX_RX_CTL_MFB_2048 0x0000
+#define AX_RX_CTL_MFB_4096 0x0100
+#define AX_RX_CTL_MFB_8192 0x0200
+#define AX_RX_CTL_MFB_16384 0x0300
+
+#define AX_DEFAULT_RX_CTL (AX_RX_CTL_SO | AX_RX_CTL_AB)
/* GPIO 0 .. 2 toggles */
#define AX_GPIO_GPO0EN 0x01 /* GPIO0 Output enable */
@@ -270,12 +269,15 @@ asix_write_cmd_async(struct usbnet *dev, u8 cmd, u16 value, u16 index,
netdev_dbg(dev->net, "asix_write_cmd_async() cmd=0x%02x value=0x%04x index=0x%04x size=%d\n",
cmd, value, index, size);
- if ((urb = usb_alloc_urb(0, GFP_ATOMIC)) == NULL) {
+
+ urb = usb_alloc_urb(0, GFP_ATOMIC);
+ if (!urb) {
netdev_err(dev->net, "Error allocating URB in write_cmd_async!\n");
return;
}
- if ((req = kmalloc(sizeof(struct usb_ctrlrequest), GFP_ATOMIC)) == NULL) {
+ req = kmalloc(sizeof(struct usb_ctrlrequest), GFP_ATOMIC);
+ if (!req) {
netdev_err(dev->net, "Failed to allocate memory for control request\n");
usb_free_urb(urb);
return;
@@ -292,7 +294,8 @@ asix_write_cmd_async(struct usbnet *dev, u8 cmd, u16 value, u16 index,
(void *)req, data, size,
asix_async_cmd_callback, req);
- if((status = usb_submit_urb(urb, GFP_ATOMIC)) < 0) {
+ status = usb_submit_urb(urb, GFP_ATOMIC);
+ if (status < 0) {
netdev_err(dev->net, "Error submitting the control message: status=%d\n",
status);
kfree(req);
@@ -534,11 +537,11 @@ static u16 asix_read_medium_status(struct usbnet *dev)
if (ret < 0) {
netdev_err(dev->net, "Error reading Medium Status register: %02x\n",
ret);
- goto out;
+ return ret; /* TODO: callers not checking for error ret */
}
- ret = le16_to_cpu(v);
-out:
- return ret;
+
+ return le16_to_cpu(v);
+
}
static int asix_write_medium_mode(struct usbnet *dev, u16 mode)
@@ -739,7 +742,7 @@ static void asix_get_drvinfo (struct net_device *net,
/* Inherit standard device info */
usbnet_get_drvinfo(net, info);
- strncpy (info->driver, driver_name, sizeof info->driver);
+ strncpy (info->driver, DRIVER_NAME, sizeof info->driver);
strncpy (info->version, DRIVER_VERSION, sizeof info->version);
info->eedump_len = data->eeprom_len;
}
@@ -884,19 +887,20 @@ static int ax88172_bind(struct usbnet *dev, struct usb_interface *intf)
/* Toggle the GPIOs in a manufacturer/model specific way */
for (i = 2; i >= 0; i--) {
- if ((ret = asix_write_cmd(dev, AX_CMD_WRITE_GPIOS,
- (gpio_bits >> (i * 8)) & 0xff, 0, 0,
- NULL)) < 0)
+ ret = asix_write_cmd(dev, AX_CMD_WRITE_GPIOS,
+ (gpio_bits >> (i * 8)) & 0xff, 0, 0, NULL);
+ if (ret < 0)
goto out;
msleep(5);
}
- if ((ret = asix_write_rx_ctl(dev, 0x80)) < 0)
+ ret = asix_write_rx_ctl(dev, 0x80);
+ if (ret < 0)
goto out;
/* Get the MAC address */
- if ((ret = asix_read_cmd(dev, AX88172_CMD_READ_NODE_ID,
- 0, 0, ETH_ALEN, buf)) < 0) {
+ ret = asix_read_cmd(dev, AX88172_CMD_READ_NODE_ID, 0, 0, ETH_ALEN, buf);
+ if (ret < 0) {
dbg("read AX_CMD_READ_NODE_ID failed: %d", ret);
goto out;
}
@@ -967,51 +971,59 @@ static int ax88772_reset(struct usbnet *dev)
int ret, embd_phy;
u16 rx_ctl;
- if ((ret = asix_write_gpio(dev,
- AX_GPIO_RSE | AX_GPIO_GPO_2 | AX_GPIO_GPO2EN, 5)) < 0)
+ ret = asix_write_gpio(dev,
+ AX_GPIO_RSE | AX_GPIO_GPO_2 | AX_GPIO_GPO2EN, 5);
+ if (ret < 0)
goto out;
embd_phy = ((asix_get_phy_addr(dev) & 0x1f) == 0x10 ? 1 : 0);
- if ((ret = asix_write_cmd(dev, AX_CMD_SW_PHY_SELECT,
- embd_phy, 0, 0, NULL)) < 0) {
+ ret = asix_write_cmd(dev, AX_CMD_SW_PHY_SELECT, embd_phy, 0, 0, NULL);
+ if (ret < 0) {
dbg("Select PHY #1 failed: %d", ret);
goto out;
}
- if ((ret = asix_sw_reset(dev, AX_SWRESET_IPPD | AX_SWRESET_PRL)) < 0)
+ ret = asix_sw_reset(dev, AX_SWRESET_IPPD | AX_SWRESET_PRL);
+ if (ret < 0)
goto out;
msleep(150);
- if ((ret = asix_sw_reset(dev, AX_SWRESET_CLEAR)) < 0)
+
+ ret = asix_sw_reset(dev, AX_SWRESET_CLEAR);
+ if (ret < 0)
goto out;
msleep(150);
if (embd_phy) {
- if ((ret = asix_sw_reset(dev, AX_SWRESET_IPRL)) < 0)
+ ret = asix_sw_reset(dev, AX_SWRESET_IPRL);
+ if (ret < 0)
goto out;
- }
- else {
- if ((ret = asix_sw_reset(dev, AX_SWRESET_PRTE)) < 0)
+ } else {
+ ret = asix_sw_reset(dev, AX_SWRESET_PRTE);
+ if (ret < 0)
goto out;
}
msleep(150);
rx_ctl = asix_read_rx_ctl(dev);
dbg("RX_CTL is 0x%04x after software reset", rx_ctl);
- if ((ret = asix_write_rx_ctl(dev, 0x0000)) < 0)
+ ret = asix_write_rx_ctl(dev, 0x0000);
+ if (ret < 0)
goto out;
rx_ctl = asix_read_rx_ctl(dev);
dbg("RX_CTL is 0x%04x setting to 0x0000", rx_ctl);
- if ((ret = asix_sw_reset(dev, AX_SWRESET_PRL)) < 0)
+ ret = asix_sw_reset(dev, AX_SWRESET_PRL);
+ if (ret < 0)
goto out;
msleep(150);
- if ((ret = asix_sw_reset(dev, AX_SWRESET_IPRL | AX_SWRESET_PRL)) < 0)
+ ret = asix_sw_reset(dev, AX_SWRESET_IPRL | AX_SWRESET_PRL);
+ if (ret < 0)
goto out;
msleep(150);
@@ -1021,18 +1033,21 @@ static int ax88772_reset(struct usbnet *dev)
ADVERTISE_ALL | ADVERTISE_CSMA);
mii_nway_restart(&dev->mii);
- if ((ret = asix_write_medium_mode(dev, AX88772_MEDIUM_DEFAULT)) < 0)
+ ret = asix_write_medium_mode(dev, AX88772_MEDIUM_DEFAULT);
+ if (ret < 0)
goto out;
- if ((ret = asix_write_cmd(dev, AX_CMD_WRITE_IPG0,
+ ret = asix_write_cmd(dev, AX_CMD_WRITE_IPG0,
AX88772_IPG0_DEFAULT | AX88772_IPG1_DEFAULT,
- AX88772_IPG2_DEFAULT, 0, NULL)) < 0) {
+ AX88772_IPG2_DEFAULT, 0, NULL);
+ if (ret < 0) {
dbg("Write IPG,IPG1,IPG2 failed: %d", ret);
goto out;
}
/* Set RX_CTL to default values with 2k buffer, and enable cactus */
- if ((ret = asix_write_rx_ctl(dev, AX_DEFAULT_RX_CTL)) < 0)
+ ret = asix_write_rx_ctl(dev, AX_DEFAULT_RX_CTL);
+ if (ret < 0)
goto out;
rx_ctl = asix_read_rx_ctl(dev);
@@ -1072,10 +1087,10 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
usbnet_get_endpoints(dev,intf);
/* Get the MAC address */
- if ((ret = asix_read_cmd(dev, AX_CMD_READ_NODE_ID,
- 0, 0, ETH_ALEN, buf)) < 0) {
+ ret = asix_read_cmd(dev, AX_CMD_READ_NODE_ID, 0, 0, ETH_ALEN, buf);
+ if (ret < 0) {
dbg("Failed to read MAC address: %d", ret);
- goto out;
+ return ret;
}
memcpy(dev->net->dev_addr, buf, ETH_ALEN);
@@ -1093,8 +1108,9 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
dev->net->netdev_ops = &ax88772_netdev_ops;
dev->net->ethtool_ops = &ax88772_ethtool_ops;
- if ((ret = ax88772_reset(dev)) < 0)
- goto out;
+ ret = ax88772_reset(dev);
+ if (ret < 0)
+ return ret;
/* Asix framing packs multiple eth frames into a 2K usb bulk transfer */
if (dev->driver_info->flags & FLAG_FRAMING_AX) {
@@ -1102,9 +1118,8 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
jumbo eth frames */
dev->rx_urb_size = 2048;
}
+
return 0;
-out:
- return ret;
}
static struct ethtool_ops ax88178_ethtool_ops = {
@@ -1262,17 +1277,15 @@ static int ax88178_reset(struct usbnet *dev)
mii_nway_restart(&dev->mii);
- if ((ret = asix_write_medium_mode(dev, AX88178_MEDIUM_DEFAULT)) < 0)
- goto out;
+ ret = asix_write_medium_mode(dev, AX88178_MEDIUM_DEFAULT);
+ if (ret < 0)
+ return ret;
- if ((ret = asix_write_rx_ctl(dev, AX_DEFAULT_RX_CTL)) < 0)
- goto out;
+ ret = asix_write_rx_ctl(dev, AX_DEFAULT_RX_CTL);
+ if (ret < 0)
+ return ret;
return 0;
-
-out:
- return ret;
-
}
static int ax88178_link_reset(struct usbnet *dev)
@@ -1390,10 +1403,10 @@ static int ax88178_bind(struct usbnet *dev, struct usb_interface *intf)
usbnet_get_endpoints(dev,intf);
/* Get the MAC address */
- if ((ret = asix_read_cmd(dev, AX_CMD_READ_NODE_ID,
- 0, 0, ETH_ALEN, buf)) < 0) {
+ ret = asix_read_cmd(dev, AX_CMD_READ_NODE_ID, 0, 0, ETH_ALEN, buf);
+ if (ret < 0) {
dbg("Failed to read MAC address: %d", ret);
- goto out;
+ return ret;
}
memcpy(dev->net->dev_addr, buf, ETH_ALEN);
@@ -1414,7 +1427,7 @@ static int ax88178_bind(struct usbnet *dev, struct usb_interface *intf)
ret = ax88178_reset(dev);
if (ret < 0)
- goto out;
+ return ret;
/* Asix framing packs multiple eth frames into a 2K usb bulk transfer */
if (dev->driver_info->flags & FLAG_FRAMING_AX) {
@@ -1422,10 +1435,8 @@ static int ax88178_bind(struct usbnet *dev, struct usb_interface *intf)
jumbo eth frames */
dev->rx_urb_size = 2048;
}
- return 0;
-out:
- return ret;
+ return 0;
}
static const struct driver_info ax8817x_info = {
@@ -1609,7 +1620,7 @@ static const struct usb_device_id products [] = {
MODULE_DEVICE_TABLE(usb, products);
static struct usb_driver asix_driver = {
- .name = "asix",
+ .name = DRIVER_NAME,
.id_table = products,
.probe = usbnet_probe,
.suspend = usbnet_suspend,
--
1.7.2.3
^ permalink raw reply related
* [PATCH 2/3] NET: fix phy init for Asix AX88178 USB (GigE)
From: Grant Grundler @ 2011-10-04 19:55 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-kernel, allan, Grant Grundler, Freddy Xin
In-Reply-To: <1317758118-24052-1-git-send-email-grundler@chromium.org>
Asix provided this patch and I've confirmed "Plugable USB2-E1000" and
"Shenzhen Winstars NWU220G" USB dongles can get a link and TX/RX data.
Signed-off-by: "Freddy Xin" <freddy@asix.com.tw>
Signed-off-by: Grant Grundler <grundler@chromium.org>
---
drivers/net/usb/asix.c | 163 ++++++++++++++++++++++++++++++------------------
1 files changed, 101 insertions(+), 62 deletions(-)
diff --git a/drivers/net/usb/asix.c b/drivers/net/usb/asix.c
index a035aa9..23ee22d 100644
--- a/drivers/net/usb/asix.c
+++ b/drivers/net/usb/asix.c
@@ -164,6 +164,8 @@ static const char driver_name [] = "asix";
#define MARVELL_CTRL_TXDELAY 0x0002
#define MARVELL_CTRL_RXDELAY 0x0080
+#define PHY_MODE_RTL8211CL 0x0004
+
/* This structure cannot exceed sizeof(unsigned long [5]) AKA 20 bytes */
struct asix_data {
u8 multi_filter[AX_MCAST_FILTER_SIZE];
@@ -1151,6 +1153,27 @@ static int marvell_phy_init(struct usbnet *dev)
return 0;
}
+static int rtl8211cl_phy_init(struct usbnet *dev)
+{
+ struct asix_data *data = (struct asix_data *)&dev->data;
+
+ netdev_dbg(dev->net, "rtl8211cl_phy_init()\n");
+
+ asix_mdio_write (dev->net, dev->mii.phy_id, 0x1f, 0x0005);
+ asix_mdio_write (dev->net, dev->mii.phy_id, 0x0c, 0);
+ asix_mdio_write (dev->net, dev->mii.phy_id, 0x01,
+ asix_mdio_read (dev->net, dev->mii.phy_id, 0x01) | 0x0080);
+ asix_mdio_write (dev->net, dev->mii.phy_id, 0x1f, 0);
+
+ if (data->ledmode == 12) {
+ asix_mdio_write (dev->net, dev->mii.phy_id, 0x1f, 0x0002);
+ asix_mdio_write (dev->net, dev->mii.phy_id, 0x1a, 0x00cb);
+ asix_mdio_write (dev->net, dev->mii.phy_id, 0x1f, 0);
+ }
+
+ return 0;
+}
+
static int marvell_led_status(struct usbnet *dev, u16 speed)
{
u16 reg = asix_mdio_read(dev->net, dev->mii.phy_id, MARVELL_LED_MANUAL);
@@ -1177,6 +1200,81 @@ static int marvell_led_status(struct usbnet *dev, u16 speed)
return 0;
}
+static int ax88178_reset(struct usbnet *dev)
+{
+ struct asix_data *data = (struct asix_data *)&dev->data;
+ int ret;
+ __le16 eeprom;
+ u8 status;
+ int gpio0 = 0;
+
+ asix_read_cmd(dev, AX_CMD_READ_GPIOS, 0, 0, 1, &status);
+ dbg("GPIO Status: 0x%04x", status);
+
+ asix_write_cmd(dev, AX_CMD_WRITE_ENABLE, 0, 0, 0, NULL);
+ asix_read_cmd(dev, AX_CMD_READ_EEPROM, 0x0017, 0, 2, &eeprom);
+ asix_write_cmd(dev, AX_CMD_WRITE_DISABLE, 0, 0, 0, NULL);
+
+ dbg("EEPROM index 0x17 is 0x%04x", eeprom);
+
+ if (eeprom == cpu_to_le16(0xffff)) {
+ data->phymode = PHY_MODE_MARVELL;
+ data->ledmode = 0;
+ gpio0 = 1;
+ } else {
+ data->phymode = le16_to_cpu(eeprom) & 7;
+ data->ledmode = le16_to_cpu(eeprom) >> 8;
+ gpio0 = (le16_to_cpu(eeprom) & 0x80) ? 0 : 1;
+ }
+ dbg("GPIO0: %d, PhyMode: %d", gpio0, data->phymode);
+
+ asix_write_gpio(dev, AX_GPIO_RSE | AX_GPIO_GPO_1 | AX_GPIO_GPO1EN, 40);
+ if ((le16_to_cpu(eeprom) >> 8) != 1) {
+ asix_write_gpio(dev, 0x003c, 30);
+ asix_write_gpio(dev, 0x001c, 300);
+ asix_write_gpio(dev, 0x003c, 30);
+ } else {
+ dbg("gpio phymode == 1 path");
+ asix_write_gpio(dev, AX_GPIO_GPO1EN, 30);
+ asix_write_gpio(dev, AX_GPIO_GPO1EN | AX_GPIO_GPO_1, 30);
+ }
+
+ asix_sw_reset(dev, 0);
+ msleep(150);
+
+ asix_sw_reset(dev, AX_SWRESET_PRL | AX_SWRESET_IPPD);
+ msleep(150);
+
+ asix_write_rx_ctl(dev, 0);
+
+ if (data->phymode == PHY_MODE_MARVELL) {
+ marvell_phy_init(dev);
+ msleep(60);
+ } else if (data->phymode == PHY_MODE_RTL8211CL)
+ rtl8211cl_phy_init(dev);
+
+ asix_mdio_write(dev->net, dev->mii.phy_id, MII_BMCR,
+ BMCR_RESET | BMCR_ANENABLE);
+ asix_mdio_write(dev->net, dev->mii.phy_id, MII_ADVERTISE,
+ ADVERTISE_ALL | ADVERTISE_CSMA | ADVERTISE_PAUSE_CAP);
+ asix_mdio_write(dev->net, dev->mii.phy_id, MII_CTRL1000,
+ ADVERTISE_1000FULL);
+
+ mii_nway_restart(&dev->mii);
+
+ if ((ret = asix_write_medium_mode(dev, AX88178_MEDIUM_DEFAULT)) < 0)
+ goto out;
+
+ if ((ret = asix_write_rx_ctl(dev, AX_DEFAULT_RX_CTL)) < 0)
+ goto out;
+
+ return 0;
+
+out:
+ return ret;
+
+}
+
static int ax88178_link_reset(struct usbnet *dev)
{
u16 mode;
@@ -1285,55 +1383,12 @@ static const struct net_device_ops ax88178_netdev_ops = {
static int ax88178_bind(struct usbnet *dev, struct usb_interface *intf)
{
- struct asix_data *data = (struct asix_data *)&dev->data;
int ret;
u8 buf[ETH_ALEN];
- __le16 eeprom;
- u8 status;
- int gpio0 = 0;
u32 phyid;
usbnet_get_endpoints(dev,intf);
- asix_read_cmd(dev, AX_CMD_READ_GPIOS, 0, 0, 1, &status);
- dbg("GPIO Status: 0x%04x", status);
-
- asix_write_cmd(dev, AX_CMD_WRITE_ENABLE, 0, 0, 0, NULL);
- asix_read_cmd(dev, AX_CMD_READ_EEPROM, 0x0017, 0, 2, &eeprom);
- asix_write_cmd(dev, AX_CMD_WRITE_DISABLE, 0, 0, 0, NULL);
-
- dbg("EEPROM index 0x17 is 0x%04x", eeprom);
-
- if (eeprom == cpu_to_le16(0xffff)) {
- data->phymode = PHY_MODE_MARVELL;
- data->ledmode = 0;
- gpio0 = 1;
- } else {
- data->phymode = le16_to_cpu(eeprom) & 7;
- data->ledmode = le16_to_cpu(eeprom) >> 8;
- gpio0 = (le16_to_cpu(eeprom) & 0x80) ? 0 : 1;
- }
- dbg("GPIO0: %d, PhyMode: %d", gpio0, data->phymode);
-
- asix_write_gpio(dev, AX_GPIO_RSE | AX_GPIO_GPO_1 | AX_GPIO_GPO1EN, 40);
- if ((le16_to_cpu(eeprom) >> 8) != 1) {
- asix_write_gpio(dev, 0x003c, 30);
- asix_write_gpio(dev, 0x001c, 300);
- asix_write_gpio(dev, 0x003c, 30);
- } else {
- dbg("gpio phymode == 1 path");
- asix_write_gpio(dev, AX_GPIO_GPO1EN, 30);
- asix_write_gpio(dev, AX_GPIO_GPO1EN | AX_GPIO_GPO_1, 30);
- }
-
- asix_sw_reset(dev, 0);
- msleep(150);
-
- asix_sw_reset(dev, AX_SWRESET_PRL | AX_SWRESET_IPPD);
- msleep(150);
-
- asix_write_rx_ctl(dev, 0);
-
/* Get the MAC address */
if ((ret = asix_read_cmd(dev, AX_CMD_READ_NODE_ID,
0, 0, ETH_ALEN, buf)) < 0) {
@@ -1357,24 +1412,8 @@ static int ax88178_bind(struct usbnet *dev, struct usb_interface *intf)
phyid = asix_get_phyid(dev);
dbg("PHYID=0x%08x", phyid);
- if (data->phymode == PHY_MODE_MARVELL) {
- marvell_phy_init(dev);
- msleep(60);
- }
-
- asix_mdio_write(dev->net, dev->mii.phy_id, MII_BMCR,
- BMCR_RESET | BMCR_ANENABLE);
- asix_mdio_write(dev->net, dev->mii.phy_id, MII_ADVERTISE,
- ADVERTISE_ALL | ADVERTISE_CSMA | ADVERTISE_PAUSE_CAP);
- asix_mdio_write(dev->net, dev->mii.phy_id, MII_CTRL1000,
- ADVERTISE_1000FULL);
-
- mii_nway_restart(&dev->mii);
-
- if ((ret = asix_write_medium_mode(dev, AX88178_MEDIUM_DEFAULT)) < 0)
- goto out;
-
- if ((ret = asix_write_rx_ctl(dev, AX_DEFAULT_RX_CTL)) < 0)
+ ret = ax88178_reset(dev);
+ if (ret < 0)
goto out;
/* Asix framing packs multiple eth frames into a 2K usb bulk transfer */
@@ -1445,7 +1484,7 @@ static const struct driver_info ax88178_info = {
.bind = ax88178_bind,
.status = asix_status,
.link_reset = ax88178_link_reset,
- .reset = ax88178_link_reset,
+ .reset = ax88178_reset,
.flags = FLAG_ETHER | FLAG_FRAMING_AX | FLAG_LINK_INTR,
.rx_fixup = asix_rx_fixup,
.tx_fixup = asix_tx_fixup,
--
1.7.2.3
^ permalink raw reply related
* [PATCH 1/3] NET: fix phy init for AX88772 USB ethernet
From: Grant Grundler @ 2011-10-04 19:55 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-kernel, allan, Grant Grundler, Freddy Xin
Fix phy initialization for AX88772 (USB 2.0 100BT). Failure was
occasionally DHCP wouldn't work after reboot or suspend/resume cycle.
Remove MONITOR_MODE. In this mode, Received packets are not buffered when
the remote wakeup is enabled.
Signed-off-by: "Freddy Xin" <freddy@asix.com.tw>
Signed-off-by: Grant Grundler <grundler@chromium.org>
Acked-by: Olof Johansson <olofj@chromium.org>
---
drivers/net/usb/asix.c | 116 +++++++++++++++++++++++++----------------------
1 files changed, 62 insertions(+), 54 deletions(-)
diff --git a/drivers/net/usb/asix.c b/drivers/net/usb/asix.c
index 6998aa6..a035aa9 100644
--- a/drivers/net/usb/asix.c
+++ b/drivers/net/usb/asix.c
@@ -36,7 +36,7 @@
#include <linux/usb/usbnet.h>
#include <linux/slab.h>
-#define DRIVER_VERSION "14-Jun-2006"
+#define DRIVER_VERSION "26-Sep-2011"
static const char driver_name [] = "asix";
/* ASIX AX8817X based USB 2.0 Ethernet Devices */
@@ -677,12 +677,6 @@ asix_get_wol(struct net_device *net, struct ethtool_wolinfo *wolinfo)
}
wolinfo->supported = WAKE_PHY | WAKE_MAGIC;
wolinfo->wolopts = 0;
- if (opt & AX_MONITOR_MODE) {
- if (opt & AX_MONITOR_LINK)
- wolinfo->wolopts |= WAKE_PHY;
- if (opt & AX_MONITOR_MAGIC)
- wolinfo->wolopts |= WAKE_MAGIC;
- }
}
static int
@@ -695,8 +689,6 @@ asix_set_wol(struct net_device *net, struct ethtool_wolinfo *wolinfo)
opt |= AX_MONITOR_LINK;
if (wolinfo->wolopts & WAKE_MAGIC)
opt |= AX_MONITOR_MAGIC;
- if (opt != 0)
- opt |= AX_MONITOR_MODE;
if (asix_write_cmd(dev, AX_CMD_WRITE_MONITOR_MODE,
opt, 0, 0, NULL) < 0)
@@ -967,36 +959,18 @@ static int ax88772_link_reset(struct usbnet *dev)
return 0;
}
-static const struct net_device_ops ax88772_netdev_ops = {
- .ndo_open = usbnet_open,
- .ndo_stop = usbnet_stop,
- .ndo_start_xmit = usbnet_start_xmit,
- .ndo_tx_timeout = usbnet_tx_timeout,
- .ndo_change_mtu = usbnet_change_mtu,
- .ndo_set_mac_address = asix_set_mac_address,
- .ndo_validate_addr = eth_validate_addr,
- .ndo_do_ioctl = asix_ioctl,
- .ndo_set_multicast_list = asix_set_multicast,
-};
-
-static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
+static int ax88772_reset(struct usbnet *dev)
{
+
int ret, embd_phy;
u16 rx_ctl;
- struct asix_data *data = (struct asix_data *)&dev->data;
- u8 buf[ETH_ALEN];
- u32 phyid;
-
- data->eeprom_len = AX88772_EEPROM_LEN;
-
- usbnet_get_endpoints(dev,intf);
if ((ret = asix_write_gpio(dev,
AX_GPIO_RSE | AX_GPIO_GPO_2 | AX_GPIO_GPO2EN, 5)) < 0)
goto out;
- /* 0x10 is the phy id of the embedded 10/100 ethernet phy */
embd_phy = ((asix_get_phy_addr(dev) & 0x1f) == 0x10 ? 1 : 0);
+
if ((ret = asix_write_cmd(dev, AX_CMD_SW_PHY_SELECT,
embd_phy, 0, 0, NULL)) < 0) {
dbg("Select PHY #1 failed: %d", ret);
@@ -1011,6 +985,7 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
goto out;
msleep(150);
+
if (embd_phy) {
if ((ret = asix_sw_reset(dev, AX_SWRESET_IPRL)) < 0)
goto out;
@@ -1029,25 +1004,6 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
rx_ctl = asix_read_rx_ctl(dev);
dbg("RX_CTL is 0x%04x setting to 0x0000", rx_ctl);
- /* Get the MAC address */
- if ((ret = asix_read_cmd(dev, AX_CMD_READ_NODE_ID,
- 0, 0, ETH_ALEN, buf)) < 0) {
- dbg("Failed to read MAC address: %d", ret);
- goto out;
- }
- memcpy(dev->net->dev_addr, buf, ETH_ALEN);
-
- /* Initialize MII structure */
- dev->mii.dev = dev->net;
- dev->mii.mdio_read = asix_mdio_read;
- dev->mii.mdio_write = asix_mdio_write;
- dev->mii.phy_id_mask = 0x1f;
- dev->mii.reg_num_mask = 0x1f;
- dev->mii.phy_id = asix_get_phy_addr(dev);
-
- phyid = asix_get_phyid(dev);
- dbg("PHYID=0x%08x", phyid);
-
if ((ret = asix_sw_reset(dev, AX_SWRESET_PRL)) < 0)
goto out;
@@ -1058,9 +1014,6 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
msleep(150);
- dev->net->netdev_ops = &ax88772_netdev_ops;
- dev->net->ethtool_ops = &ax88772_ethtool_ops;
-
asix_mdio_write(dev->net, dev->mii.phy_id, MII_BMCR, BMCR_RESET);
asix_mdio_write(dev->net, dev->mii.phy_id, MII_ADVERTISE,
ADVERTISE_ALL | ADVERTISE_CSMA);
@@ -1086,6 +1039,61 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
rx_ctl = asix_read_medium_status(dev);
dbg("Medium Status is 0x%04x after all initializations", rx_ctl);
+ return 0;
+
+out:
+ return ret;
+
+}
+
+static const struct net_device_ops ax88772_netdev_ops = {
+ .ndo_open = usbnet_open,
+ .ndo_stop = usbnet_stop,
+ .ndo_start_xmit = usbnet_start_xmit,
+ .ndo_tx_timeout = usbnet_tx_timeout,
+ .ndo_change_mtu = usbnet_change_mtu,
+ .ndo_set_mac_address = asix_set_mac_address,
+ .ndo_validate_addr = eth_validate_addr,
+ .ndo_do_ioctl = asix_ioctl,
+ .ndo_set_multicast_list = asix_set_multicast,
+};
+
+static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
+{
+ int ret;
+ struct asix_data *data = (struct asix_data *)&dev->data;
+ u8 buf[ETH_ALEN];
+ u32 phyid;
+
+ data->eeprom_len = AX88772_EEPROM_LEN;
+
+ usbnet_get_endpoints(dev,intf);
+
+ /* Get the MAC address */
+ if ((ret = asix_read_cmd(dev, AX_CMD_READ_NODE_ID,
+ 0, 0, ETH_ALEN, buf)) < 0) {
+ dbg("Failed to read MAC address: %d", ret);
+ goto out;
+ }
+ memcpy(dev->net->dev_addr, buf, ETH_ALEN);
+
+ /* Initialize MII structure */
+ dev->mii.dev = dev->net;
+ dev->mii.mdio_read = asix_mdio_read;
+ dev->mii.mdio_write = asix_mdio_write;
+ dev->mii.phy_id_mask = 0x1f;
+ dev->mii.reg_num_mask = 0x1f;
+ dev->mii.phy_id = asix_get_phy_addr(dev);
+
+ phyid = asix_get_phyid(dev);
+ dbg("PHYID=0x%08x", phyid);
+
+ dev->net->netdev_ops = &ax88772_netdev_ops;
+ dev->net->ethtool_ops = &ax88772_ethtool_ops;
+
+ if ((ret = ax88772_reset(dev)) < 0)
+ goto out;
+
/* Asix framing packs multiple eth frames into a 2K usb bulk transfer */
if (dev->driver_info->flags & FLAG_FRAMING_AX) {
/* hard_mtu is still the default - the device does not support
@@ -1093,7 +1101,6 @@ static int ax88772_bind(struct usbnet *dev, struct usb_interface *intf)
dev->rx_urb_size = 2048;
}
return 0;
-
out:
return ret;
}
@@ -1427,7 +1434,7 @@ static const struct driver_info ax88772_info = {
.bind = ax88772_bind,
.status = asix_status,
.link_reset = ax88772_link_reset,
- .reset = ax88772_link_reset,
+ .reset = ax88772_reset,
.flags = FLAG_ETHER | FLAG_FRAMING_AX | FLAG_LINK_INTR,
.rx_fixup = asix_rx_fixup,
.tx_fixup = asix_tx_fixup,
@@ -1585,6 +1592,7 @@ static void __exit asix_exit(void)
module_exit(asix_exit);
MODULE_AUTHOR("David Hollis");
+MODULE_VERSION(DRIVER_VERSION);
MODULE_DESCRIPTION("ASIX AX8817X based USB 2.0 Ethernet Devices");
MODULE_LICENSE("GPL");
--
1.7.2.3
^ permalink raw reply related
* RE: big picture UDP/IP performance question re 2.6.18 -> 2.6.32
From: Loke, Chetan @ 2011-10-04 19:54 UTC (permalink / raw)
To: starlight, Eric Dumazet
Cc: linux-kernel, netdev, Willy Tarreau, Peter Zijlstra
In-Reply-To: <6.2.5.6.2.20111002110545.03a82460@binnacle.cx>
> Sometimes I think it might be nice to run without
> an OS at all :-)
>
> We are looking at various kernel bypass approaches
> for receiving packets and it's in my mind that
> one of these may get us to running on newer
> kernels without giving up performance.
>
If permissible in your environment, try running PACKET:tpacket_v3.
Also, I always wondered about the following in the output of:
net/ptype
dada edsa_rcv+0x0/0x246
001b dsa_rcv+0x0/0x225
and so when I'm fine tuning apps in virtual environments I nuke the
above during kernel compilation.
Chetan Loke
^ permalink raw reply
* Re: big picture UDP/IP performance question re 2.6.18 -> 2.6.32
From: Serge Belyshev @ 2011-10-04 19:49 UTC (permalink / raw)
To: Christoph Lameter
Cc: Joe Perches, starlight, Eric Dumazet, linux-kernel, netdev,
Willy Tarreau, Peter Zijlstra, Stephen Hemminger
In-Reply-To: <alpine.DEB.2.00.1110041442020.15871@router.home>
Christoph Lameter <cl@gentwo.org> writes:
> No. BFS has no NUMA support.
BFS has NUMA support since v0.300, 1st oct. 2009.
^ permalink raw reply
* Re: big picture UDP/IP performance question re 2.6.18 -> 2.6.32
From: starlight @ 2011-10-04 19:45 UTC (permalink / raw)
To: Joe Perches, Christoph Lameter
Cc: Eric Dumazet, linux-kernel, netdev, Willy Tarreau, Peter Zijlstra,
Stephen Hemminger
In-Reply-To: <1317757118.3580.24.camel@Joe-Laptop>
At 12:38 PM 10/4/2011 -0700, Joe Perches wrote:
>On Tue, 2011-10-04 at 14:16 -0500, Christoph Lameter wrote:
>> On Mon, 3 Oct 2011, starlight@binnacle.cx wrote:
>> > I've come to the conclusion that Eric is right
>> > and the primary issue is an increase in the
>> > cost of scheduler context switches. Have
>> > been watching this number and it has held
>> > pretty close to 200k/sec under all scenarios
>> > and kernel versions, so it has to be
>> > a longer code-path, bigger cache pressure
>> > or both in the scheduler. Sadly this makes
>> > newer kernels a no-go for us.
>> We had similar experiences. Basically latency
>> constantly gets screwed up by the new fancy
>> features being added to the scheduler and network
>> subsystem (most notorious is the new "fair"
>> scheduler, 2.6.23 made a big
>> step down).
>
>Idly curious, have you compared bfs performance?
>http://ck.kolivas.org/patches/bfs/bfs-faq.txt
No but it certainly does look very interesting.
Looks like trying it out not much more work than
patching and building a kernel and running the
benchmark. Will take a look and report back
if I do. Little busy at present but should
have time in a week or so.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox