* order 7 allocations from xt_recent
From: Dave Jones @ 2013-01-03 16:43 UTC (permalink / raw)
To: netdev; +Cc: h.reindl, Fedora Kernel Team
We had a report from a user that shows this code trying
to do enormous allocations, which isn't going to work too well..
iptables: page allocation failure: order:7, mode:0xc0d0
Pid: 2822, comm: iptables Not tainted 3.6.10-2.fc17.x86_64 #1
Call Trace:
[<ffffffff8113130b>] warn_alloc_failed+0xeb/0x150
[<ffffffff81616576>] ? __alloc_pages_direct_compact+0x17e/0x190
[<ffffffff81135196>] __alloc_pages_nodemask+0x736/0x990
[<ffffffff811710e0>] alloc_pages_current+0xb0/0x120
[<ffffffff8113022a>] __get_free_pages+0x2a/0x80
[<ffffffff811786d9>] kmalloc_order_trace+0x39/0xb0
[<ffffffff8117ae3a>] __kmalloc+0x16a/0x1a0
[<ffffffff8118aa7c>] ? mem_cgroup_bad_page_check+0x1c/0x30
[<ffffffff81134563>] ? get_page_from_freelist+0x453/0x950
[<ffffffffa007696e>] recent_mt_check.isra.6+0x16e/0x2c0 [xt_recent]
[<ffffffffa0076b4b>] recent_mt_check_v0+0x6b/0xa0 [xt_recent]
[<ffffffff8153fdda>] xt_check_match+0xaa/0x1e0
[<ffffffff8153f3ab>] ? xt_find_match+0x11b/0x130
[<ffffffff8153f3ab>] ? xt_find_match+0x11b/0x130
[<ffffffff8159257c>] check_match+0x3c/0x50
[<ffffffff81593ccb>] translate_table+0x39b/0x5b0
[<ffffffff815956f3>] do_ipt_set_ctl+0x133/0x200
[<ffffffff8153e10b>] nf_setsockopt+0x6b/0x90
[<ffffffff8161f236>] ? _raw_spin_lock_bh+0x16/0x40
[<ffffffff8154e41f>] ip_setsockopt+0x8f/0xa0
[<ffffffff8156f49d>] raw_setsockopt+0x1d/0x30
[<ffffffff814fcf14>] sock_common_setsockopt+0x14/0x20
[<ffffffff814fc23c>] sys_setsockopt+0x7c/0xe0
[<ffffffff816270e9>] system_call_fastpath+0x16/0x1b
which looks like it's this..
t = kzalloc(sizeof(*t) + sizeof(t->iphash[0]) * ip_list_hash_size,
GFP_KERNEL);
Which is initialised thus..
ip_list_hash_size = 1 << fls(ip_list_tot);
And ip_list_tot is 10000 in this case. Hmm ?
Complete report and setup described in his bug report at https://bugzilla.redhat.com/show_bug.cgi?id=890715
Dave
^ permalink raw reply
* Re: [PATCH v2] tcp: split tcp_ecn sysctl knob to distinguish between IPv4 and IPv6
From: Stephen Hemminger @ 2013-01-03 16:22 UTC (permalink / raw)
To: Hannes Frederic Sowa; +Cc: David Miller, netdev
In-Reply-To: <20130103125812.GJ4464@order.stressinduktion.org>
On Thu, 3 Jan 2013 13:58:12 +0100
Hannes Frederic Sowa <hannes@stressinduktion.org> wrote:
> On Thu, Jan 03, 2013 at 04:53:50AM -0800, David Miller wrote:
> > From: Hannes Frederic Sowa <hannes@stressinduktion.org>
> > Date: Thu, 3 Jan 2013 13:28:44 +0100
> >
> > > On Thu, Jan 03, 2013 at 01:00:09AM +0100, Hannes Frederic Sowa wrote:
> > >> ECN could be more reliable when used with IPv6 (I don't have proofs). For
> > >> people who want to try ECN with IPv6 but still have problems connecting
> > >> to destinations because of broken IPv4 routers this switch allows one
> > >> to enable ECN just for IPv6.
> > >>
> > >> Perhaps ECN could be enabled by default in future.
> > >>
> > >> No code changes since initial submission, just added documentation.
> > >
> > > May I ask what lead to the rejection of the patch?
> >
> > I don't see any value in splitting up ipv4 and ipv6.
> >
> > I think we should turn the existing knob on unliaterally
> > by default. That would be a much better patch.
>
> Oh, I would be happy to do that. :)
Perhaps, ECN should be a property of the route.
There is RTAX_FEATURE_ECN but it appears to have been lost
as part of the removal of route cache.
^ permalink raw reply
* Re: [PATCH] bgmac: driver for GBit MAC core on BCMA bus
From: Rafał Miłecki @ 2013-01-03 15:57 UTC (permalink / raw)
To: Francois Romieu; +Cc: netdev, David S. Miller
In-Reply-To: <20121226231721.GA8243@electric-eye.fr.zoreil.com>
2012/12/27 Francois Romieu <romieu@fr.zoreil.com>:
> Rafał Miłecki <zajec5@gmail.com> :
> [...]
>> diff --git a/drivers/net/ethernet/broadcom/bgmac.c b/drivers/net/ethernet/broadcom/bgmac.c
> [...]
>> +static void bgmac_dma_tx_reset(struct bgmac *bgmac, struct bgmac_dma_ring *ring)
>> +{
>> + u32 val;
>> + int i;
>> +
>> + if (!ring->mmio_base)
>> + return;
>
> static int bgmac_probe(struct bcma_device *core)
> [...]
> bgmac_chip_reset(bgmac);
>
> err = bgmac_dma_alloc(bgmac);
>
> mmio_base is only set in bgmac_dma_alloc.
>
> -> remove the useless bgmac_chip_reset() in bgmac_probe() and the driver
> won't need to test for !ring->mmio_base.
bgmac_chip_reset also does different things, so I can't just remove
bgmac_chip_reset from bgmac_probe. I could add some parameter to
bgmac_chip_reset to let it know if DMA should be reset as well, but
I'm note sure about the advantage.
>> + if (skb->len > BGMAC_DESC_CTL1_LEN) {
>> + bgmac_err(bgmac, "Too long skb (%d)\n", skb->len);
>> + return NETDEV_TX_BUSY;
>> + }
>
> The driver will loop if you don't stop queuing. It won't help.
I hoped net layer will drop the packed after few failed xmit. OK, I'll
add netif_stop_queue here.
>> +
>> + if (ring->start <= ring->end)
>> + free_slots = ring->start - ring->end + BGMAC_TX_RING_SLOTS;
>> + else
>> + free_slots = ring->start - ring->end;
>> + if (free_slots <= 1) {
>> + bgmac_err(bgmac, "No free slots on ring 0x%X!\n", ring->mmio_base);
>> + netif_stop_queue(bgmac->net_dev);
>> + return NETDEV_TX_BUSY;
>> + }
>
> The driver should stop queueing after its processing if it detects that
> it can't handle more requests.
I think it's what the code does... netif_stop_queue and return
NETDEV_TX_BUSY. Anything wrong about it?
>> + slot = &ring->slots[ring->end];
>> + slot->skb = skb;
>> + slot->dma_addr = dma_map_single(dma_dev, skb->data, skb->len, DMA_TO_DEVICE);
>> + if (dma_mapping_error(dma_dev, slot->dma_addr)) {
>> + bgmac_err(bgmac, "Mapping error of skb on ring 0x%X\n", ring->mmio_base);
>> + return NETDEV_TX_BUSY;
>> + }
>
> Why don't you drop the packet and return TX_OK ?
Didn't know about that and there isn't any nice NAPI tutorial for developers :(
>> + dma_desc->addr_low = cpu_to_le32(lower_32_bits(slot->dma_addr));
>> + dma_desc->addr_high = cpu_to_le32(upper_32_bits(slot->dma_addr));
>> + dma_desc->ctl0 = cpu_to_le32(ctl0);
>> + dma_desc->ctl1 = cpu_to_le32(ctl1);
>> +
>> + wmb();
>
> Are we sure that the hardware will never read a descriptor where ctl0 is
> set and ctl1 is not (start_xmit, dma starts, competing start_xmit, oops) ?
Do you mean more that one ->ndo_start_xmit(...) can be executed at a
time? So I need mutex for it?
--
Rafał
^ permalink raw reply
* Re: [RFC PATCH net-next 3/3] vxlan: remove NETIF_F_NETNS_LOCAL from vxlan device features.
From: Stephen Hemminger @ 2013-01-03 15:45 UTC (permalink / raw)
To: Rami Rosen; +Cc: davem, netdev, ebiederm
In-Reply-To: <1357205121-4700-4-git-send-email-ramirose@gmail.com>
On Thu, 3 Jan 2013 11:25:21 +0200
Rami Rosen <ramirose@gmail.com> wrote:
> There is no need for NETIF_F_NETNS_LOCAL for vxlan device; this patch removes it.
>
> Signed-off-by: Rami Rosen <ramirose@gmail.com>
> ---
> drivers/net/vxlan.c | 1 -
> 1 file changed, 1 deletion(-)
>
> diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
> index 40f2cc1..385f743 100644
> --- a/drivers/net/vxlan.c
> +++ b/drivers/net/vxlan.c
> @@ -1185,7 +1185,6 @@ static void vxlan_setup(struct net_device *dev)
>
> dev->tx_queue_len = 0;
> dev->features |= NETIF_F_LLTX;
> - dev->features |= NETIF_F_NETNS_LOCAL;
> dev->features |= NETIF_F_SG | NETIF_F_HW_CSUM;
> dev->features |= NETIF_F_RXCSUM;
>
Vxlan works over UDP socket, and the UDP socket is part of the namespace
it is created in. Moving the vxlan device does not move that state.
^ permalink raw reply
* Re: [PATCH net-next v2 2/3] netpoll: add IPv6 support
From: Cong Wang @ 2013-01-03 15:08 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20130103.015453.730687035214069794.davem@davemloft.net>
On Thu, 2013-01-03 at 01:54 -0800, David Miller wrote:
> From: Cong Wang <amwang@redhat.com>
> Date: Thu, 3 Jan 2013 12:50:22 +0800
>
> > -static inline int udp6_csum_init(struct sk_buff *skb, struct udphdr *uh,
> > - int proto)
> > +int udp6_csum_init(struct sk_buff *skb, struct udphdr *uh, int proto)
> > {
> > int err;
> >
> > --
> > 1.7.7.6
> >
>
> With IPV6=m and NETPOLL=y using this symbol will not work.
>
Good point!
One solution is to introduce a new Kconfig option:
+config NETPOLL6
+ depends on (IPV6 || IPV6=n)
+ def_bool NETPOLL
+
and move these IPV6 code under #ifdef CONFIG_NETPOLL6.
^ permalink raw reply
* Re: [PATCH 4/4 v3] net/smsc911x: Provide common clock functionality
From: Linus Walleij @ 2013-01-03 14:28 UTC (permalink / raw)
To: Lee Jones
Cc: Russell King - ARM Linux, Steve Glendinning, Robert Marklund,
linus.walleij, arnd, netdev, linux-kernel, linux-arm-kernel
In-Reply-To: <20130103111450.GC14714@gmail.com>
On Thu, Jan 3, 2013 at 12:14 PM, Lee Jones <lee.jones@linaro.org> wrote:
> Some platforms provide clocks which require enabling before the
> SMSC911x chip will power on. This patch uses the new common clk
> framework to do just that. If no clock is provided, it will just
> be ignored and the driver will continue to assume that no clock
> is required for the chip to run successfully.
>
> Cc: Steve Glendinning <steve.glendinning@shawell.net>
> Cc: netdev@vger.kernel.org
> Signed-off-by: Lee Jones <lee.jones@linaro.org>
Looks all right to me now:
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Yours,
Linus Walleij
^ permalink raw reply
* [PATCH net-next/master] ethernet/broadcom/tg3: Fix sparse warning: constant 0x7fffffffffffffff is so big it is long long
From: Peter Huewe @ 2013-01-03 14:23 UTC (permalink / raw)
To: Matt Carlson
Cc: Nithin Nayak Sujir, Michael Chan, netdev, linux-kernel,
Fengguang Wu, kbuild, Peter Huewe
In-Reply-To: <50beaf6a.HF5L0P4z6BqrUx/x%fengguang.wu@intel.com>
Sparse complains that:
drivers/net/ethernet/broadcom/tg3.c:5670:55: sparse: constant
0x7fffffffffffffff is so big it is long long (on x86/32 bit)
so we suffix the constant with LL in the header file.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
---
drivers/net/ethernet/broadcom/tg3.h | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/tg3.h b/drivers/net/ethernet/broadcom/tg3.h
index d330e81..247d954 100644
--- a/drivers/net/ethernet/broadcom/tg3.h
+++ b/drivers/net/ethernet/broadcom/tg3.h
@@ -774,7 +774,7 @@
#define SG_DIG_AUTONEG_ERROR 0x00000001
#define TG3_TX_TSTAMP_LSB 0x000005c0
#define TG3_TX_TSTAMP_MSB 0x000005c4
-#define TG3_TSTAMP_MASK 0x7fffffffffffffff
+#define TG3_TSTAMP_MASK 0x7fffffffffffffffLL
/* 0x5c8 --> 0x600 unused */
#define MAC_TX_MAC_STATE_BASE 0x00000600 /* 16 bytes */
#define MAC_RX_MAC_STATE_BASE 0x00000610 /* 20 bytes */
--
1.7.8.6
^ permalink raw reply related
* Re: clk dereference in drivers/net/ethernet/ti/cpts.c
From: Richard Cochran @ 2013-01-03 14:00 UTC (permalink / raw)
To: Julia Lawall; +Cc: davem, tony, netdev, linux-kernel
In-Reply-To: <alpine.DEB.2.02.1301031040320.1989@hadrien>
On Thu, Jan 03, 2013 at 11:20:52AM +0100, Julia Lawall wrote:
> There has been a discussion recently about how the result of get_clk
> should be an opaque handle, not a value that can be dereferenced:
>
> https://lkml.org/lkml/2012/12/20/105
>
> There is such a dereference in drivers/net/ethernet/ti/cpts.c, in the
> function cpts_clk_init:
>
> cpts->freq = cpts->refclk->recalc(cpts->refclk);
>
> It was not obvious to me, however, what API function should be used
> instead, so I am just reporting the (potential) problem.
This issue has been fixed in v3.8-rc2.
Thanks,
Richard
^ permalink raw reply
* Re: ppoll() stuck on POLLIN while TCP peer is sending
From: Eric Dumazet @ 2013-01-03 13:41 UTC (permalink / raw)
To: Eric Wong
Cc: Mel Gorman, linux-mm, netdev, linux-kernel, Rik van Riel,
Minchan Kim, Andrew Morton, Linus Torvalds
In-Reply-To: <20130102204712.GA17806@dcvr.yhbt.net>
On Wed, 2013-01-02 at 20:47 +0000, Eric Wong wrote:
> Eric Wong <normalperson@yhbt.net> wrote:
> > [1] my full setup is very strange.
> >
> > Other than the FUSE component I forgot to mention, little depends on
> > the kernel. With all this, the standalone toosleepy can get stuck.
> > I'll try to reproduce it with less...
>
> I just confirmed my toosleepy processes will get stuck while just
> doing "rsync -a" between local disks. So this does not depend on
> sendfile or FUSE to reproduce.
> --
How do you tell your 'toosleepy' is stuck ?
If reading its output, you should change its logic, there is no
guarantee the recv() will deliver exactly 16384 bytes each round.
With the following patch, I cant reproduce the 'apparent stuck'
diff --git a/toosleepy.c b/toosleepy.c
index e64b7cd..df3610f 100644
--- a/toosleepy.c
+++ b/toosleepy.c
@@ -15,6 +15,7 @@
#include <fcntl.h>
#include <assert.h>
#include <limits.h>
+#include <time.h>
struct receiver {
int rfd;
@@ -53,6 +54,7 @@ static void * recv_loop(void *p)
ssize_t r, s;
size_t received = 0;
size_t sent = 0;
+ time_t t0 = time(NULL), t1;
for (;;) {
r = recv(rcvr->rfd, buf, sizeof(buf), 0);
@@ -80,9 +82,12 @@ static void * recv_loop(void *p)
write(-1, buf, sizeof(buf));
}
}
- if ((received % (sizeof(buf) * sizeof(buf) * 16) == 0))
+ t1 = time(NULL);
+ if (t1 != t0) {
dprintf(2, " %d progress: %zu\n",
rcvr->rfd, received);
+ t0 = t1;
+ }
}
dprintf(2, "%d got: %zu\n", rcvr->rfd, received);
if (rcvr->sfd >= 0) {
^ permalink raw reply related
* Re: [PATCH net-next] softirq: reduce latencies
From: Eric Dumazet @ 2013-01-03 13:31 UTC (permalink / raw)
To: sedat.dilek; +Cc: netdev, LKML
In-Reply-To: <CA+icZUVEvzyrPsQG1QWLnHxhtaewaWTOVzBpTnLqjKuHMjjFHg@mail.gmail.com>
On Thu, 2013-01-03 at 14:12 +0100, Sedat Dilek wrote:
> Hi Eric,
>
> your patch from [2] applies cleanly on top of Linux v3.8-rc2.
> I would like to test it.
> In [1] you were talking about benchmarks you did.
> Can you describe them or provide a testcase (script etc.)?
> You made only network testing?
Yes I did network testing :
- net_rx_action() softirq handler is the typical function that can
consume 2 ms per call.
I did some netperf sessions, with multiqueue 10G nics, tuned to that IRQ
would be handled by few cpus.
(check /proc/irq/*/eth0-$QUEUE/../smp_affinity )
Another way to make the softirq processing use more cpu cycles is by
adding a fake iptable setup like :
for n in `seq 1 100`
do
iptables -I INPUT
done
A common network load is to launch ~200 concurrent TCP_RR netperf
sessions like the following
netperf -H remote_host -t TCP_RR -l 1000
And then you can launch some netperf asking P99_LATENCY results :
netperf -H remote_host -t TCP_RR -- -k P99_LATENCY
You can play with taskset or netperf option -T to force
netperf/netserver running on given cpus.
^ permalink raw reply
* Re: [PATCH v2] tcp: split tcp_ecn sysctl knob to distinguish between IPv4 and IPv6
From: Eric Dumazet @ 2013-01-03 13:13 UTC (permalink / raw)
To: Hannes Frederic Sowa; +Cc: netdev
In-Reply-To: <20130103125736.GI4464@order.stressinduktion.org>
On Thu, 2013-01-03 at 13:57 +0100, Hannes Frederic Sowa wrote:
> Currently not, but I will try. :)
>
> Is the direction to split the sysctl ok? Or should it be a unified sysctl for
> ipv4/ipv6 and only be separate per namespace?
>
I fixed some ECN glitches in IPv6 in the past, I am not sure we can
assume IPv6 gear is ECN bug free. In fact it might be the opposite.
So adding a per net sysctl would allow a better tuning for experiments,
and allow some applications to use a private netns and change their
sysctl.
^ permalink raw reply
* Re: [PATCH net-next] softirq: reduce latencies
From: Sedat Dilek @ 2013-01-03 13:12 UTC (permalink / raw)
To: Eric Dumazet; +Cc: netdev, LKML
Hi Eric,
your patch from [2] applies cleanly on top of Linux v3.8-rc2.
I would like to test it.
In [1] you were talking about benchmarks you did.
Can you describe them or provide a testcase (script etc.)?
You made only network testing?
Thanks in advance.
Regards,
- Sedat -
[1] http://marc.info/?l=linux-kernel&m=135721614718434&w=2
[2] https://patchwork.kernel.org/patch/1927531/
^ permalink raw reply
* [PATCH 2/2] smsc95xx: enable dynamic autosuspend
From: Steve Glendinning @ 2013-01-03 13:00 UTC (permalink / raw)
To: netdev; +Cc: ming.lei, oneukum, gregkh, Steve Glendinning
In-Reply-To: <1357218016-2586-1-git-send-email-steve.glendinning@shawell.net>
This patch enables USB dynamic autosuspend for LAN9500A. This
saves very little power in itself, but it allows power saving
in upstream hubs/hosts.
The earlier devices in this family (LAN9500/9512/9514) do not
support this feature.
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
---
drivers/net/usb/smsc95xx.c | 151 +++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 150 insertions(+), 1 deletion(-)
diff --git a/drivers/net/usb/smsc95xx.c b/drivers/net/usb/smsc95xx.c
index 124e67f..6a74a68 100644
--- a/drivers/net/usb/smsc95xx.c
+++ b/drivers/net/usb/smsc95xx.c
@@ -55,6 +55,13 @@
#define FEATURE_PHY_NLP_CROSSOVER (0x02)
#define FEATURE_AUTOSUSPEND (0x04)
+#define SUSPEND_SUSPEND0 (0x01)
+#define SUSPEND_SUSPEND1 (0x02)
+#define SUSPEND_SUSPEND2 (0x04)
+#define SUSPEND_SUSPEND3 (0x08)
+#define SUSPEND_ALLMODES (SUSPEND_SUSPEND0 | SUSPEND_SUSPEND1 | \
+ SUSPEND_SUSPEND2 | SUSPEND_SUSPEND3)
+
struct smsc95xx_priv {
u32 mac_cr;
u32 hash_hi;
@@ -62,6 +69,7 @@ struct smsc95xx_priv {
u32 wolopts;
spinlock_t mac_cr_lock;
u8 features;
+ u8 suspend_flags;
};
static bool turbo_mode = true;
@@ -1242,6 +1250,8 @@ static int smsc95xx_enter_suspend0(struct usbnet *dev)
/* read back PM_CTRL */
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
+ pdata->suspend_flags |= SUSPEND_SUSPEND0;
+
return ret;
}
@@ -1286,11 +1296,14 @@ static int smsc95xx_enter_suspend1(struct usbnet *dev)
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
+ pdata->suspend_flags |= SUSPEND_SUSPEND1;
+
return ret;
}
static int smsc95xx_enter_suspend2(struct usbnet *dev)
{
+ struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
u32 val;
int ret;
@@ -1303,9 +1316,97 @@ static int smsc95xx_enter_suspend2(struct usbnet *dev)
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
+ pdata->suspend_flags |= SUSPEND_SUSPEND2;
+
return ret;
}
+static int smsc95xx_enter_suspend3(struct usbnet *dev)
+{
+ struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
+ u32 val;
+ int ret;
+
+ ret = smsc95xx_read_reg_nopm(dev, RX_FIFO_INF, &val);
+ if (ret < 0)
+ return ret;
+
+ if (val & 0xFFFF) {
+ netdev_info(dev->net, "rx fifo not empty in autosuspend\n");
+ return -EBUSY;
+ }
+
+ ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
+ if (ret < 0)
+ return ret;
+
+ val &= ~(PM_CTL_SUS_MODE_ | PM_CTL_WUPS_ | PM_CTL_PHY_RST_);
+ val |= PM_CTL_SUS_MODE_3 | PM_CTL_RES_CLR_WKP_STS;
+
+ ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
+ if (ret < 0)
+ return ret;
+
+ /* clear wol status */
+ val &= ~PM_CTL_WUPS_;
+ val |= PM_CTL_WUPS_WOL_;
+
+ ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
+ if (ret < 0)
+ return ret;
+
+ pdata->suspend_flags |= SUSPEND_SUSPEND3;
+
+ return 0;
+}
+
+static int smsc95xx_autosuspend(struct usbnet *dev, u32 link_up)
+{
+ struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
+ int ret;
+
+ if (!netif_running(dev->net)) {
+ /* interface is ifconfig down so fully power down hw */
+ netdev_dbg(dev->net, "autosuspend entering SUSPEND2\n");
+ return smsc95xx_enter_suspend2(dev);
+ }
+
+ if (!link_up) {
+ /* link is down so enter EDPD mode, but only if device can
+ * reliably resume from it. This check should be redundant
+ * as current FEATURE_AUTOSUSPEND parts also support
+ * FEATURE_PHY_NLP_CROSSOVER but it's included for clarity */
+ if (!(pdata->features & FEATURE_PHY_NLP_CROSSOVER)) {
+ netdev_warn(dev->net, "EDPD not supported\n");
+ return -EBUSY;
+ }
+
+ netdev_dbg(dev->net, "autosuspend entering SUSPEND1\n");
+
+ /* enable PHY wakeup events for if cable is attached */
+ ret = smsc95xx_enable_phy_wakeup_interrupts(dev,
+ PHY_INT_MASK_ANEG_COMP_);
+ if (ret < 0) {
+ netdev_warn(dev->net, "error enabling PHY wakeup ints\n");
+ return ret;
+ }
+
+ netdev_info(dev->net, "entering SUSPEND1 mode\n");
+ return smsc95xx_enter_suspend1(dev);
+ }
+
+ /* enable PHY wakeup events so we remote wakeup if cable is pulled */
+ ret = smsc95xx_enable_phy_wakeup_interrupts(dev,
+ PHY_INT_MASK_LINK_DOWN_);
+ if (ret < 0) {
+ netdev_warn(dev->net, "error enabling PHY wakeup ints\n");
+ return ret;
+ }
+
+ netdev_dbg(dev->net, "autosuspend entering SUSPEND3\n");
+ return smsc95xx_enter_suspend3(dev);
+}
+
static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
{
struct usbnet *dev = usb_get_intfdata(intf);
@@ -1313,15 +1414,35 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
u32 val, link_up;
int ret;
+ /* TODO: don't indicate this feature to usb framework if
+ * our current hardware doesn't have the capability
+ */
+ if ((message.event == PM_EVENT_AUTO_SUSPEND) &&
+ (!(pdata->features & FEATURE_AUTOSUSPEND))) {
+ netdev_warn(dev->net, "autosuspend not supported\n");
+ return -EBUSY;
+ }
+
ret = usbnet_suspend(intf, message);
if (ret < 0) {
netdev_warn(dev->net, "usbnet_suspend error\n");
return ret;
}
+ if (pdata->suspend_flags) {
+ netdev_warn(dev->net, "error during last resume\n");
+ pdata->suspend_flags = 0;
+ }
+
/* determine if link is up using only _nopm functions */
link_up = smsc95xx_link_ok_nopm(dev);
+ if (message.event == PM_EVENT_AUTO_SUSPEND) {
+ ret = smsc95xx_autosuspend(dev, link_up);
+ goto done;
+ }
+
+ /* if we get this far we're not autosuspending */
/* if no wol options set, or if link is down and we're not waking on
* PHY activity, enter lowest power SUSPEND2 mode
*/
@@ -1552,12 +1673,18 @@ static int smsc95xx_resume(struct usb_interface *intf)
{
struct usbnet *dev = usb_get_intfdata(intf);
struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
+ u8 suspend_flags = pdata->suspend_flags;
int ret;
u32 val;
BUG_ON(!dev);
- if (pdata->wolopts) {
+ netdev_dbg(dev->net, "resume suspend_flags=0x%02x\n", suspend_flags);
+
+ /* do this first to ensure it's cleared even in error case */
+ pdata->suspend_flags = 0;
+
+ if (suspend_flags & SUSPEND_ALLMODES) {
/* clear wake-up sources */
ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
if (ret < 0)
@@ -1741,6 +1868,26 @@ static struct sk_buff *smsc95xx_tx_fixup(struct usbnet *dev,
return skb;
}
+static int smsc95xx_manage_power(struct usbnet *dev, int on)
+{
+ struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
+
+ dev->intf->needs_remote_wakeup = on;
+
+ if (pdata->features & FEATURE_AUTOSUSPEND)
+ return 0;
+
+ /* this chip revision doesn't support autosuspend */
+ netdev_info(dev->net, "hardware doesn't support USB autosuspend\n");
+
+ if (on)
+ usb_autopm_get_interface_no_resume(dev->intf);
+ else
+ usb_autopm_put_interface(dev->intf);
+
+ return 0;
+}
+
static const struct driver_info smsc95xx_info = {
.description = "smsc95xx USB 2.0 Ethernet",
.bind = smsc95xx_bind,
@@ -1750,6 +1897,7 @@ static const struct driver_info smsc95xx_info = {
.rx_fixup = smsc95xx_rx_fixup,
.tx_fixup = smsc95xx_tx_fixup,
.status = smsc95xx_status,
+ .manage_power = smsc95xx_manage_power,
.flags = FLAG_ETHER | FLAG_SEND_ZLP | FLAG_LINK_INTR,
};
@@ -1857,6 +2005,7 @@ static struct usb_driver smsc95xx_driver = {
.reset_resume = smsc95xx_resume,
.disconnect = usbnet_disconnect,
.disable_hub_initiated_lpm = 1,
+ .supports_autosuspend = 1,
};
module_usb_driver(smsc95xx_driver);
--
1.7.10.4
^ permalink raw reply related
* [PATCH 1/2] smsc95xx: eliminate duplicate warnings on io failure
From: Steve Glendinning @ 2013-01-03 13:00 UTC (permalink / raw)
To: netdev; +Cc: ming.lei, oneukum, gregkh, Steve Glendinning, Joe Perches
In-Reply-To: <1357218016-2586-1-git-send-email-steve.glendinning@shawell.net>
The register read/write functions already log a warning if
an access fails, so this patch removes the additional warnings
logged by callers that don't add any more information.
This patch makes the resulting driver smaller by not containing
as many warning strings.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
---
drivers/net/usb/smsc95xx.c | 284 +++++++++++---------------------------------
1 file changed, 67 insertions(+), 217 deletions(-)
diff --git a/drivers/net/usb/smsc95xx.c b/drivers/net/usb/smsc95xx.c
index 9b73670..124e67f 100644
--- a/drivers/net/usb/smsc95xx.c
+++ b/drivers/net/usb/smsc95xx.c
@@ -513,10 +513,8 @@ static int smsc95xx_phy_update_flowcontrol(struct usbnet *dev, u8 duplex,
u32 flow, afc_cfg = 0;
int ret = smsc95xx_read_reg(dev, AFC_CFG, &afc_cfg);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading AFC_CFG\n");
+ if (ret < 0)
return ret;
- }
if (duplex == DUPLEX_FULL) {
u8 cap = mii_resolve_flowctrl_fdx(lcladv, rmtadv);
@@ -541,16 +539,10 @@ static int smsc95xx_phy_update_flowcontrol(struct usbnet *dev, u8 duplex,
}
ret = smsc95xx_write_reg(dev, FLOW, flow);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing FLOW\n");
- return ret;
- }
-
- ret = smsc95xx_write_reg(dev, AFC_CFG, afc_cfg);
if (ret < 0)
- netdev_warn(dev->net, "Error writing AFC_CFG\n");
+ return ret;
- return ret;
+ return smsc95xx_write_reg(dev, AFC_CFG, afc_cfg);
}
static int smsc95xx_link_reset(struct usbnet *dev)
@@ -564,16 +556,12 @@ static int smsc95xx_link_reset(struct usbnet *dev)
/* clear interrupt status */
ret = smsc95xx_mdio_read(dev->net, mii->phy_id, PHY_INT_SRC);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PHY_INT_SRC\n");
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_write_reg(dev, INT_STS, INT_STS_CLEAR_ALL_);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing INT_STS\n");
+ if (ret < 0)
return ret;
- }
mii_check_media(mii, 1, 1);
mii_ethtool_gset(&dev->mii, &ecmd);
@@ -595,10 +583,8 @@ static int smsc95xx_link_reset(struct usbnet *dev)
spin_unlock_irqrestore(&pdata->mac_cr_lock, flags);
ret = smsc95xx_write_reg(dev, MAC_CR, pdata->mac_cr);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing MAC_CR\n");
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_phy_update_flowcontrol(dev, ecmd.duplex, lcladv, rmtadv);
if (ret < 0)
@@ -638,10 +624,8 @@ static int smsc95xx_set_features(struct net_device *netdev,
int ret;
ret = smsc95xx_read_reg(dev, COE_CR, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read COE_CR: %d\n", ret);
+ if (ret < 0)
return ret;
- }
if (features & NETIF_F_HW_CSUM)
read_buf |= Tx_COE_EN_;
@@ -654,10 +638,8 @@ static int smsc95xx_set_features(struct net_device *netdev,
read_buf &= ~Rx_COE_EN_;
ret = smsc95xx_write_reg(dev, COE_CR, read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write COE_CR: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, hw, dev->net, "COE_CR = 0x%08x\n", read_buf);
return 0;
@@ -800,16 +782,10 @@ static int smsc95xx_set_mac_address(struct usbnet *dev)
int ret;
ret = smsc95xx_write_reg(dev, ADDRL, addr_lo);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write ADDRL: %d\n", ret);
- return ret;
- }
-
- ret = smsc95xx_write_reg(dev, ADDRH, addr_hi);
if (ret < 0)
- netdev_warn(dev->net, "Failed to write ADDRH: %d\n", ret);
+ return ret;
- return ret;
+ return smsc95xx_write_reg(dev, ADDRH, addr_hi);
}
/* starts the TX path */
@@ -825,17 +801,11 @@ static int smsc95xx_start_tx_path(struct usbnet *dev)
spin_unlock_irqrestore(&pdata->mac_cr_lock, flags);
ret = smsc95xx_write_reg(dev, MAC_CR, pdata->mac_cr);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write MAC_CR: %d\n", ret);
+ if (ret < 0)
return ret;
- }
/* Enable Tx at SCSRs */
- ret = smsc95xx_write_reg(dev, TX_CFG, TX_CFG_ON_);
- if (ret < 0)
- netdev_warn(dev->net, "Failed to write TX_CFG: %d\n", ret);
-
- return ret;
+ return smsc95xx_write_reg(dev, TX_CFG, TX_CFG_ON_);
}
/* Starts the Receive path */
@@ -843,17 +813,12 @@ static int smsc95xx_start_rx_path(struct usbnet *dev, int in_pm)
{
struct smsc95xx_priv *pdata = (struct smsc95xx_priv *)(dev->data[0]);
unsigned long flags;
- int ret;
spin_lock_irqsave(&pdata->mac_cr_lock, flags);
pdata->mac_cr |= MAC_CR_RXEN_;
spin_unlock_irqrestore(&pdata->mac_cr_lock, flags);
- ret = __smsc95xx_write_reg(dev, MAC_CR, pdata->mac_cr, in_pm);
- if (ret < 0)
- netdev_warn(dev->net, "Failed to write MAC_CR: %d\n", ret);
-
- return ret;
+ return __smsc95xx_write_reg(dev, MAC_CR, pdata->mac_cr, in_pm);
}
static int smsc95xx_phy_initialize(struct usbnet *dev)
@@ -910,19 +875,15 @@ static int smsc95xx_reset(struct usbnet *dev)
netif_dbg(dev, ifup, dev->net, "entering smsc95xx_reset\n");
ret = smsc95xx_write_reg(dev, HW_CFG, HW_CFG_LRST_);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write HW_CFG_LRST_ bit in HW_CFG\n");
+ if (ret < 0)
return ret;
- }
timeout = 0;
do {
msleep(10);
ret = smsc95xx_read_reg(dev, HW_CFG, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read HW_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
timeout++;
} while ((read_buf & HW_CFG_LRST_) && (timeout < 100));
@@ -932,19 +893,15 @@ static int smsc95xx_reset(struct usbnet *dev)
}
ret = smsc95xx_write_reg(dev, PM_CTRL, PM_CTL_PHY_RST_);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write PM_CTRL: %d\n", ret);
+ if (ret < 0)
return ret;
- }
timeout = 0;
do {
msleep(10);
ret = smsc95xx_read_reg(dev, PM_CTRL, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read PM_CTRL: %d\n", ret);
+ if (ret < 0)
return ret;
- }
timeout++;
} while ((read_buf & PM_CTL_PHY_RST_) && (timeout < 100));
@@ -961,10 +918,8 @@ static int smsc95xx_reset(struct usbnet *dev)
dev->net->dev_addr);
ret = smsc95xx_read_reg(dev, HW_CFG, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read HW_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, ifup, dev->net, "Read Value from HW_CFG : 0x%08x\n",
read_buf);
@@ -972,16 +927,12 @@ static int smsc95xx_reset(struct usbnet *dev)
read_buf |= HW_CFG_BIR_;
ret = smsc95xx_write_reg(dev, HW_CFG, read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write HW_CFG_BIR_ bit in HW_CFG\n");
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_read_reg(dev, HW_CFG, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read HW_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, ifup, dev->net,
"Read Value from HW_CFG after writing HW_CFG_BIR_: 0x%08x\n",
@@ -1002,42 +953,32 @@ static int smsc95xx_reset(struct usbnet *dev)
(ulong)dev->rx_urb_size);
ret = smsc95xx_write_reg(dev, BURST_CAP, burst_cap);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write BURST_CAP: %d\n", ret);
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_read_reg(dev, BURST_CAP, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read BURST_CAP: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, ifup, dev->net,
"Read Value from BURST_CAP after writing: 0x%08x\n",
read_buf);
ret = smsc95xx_write_reg(dev, BULK_IN_DLY, DEFAULT_BULK_IN_DELAY);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write BULK_IN_DLY: %d\n", ret);
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_read_reg(dev, BULK_IN_DLY, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read BULK_IN_DLY: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, ifup, dev->net,
"Read Value from BULK_IN_DLY after writing: 0x%08x\n",
read_buf);
ret = smsc95xx_read_reg(dev, HW_CFG, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read HW_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, ifup, dev->net, "Read Value from HW_CFG: 0x%08x\n",
read_buf);
@@ -1051,69 +992,51 @@ static int smsc95xx_reset(struct usbnet *dev)
read_buf |= NET_IP_ALIGN << 9;
ret = smsc95xx_write_reg(dev, HW_CFG, read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write HW_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_read_reg(dev, HW_CFG, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read HW_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, ifup, dev->net,
"Read Value from HW_CFG after writing: 0x%08x\n", read_buf);
ret = smsc95xx_write_reg(dev, INT_STS, INT_STS_CLEAR_ALL_);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write INT_STS: %d\n", ret);
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_read_reg(dev, ID_REV, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read ID_REV: %d\n", ret);
+ if (ret < 0)
return ret;
- }
netif_dbg(dev, ifup, dev->net, "ID_REV = 0x%08x\n", read_buf);
/* Configure GPIO pins as LED outputs */
write_buf = LED_GPIO_CFG_SPD_LED | LED_GPIO_CFG_LNK_LED |
LED_GPIO_CFG_FDX_LED;
ret = smsc95xx_write_reg(dev, LED_GPIO_CFG, write_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write LED_GPIO_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
/* Init Tx */
ret = smsc95xx_write_reg(dev, FLOW, 0);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write FLOW: %d\n", ret);
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_write_reg(dev, AFC_CFG, AFC_CFG_DEFAULT);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write AFC_CFG: %d\n", ret);
+ if (ret < 0)
return ret;
- }
/* Don't need mac_cr_lock during initialisation */
ret = smsc95xx_read_reg(dev, MAC_CR, &pdata->mac_cr);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read MAC_CR: %d\n", ret);
+ if (ret < 0)
return ret;
- }
/* Init Rx */
/* Set Vlan */
ret = smsc95xx_write_reg(dev, VLAN1, (u32)ETH_P_8021Q);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write VLAN1: %d\n", ret);
+ if (ret < 0)
return ret;
- }
/* Enable or disable checksum offload engines */
ret = smsc95xx_set_features(dev->net, dev->net->features);
@@ -1131,19 +1054,15 @@ static int smsc95xx_reset(struct usbnet *dev)
}
ret = smsc95xx_read_reg(dev, INT_EP_CTL, &read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read INT_EP_CTL: %d\n", ret);
+ if (ret < 0)
return ret;
- }
/* enable PHY interrupts */
read_buf |= INT_EP_CTL_PHY_INT_;
ret = smsc95xx_write_reg(dev, INT_EP_CTL, read_buf);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to write INT_EP_CTL: %d\n", ret);
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_start_tx_path(dev);
if (ret < 0) {
@@ -1213,10 +1132,8 @@ static int smsc95xx_bind(struct usbnet *dev, struct usb_interface *intf)
/* detect device revision as different features may be available */
ret = smsc95xx_read_reg(dev, ID_REV, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Failed to read ID_REV: %d\n", ret);
+ if (ret < 0)
return ret;
- }
val >>= 16;
if ((val == ID_REV_CHIP_ID_9500A_) || (val == ID_REV_CHIP_ID_9530_) ||
@@ -1261,17 +1178,13 @@ static int smsc95xx_enable_phy_wakeup_interrupts(struct usbnet *dev, u16 mask)
/* read to clear */
ret = smsc95xx_mdio_read_nopm(dev->net, mii->phy_id, PHY_INT_SRC);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PHY_INT_SRC\n");
+ if (ret < 0)
return ret;
- }
/* enable interrupt source */
ret = smsc95xx_mdio_read_nopm(dev->net, mii->phy_id, PHY_INT_MASK);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PHY_INT_MASK\n");
+ if (ret < 0)
return ret;
- }
ret |= mask;
@@ -1287,16 +1200,12 @@ static int smsc95xx_link_ok_nopm(struct usbnet *dev)
/* first, a dummy read, needed to latch some MII phys */
ret = smsc95xx_mdio_read_nopm(dev->net, mii->phy_id, MII_BMSR);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading MII_BMSR\n");
+ if (ret < 0)
return ret;
- }
ret = smsc95xx_mdio_read_nopm(dev->net, mii->phy_id, MII_BMSR);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading MII_BMSR\n");
+ if (ret < 0)
return ret;
- }
return !!(ret & BMSR_LSTATUS);
}
@@ -1308,19 +1217,15 @@ static int smsc95xx_enter_suspend0(struct usbnet *dev)
int ret;
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
val &= (~(PM_CTL_SUS_MODE_ | PM_CTL_WUPS_ | PM_CTL_PHY_RST_));
val |= PM_CTL_SUS_MODE_0;
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
/* clear wol status */
val &= ~PM_CTL_WUPS_;
@@ -1331,15 +1236,11 @@ static int smsc95xx_enter_suspend0(struct usbnet *dev)
val |= PM_CTL_WUPS_ED_;
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
/* read back PM_CTRL */
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
- if (ret < 0)
- netdev_warn(dev->net, "Error reading PM_CTRL\n");
return ret;
}
@@ -1360,10 +1261,8 @@ static int smsc95xx_enter_suspend1(struct usbnet *dev)
/* enable energy detect power-down mode */
ret = smsc95xx_mdio_read_nopm(dev->net, mii->phy_id, PHY_MODE_CTRL_STS);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PHY_MODE_CTRL_STS\n");
+ if (ret < 0)
return ret;
- }
ret |= MODE_CTRL_STS_EDPWRDOWN_;
@@ -1371,27 +1270,21 @@ static int smsc95xx_enter_suspend1(struct usbnet *dev)
/* enter SUSPEND1 mode */
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
val &= ~(PM_CTL_SUS_MODE_ | PM_CTL_WUPS_ | PM_CTL_PHY_RST_);
val |= PM_CTL_SUS_MODE_1;
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
/* clear wol status, enable energy detection */
val &= ~PM_CTL_WUPS_;
val |= (PM_CTL_WUPS_ED_ | PM_CTL_ED_EN_);
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0)
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
return ret;
}
@@ -1402,17 +1295,13 @@ static int smsc95xx_enter_suspend2(struct usbnet *dev)
int ret;
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
val &= ~(PM_CTL_SUS_MODE_ | PM_CTL_WUPS_ | PM_CTL_PHY_RST_);
val |= PM_CTL_SUS_MODE_2;
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0)
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
return ret;
}
@@ -1442,32 +1331,24 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
/* disable energy detect (link up) & wake up events */
ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading WUCSR\n");
+ if (ret < 0)
goto done;
- }
val &= ~(WUCSR_MPEN_ | WUCSR_WAKE_EN_);
ret = smsc95xx_write_reg_nopm(dev, WUCSR, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUCSR\n");
+ if (ret < 0)
goto done;
- }
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PM_CTRL\n");
+ if (ret < 0)
goto done;
- }
val &= ~(PM_CTL_ED_EN_ | PM_CTL_WOL_EN_);
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ if (ret < 0)
goto done;
- }
ret = smsc95xx_enter_suspend2(dev);
goto done;
@@ -1565,7 +1446,6 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
for (i = 0; i < (wuff_filter_count * 4); i++) {
ret = smsc95xx_write_reg_nopm(dev, WUFF, filter_mask[i]);
if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUFF\n");
kfree(filter_mask);
goto done;
}
@@ -1574,67 +1454,51 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
for (i = 0; i < (wuff_filter_count / 4); i++) {
ret = smsc95xx_write_reg_nopm(dev, WUFF, command[i]);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUFF\n");
+ if (ret < 0)
goto done;
- }
}
for (i = 0; i < (wuff_filter_count / 4); i++) {
ret = smsc95xx_write_reg_nopm(dev, WUFF, offset[i]);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUFF\n");
+ if (ret < 0)
goto done;
- }
}
for (i = 0; i < (wuff_filter_count / 2); i++) {
ret = smsc95xx_write_reg_nopm(dev, WUFF, crc[i]);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUFF\n");
+ if (ret < 0)
goto done;
- }
}
/* clear any pending pattern match packet status */
ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading WUCSR\n");
+ if (ret < 0)
goto done;
- }
val |= WUCSR_WUFR_;
ret = smsc95xx_write_reg_nopm(dev, WUCSR, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUCSR\n");
+ if (ret < 0)
goto done;
- }
}
if (pdata->wolopts & WAKE_MAGIC) {
/* clear any pending magic packet status */
ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading WUCSR\n");
+ if (ret < 0)
goto done;
- }
val |= WUCSR_MPR_;
ret = smsc95xx_write_reg_nopm(dev, WUCSR, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUCSR\n");
+ if (ret < 0)
goto done;
- }
}
/* enable/disable wakeup sources */
ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading WUCSR\n");
+ if (ret < 0)
goto done;
- }
if (pdata->wolopts & (WAKE_BCAST | WAKE_MCAST | WAKE_ARP | WAKE_UCAST)) {
netdev_info(dev->net, "enabling pattern match wakeup\n");
@@ -1653,17 +1517,13 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
}
ret = smsc95xx_write_reg_nopm(dev, WUCSR, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUCSR\n");
+ if (ret < 0)
goto done;
- }
/* enable wol wakeup source */
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PM_CTRL\n");
+ if (ret < 0)
goto done;
- }
val |= PM_CTL_WOL_EN_;
@@ -1672,10 +1532,8 @@ static int smsc95xx_suspend(struct usb_interface *intf, pm_message_t message)
val |= PM_CTL_ED_EN_;
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ if (ret < 0)
goto done;
- }
/* enable receiver to enable frame reception */
smsc95xx_start_rx_path(dev, 1);
@@ -1702,34 +1560,26 @@ static int smsc95xx_resume(struct usb_interface *intf)
if (pdata->wolopts) {
/* clear wake-up sources */
ret = smsc95xx_read_reg_nopm(dev, WUCSR, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading WUCSR\n");
+ if (ret < 0)
return ret;
- }
val &= ~(WUCSR_WAKE_EN_ | WUCSR_MPEN_);
ret = smsc95xx_write_reg_nopm(dev, WUCSR, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing WUCSR\n");
+ if (ret < 0)
return ret;
- }
/* clear wake-up status */
ret = smsc95xx_read_reg_nopm(dev, PM_CTRL, &val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error reading PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
val &= ~PM_CTL_WOL_EN_;
val |= PM_CTL_WUPS_;
ret = smsc95xx_write_reg_nopm(dev, PM_CTRL, val);
- if (ret < 0) {
- netdev_warn(dev->net, "Error writing PM_CTRL\n");
+ if (ret < 0)
return ret;
- }
}
ret = usbnet_resume(intf);
--
1.7.10.4
^ permalink raw reply related
* [PATCH 0/2] smsc95xx enhancements
From: Steve Glendinning @ 2013-01-03 13:00 UTC (permalink / raw)
To: netdev; +Cc: ming.lei, oneukum, gregkh, Steve Glendinning
Two driver enhancements for net-next. There's ongoing discussion around
the best way to improve autosuspend moving forwards but in the meantime
the second patch in this set enables autosuspend for supported parts in
most situations.
Steve Glendinning (2):
smsc95xx: eliminate duplicate warnings on io failure
smsc95xx: enable dynamic autosuspend
drivers/net/usb/smsc95xx.c | 435 ++++++++++++++++++++++----------------------
1 file changed, 217 insertions(+), 218 deletions(-)
--
1.7.10.4
^ permalink raw reply
* Re: [PATCH v2] tcp: split tcp_ecn sysctl knob to distinguish between IPv4 and IPv6
From: Hannes Frederic Sowa @ 2013-01-03 12:58 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20130103.045350.1322196416782379673.davem@davemloft.net>
On Thu, Jan 03, 2013 at 04:53:50AM -0800, David Miller wrote:
> From: Hannes Frederic Sowa <hannes@stressinduktion.org>
> Date: Thu, 3 Jan 2013 13:28:44 +0100
>
> > On Thu, Jan 03, 2013 at 01:00:09AM +0100, Hannes Frederic Sowa wrote:
> >> ECN could be more reliable when used with IPv6 (I don't have proofs). For
> >> people who want to try ECN with IPv6 but still have problems connecting
> >> to destinations because of broken IPv4 routers this switch allows one
> >> to enable ECN just for IPv6.
> >>
> >> Perhaps ECN could be enabled by default in future.
> >>
> >> No code changes since initial submission, just added documentation.
> >
> > May I ask what lead to the rejection of the patch?
>
> I don't see any value in splitting up ipv4 and ipv6.
>
> I think we should turn the existing knob on unliaterally
> by default. That would be a much better patch.
Oh, I would be happy to do that. :)
^ permalink raw reply
* Re: [PATCH v2] tcp: split tcp_ecn sysctl knob to distinguish between IPv4 and IPv6
From: Hannes Frederic Sowa @ 2013-01-03 12:57 UTC (permalink / raw)
To: Eric Dumazet; +Cc: netdev
In-Reply-To: <1357217260.21409.24224.camel@edumazet-glaptop>
On Thu, Jan 03, 2013 at 04:47:40AM -0800, Eric Dumazet wrote:
> On Thu, 2013-01-03 at 01:00 +0100, Hannes Frederic Sowa wrote:
> > ECN could be more reliable when used with IPv6 (I don't have proofs). For
> > people who want to try ECN with IPv6 but still have problems connecting
> > to destinations because of broken IPv4 routers this switch allows one
> > to enable ECN just for IPv6.
> >
> > Perhaps ECN could be enabled by default in future.
> >
> > No code changes since initial submission, just added documentation.
> >
> > Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
> > ---
>
> It would be nice to make the ecn sysctl per net namespace.
>
> Do you have an idea of how to adapt your patch to get this as well ?
Currently not, but I will try. :)
Is the direction to split the sysctl ok? Or should it be a unified sysctl for
ipv4/ipv6 and only be separate per namespace?
Thanks!
^ permalink raw reply
* Re: [PATCH v2] tcp: split tcp_ecn sysctl knob to distinguish between IPv4 and IPv6
From: David Miller @ 2013-01-03 12:53 UTC (permalink / raw)
To: hannes; +Cc: netdev
In-Reply-To: <20130103122844.GH4464@order.stressinduktion.org>
From: Hannes Frederic Sowa <hannes@stressinduktion.org>
Date: Thu, 3 Jan 2013 13:28:44 +0100
> On Thu, Jan 03, 2013 at 01:00:09AM +0100, Hannes Frederic Sowa wrote:
>> ECN could be more reliable when used with IPv6 (I don't have proofs). For
>> people who want to try ECN with IPv6 but still have problems connecting
>> to destinations because of broken IPv4 routers this switch allows one
>> to enable ECN just for IPv6.
>>
>> Perhaps ECN could be enabled by default in future.
>>
>> No code changes since initial submission, just added documentation.
>
> May I ask what lead to the rejection of the patch?
I don't see any value in splitting up ipv4 and ipv6.
I think we should turn the existing knob on unliaterally
by default. That would be a much better patch.
^ permalink raw reply
* Re: [PATCH v2] tcp: split tcp_ecn sysctl knob to distinguish between IPv4 and IPv6
From: Eric Dumazet @ 2013-01-03 12:47 UTC (permalink / raw)
To: Hannes Frederic Sowa; +Cc: netdev
In-Reply-To: <20130103000009.GG4464@order.stressinduktion.org>
On Thu, 2013-01-03 at 01:00 +0100, Hannes Frederic Sowa wrote:
> ECN could be more reliable when used with IPv6 (I don't have proofs). For
> people who want to try ECN with IPv6 but still have problems connecting
> to destinations because of broken IPv4 routers this switch allows one
> to enable ECN just for IPv6.
>
> Perhaps ECN could be enabled by default in future.
>
> No code changes since initial submission, just added documentation.
>
> Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
> ---
It would be nice to make the ecn sysctl per net namespace.
Do you have an idea of how to adapt your patch to get this as well ?
^ permalink raw reply
* Re: [PATCH v2 net-next] team: add ethtool support
From: Flavio Leitner @ 2013-01-03 12:32 UTC (permalink / raw)
To: Ben Hutchings; +Cc: netdev, Jiri Pirko
In-Reply-To: <1356992186.31299.1.camel@deadeye.wl.decadent.org.uk>
On Mon, Dec 31, 2012 at 10:16:26PM +0000, Ben Hutchings wrote:
> On Sun, 2012-12-30 at 00:37 -0200, Flavio Leitner wrote:
> > This patch adds few ethtool operations to team driver.
> >
> > Signed-off-by: Flavio Leitner <fbl@redhat.com>
> > ---
> > v2 - removed generic statistics from ethtool
> >
> > drivers/net/team/team.c | 17 +++++++++++++++++
> > 1 file changed, 17 insertions(+)
> >
> > diff --git a/drivers/net/team/team.c b/drivers/net/team/team.c
> > index ad86660..7665a088 100644
> > --- a/drivers/net/team/team.c
> > +++ b/drivers/net/team/team.c
> [...]
> > +static void team_ethtool_get_drvinfo(struct net_device *dev,
> > + struct ethtool_drvinfo *drvinfo)
> > +{
> > + strncpy(drvinfo->driver, DRV_NAME, 32);
> > + strncpy(drvinfo->version, UTS_RELEASE, 32);
> > +}
> [...]
>
> These must be null-terminated, so use strlcpy() not strncpy().
Ok, I will post another patch fixing this and few other drivers as
well.
Thank you!
--
fbl
^ permalink raw reply
* [PATCH net-next] softirq: reduce latencies
From: Eric Dumazet @ 2013-01-03 12:28 UTC (permalink / raw)
To: David Miller, Andrew Morton
Cc: netdev, linux-kernel@vger.kernel.org, Tom Herbert
From: Eric Dumazet <edumazet@google.com>
In various network workloads, __do_softirq() latencies can be up
to 20 ms if HZ=1000, and 200 ms if HZ=100.
This is because we iterate 10 times in the softirq dispatcher,
and some actions can consume a lot of cycles.
This patch changes the fallback to ksoftirqd condition to :
- A time limit of 2 ms.
- need_resched() being set on current task
When one of this condition is met, we wakeup ksoftirqd for further
softirq processing if we still have pending softirqs.
I ran several benchmarks and got no significant difference in
throughput, but a very significant reduction of maximum latencies.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: David Miller <davem@davemloft.net>
Cc: Tom Herbert <therbert@google.com>
---
kernel/softirq.c | 17 +++++++++--------
1 file changed, 9 insertions(+), 8 deletions(-)
diff --git a/kernel/softirq.c b/kernel/softirq.c
index ed567ba..64d61ea 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -195,21 +195,21 @@ void local_bh_enable_ip(unsigned long ip)
EXPORT_SYMBOL(local_bh_enable_ip);
/*
- * We restart softirq processing MAX_SOFTIRQ_RESTART times,
- * and we fall back to softirqd after that.
+ * We restart softirq processing for at most 2 ms,
+ * and if need_resched() is not set.
*
- * This number has been established via experimentation.
+ * These limits have been established via experimentation.
* The two things to balance is latency against fairness -
* we want to handle softirqs as soon as possible, but they
* should not be able to lock up the box.
*/
-#define MAX_SOFTIRQ_RESTART 10
+#define MAX_SOFTIRQ_TIME min(1, (2*HZ/1000))
asmlinkage void __do_softirq(void)
{
struct softirq_action *h;
__u32 pending;
- int max_restart = MAX_SOFTIRQ_RESTART;
+ unsigned long end = jiffies + MAX_SOFTIRQ_TIME;
int cpu;
unsigned long old_flags = current->flags;
@@ -264,11 +264,12 @@ restart:
local_irq_disable();
pending = local_softirq_pending();
- if (pending && --max_restart)
- goto restart;
+ if (pending) {
+ if (time_before(jiffies, end) && !need_resched())
+ goto restart;
- if (pending)
wakeup_softirqd();
+ }
lockdep_softirq_exit();
^ permalink raw reply related
* Re: [PATCH v2] tcp: split tcp_ecn sysctl knob to distinguish between IPv4 and IPv6
From: Hannes Frederic Sowa @ 2013-01-03 12:28 UTC (permalink / raw)
To: netdev; +Cc: davem
In-Reply-To: <20130103000009.GG4464@order.stressinduktion.org>
On Thu, Jan 03, 2013 at 01:00:09AM +0100, Hannes Frederic Sowa wrote:
> ECN could be more reliable when used with IPv6 (I don't have proofs). For
> people who want to try ECN with IPv6 but still have problems connecting
> to destinations because of broken IPv4 routers this switch allows one
> to enable ECN just for IPv6.
>
> Perhaps ECN could be enabled by default in future.
>
> No code changes since initial submission, just added documentation.
May I ask what lead to the rejection of the patch?
^ permalink raw reply
* [PATCH -v2 25/26] net: rename random32 to prandom
From: Akinobu Mita @ 2013-01-03 12:19 UTC (permalink / raw)
To: linux-kernel, akpm; +Cc: Akinobu Mita, David S. Miller, netdev
In-Reply-To: <1357215562-6288-1-git-send-email-akinobu.mita@gmail.com>
Commit 496f2f93b1cc286f5a4f4f9acdc1e5314978683f ("random32: rename
random32 to prandom") renamed random32() and srandom32() to prandom_u32()
and prandom_seed() respectively.
net_random() and net_srandom() need to be redefined with prandom_* in
order to finish the naming transition.
While I'm at it, enclose macro argument of net_srandom() with parenthesis.
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
---
* New patch from v2
include/linux/net.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/include/linux/net.h b/include/linux/net.h
index aa16731..99c9f0c 100644
--- a/include/linux/net.h
+++ b/include/linux/net.h
@@ -240,8 +240,8 @@ do { \
#define net_dbg_ratelimited(fmt, ...) \
net_ratelimited_function(pr_debug, fmt, ##__VA_ARGS__)
-#define net_random() random32()
-#define net_srandom(seed) srandom32((__force u32)seed)
+#define net_random() prandom_u32()
+#define net_srandom(seed) prandom_seed((__force u32)(seed))
extern int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
struct kvec *vec, size_t num, size_t len);
--
1.7.11.7
^ permalink raw reply related
* [PATCH -v2 20/26] net/sunrpc: rename random32() to prandom_u32()
From: Akinobu Mita @ 2013-01-03 12:19 UTC (permalink / raw)
To: linux-kernel, akpm
Cc: Akinobu Mita, J. Bruce Fields, Trond Myklebust, David S. Miller,
netdev, linux-nfs
In-Reply-To: <1357215562-6288-1-git-send-email-akinobu.mita@gmail.com>
Use more preferable function name which implies using a pseudo-random
number generator.
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
Cc: linux-nfs@vger.kernel.org
---
* Change from v1
- leave net_random() callers as-is because that is a useful indirection
net/sunrpc/auth_gss/gss_krb5_wrap.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/sunrpc/auth_gss/gss_krb5_wrap.c b/net/sunrpc/auth_gss/gss_krb5_wrap.c
index 107c452..daf301f 100644
--- a/net/sunrpc/auth_gss/gss_krb5_wrap.c
+++ b/net/sunrpc/auth_gss/gss_krb5_wrap.c
@@ -130,8 +130,8 @@ gss_krb5_make_confounder(char *p, u32 conflen)
/* initialize to random value */
if (i == 0) {
- i = random32();
- i = (i << 32) | random32();
+ i = prandom_u32();
+ i = (i << 32) | prandom_u32();
}
switch (conflen) {
--
1.7.11.7
^ permalink raw reply related
* [PATCH -v2 24/26] net/core: remove duplicate statements by do-while loop
From: Akinobu Mita @ 2013-01-03 12:19 UTC (permalink / raw)
To: linux-kernel, akpm; +Cc: Akinobu Mita, David S. Miller, netdev
In-Reply-To: <1357215562-6288-1-git-send-email-akinobu.mita@gmail.com>
Remove duplicate statements by using do-while loop instead of while loop.
- A;
- while (e) {
+ do {
A;
- }
+ } while (e);
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
---
No change from v1
net/core/pktgen.c | 15 ++++++---------
1 file changed, 6 insertions(+), 9 deletions(-)
diff --git a/net/core/pktgen.c b/net/core/pktgen.c
index d23ec86..48a3a29 100644
--- a/net/core/pktgen.c
+++ b/net/core/pktgen.c
@@ -2379,18 +2379,15 @@ static void mod_cur_headers(struct pktgen_dev *pkt_dev)
__be32 s;
if (pkt_dev->flags & F_IPDST_RND) {
- t = prandom_u32() % (imx - imn) + imn;
- s = htonl(t);
-
- while (ipv4_is_loopback(s) ||
- ipv4_is_multicast(s) ||
- ipv4_is_lbcast(s) ||
- ipv4_is_zeronet(s) ||
- ipv4_is_local_multicast(s)) {
+ do {
t = prandom_u32() %
(imx - imn) + imn;
s = htonl(t);
- }
+ } while (ipv4_is_loopback(s) ||
+ ipv4_is_multicast(s) ||
+ ipv4_is_lbcast(s) ||
+ ipv4_is_zeronet(s) ||
+ ipv4_is_local_multicast(s));
pkt_dev->cur_daddr = s;
} else {
t = ntohl(pkt_dev->cur_daddr);
--
1.7.11.7
^ permalink raw reply related
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