* [PATCH net v2] net: smc91x: Fix build without gpiolib
From: Tobias Klauser @ 2014-12-12 16:07 UTC (permalink / raw)
To: Nicolas Pitre; +Cc: David S. Miller, Tony Lindgren, netdev
If GPIOLIB=n the following build errors occur:
drivers/net/ethernet/smsc/smc91x.c: In function 'try_toggle_control_gpio':
drivers/net/ethernet/smsc/smc91x.c:2204:2: error: implicit declaration of function 'devm_gpiod_get_index' [-Werror=implicit-function-declaration]
drivers/net/ethernet/smsc/smc91x.c:2204:7: warning: assignment makes pointer from integer without a cast [enabled by default]
drivers/net/ethernet/smsc/smc91x.c:2213:2: error: implicit declaration of function 'gpiod_direction_output' [-Werror=implicit-function-declaration]
drivers/net/ethernet/smsc/smc91x.c:2216:3: error: implicit declaration of function 'devm_gpiod_put' [-Werror=implicit-function-declaration]
drivers/net/ethernet/smsc/smc91x.c:2222:2: error: implicit declaration of function 'gpiod_set_value_cansleep' [-Werror=implicit-function-declaration]
Since the toggling of the GPIOs is an optional feature, define
try_toggle_control_gpio only if GPIOLIB is enabled.
Fixes: 7d2911c4381 ("net: smc91x: Fix gpios for device tree based booting")
Cc: Tony Lindgren <tony@atomide.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
---
drivers/net/ethernet/smsc/smc91x.c | 12 +++++++++++-
1 file changed, 11 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/smsc/smc91x.c b/drivers/net/ethernet/smsc/smc91x.c
index 6cc3cf6..050bcb6 100644
--- a/drivers/net/ethernet/smsc/smc91x.c
+++ b/drivers/net/ethernet/smsc/smc91x.c
@@ -2193,6 +2193,7 @@ MODULE_DEVICE_TABLE(of, smc91x_match);
/**
* of_try_set_control_gpio - configure a gpio if it exists
*/
+#ifdef CONFIG_GPIOLIB
static int try_toggle_control_gpio(struct device *dev,
struct gpio_desc **desc,
const char *name, int index,
@@ -2224,7 +2225,16 @@ static int try_toggle_control_gpio(struct device *dev,
return 0;
}
-#endif
+#else
+static int try_toggle_control_gpio(struct device *dev,
+ struct gpio_desc **desc,
+ const char *name, int index,
+ int value, unsigned int nsdelay)
+{
+ return 0;
+}
+#endif /* CONFIG_GPIOLIB */
+#endif /* CONFIG_OF */
/*
* smc_init(void)
--
2.2.0
^ permalink raw reply related
* Re: [net PATCH] fib_trie: Fix trie balancing issue if new node pushes down existing node
From: Alexander Duyck @ 2014-12-12 16:09 UTC (permalink / raw)
To: David Miller, alexander.h.duyck; +Cc: netdev
In-Reply-To: <20141212.105425.2023800747122061277.davem@davemloft.net>
On 12/12/2014 07:54 AM, David Miller wrote:
> From: David Miller <davem@davemloft.net>
> Date: Thu, 11 Dec 2014 21:32:16 -0500 (EST)
>
>> From: Alexander Duyck <alexander.h.duyck@redhat.com>
>> Date: Wed, 10 Dec 2014 21:49:22 -0800
>>
>>> This patch addresses an issue with the level compression of the fib_trie.
>>> Specifically in the case of adding a new leaf that triggers a new node to
>>> be added that takes the place of the old node. The result is a trie where
>>> the 1 child tnode is on one side and one leaf is on the other which gives
>>> you a very deep trie. Below is the script I used to generate a trie on
>>> dummy0 with a 10.X.X.X family of addresses.
>> ...
>>> What this fix does is start the rebalance at the newly created tnode
>>> instead of at the parent tnode. This way if there is a gap between the
>>> parent and the new node it doesn't prevent the new tnode from being
>>> coalesced with any pre-existing nodes that may have been pushed into one
>>> of the new nodes child branches.
>>>
>>> Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
>> One has to be mindful with this code that what it's doing now might
>> be intentional. For example, it might be doing things this way
>> on purpose in order to minimize rebalancing during route flaps.
>>
>> Barring anything like that, I think your change is fine.
> Alex, in case it isn't clear, I'm hoping that you might have some
> thoughts on this aspect of your changes. :)
Route flaps should at most trigger an inflate without a deflate due to
the way the resize logic works. If this occurs often it would be useful
since then there would already be space in the tree for the next time
around.
- Alex
^ permalink raw reply
* Re: [PATCH net v2] net: smc91x: Fix build without gpiolib
From: Tony Lindgren @ 2014-12-12 16:11 UTC (permalink / raw)
To: Tobias Klauser; +Cc: Nicolas Pitre, David S. Miller, netdev
In-Reply-To: <1418400429-6680-1-git-send-email-tklauser@distanz.ch>
* Tobias Klauser <tklauser@distanz.ch> [141212 08:09]:
> If GPIOLIB=n the following build errors occur:
>
> drivers/net/ethernet/smsc/smc91x.c: In function 'try_toggle_control_gpio':
> drivers/net/ethernet/smsc/smc91x.c:2204:2: error: implicit declaration of function 'devm_gpiod_get_index' [-Werror=implicit-function-declaration]
> drivers/net/ethernet/smsc/smc91x.c:2204:7: warning: assignment makes pointer from integer without a cast [enabled by default]
> drivers/net/ethernet/smsc/smc91x.c:2213:2: error: implicit declaration of function 'gpiod_direction_output' [-Werror=implicit-function-declaration]
> drivers/net/ethernet/smsc/smc91x.c:2216:3: error: implicit declaration of function 'devm_gpiod_put' [-Werror=implicit-function-declaration]
> drivers/net/ethernet/smsc/smc91x.c:2222:2: error: implicit declaration of function 'gpiod_set_value_cansleep' [-Werror=implicit-function-declaration]
>
> Since the toggling of the GPIOs is an optional feature, define
> try_toggle_control_gpio only if GPIOLIB is enabled.
Oops sorry about that, I guess my systems have it all selected so
randconfig builds did not expose it.
Would it make sense to add the missing gpio stub into
include/linux/gpio/consumer.h instead?
Regards,
Tony
> Fixes: 7d2911c4381 ("net: smc91x: Fix gpios for device tree based booting")
> Cc: Tony Lindgren <tony@atomide.com>
> Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
> ---
> drivers/net/ethernet/smsc/smc91x.c | 12 +++++++++++-
> 1 file changed, 11 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/smsc/smc91x.c b/drivers/net/ethernet/smsc/smc91x.c
> index 6cc3cf6..050bcb6 100644
> --- a/drivers/net/ethernet/smsc/smc91x.c
> +++ b/drivers/net/ethernet/smsc/smc91x.c
> @@ -2193,6 +2193,7 @@ MODULE_DEVICE_TABLE(of, smc91x_match);
> /**
> * of_try_set_control_gpio - configure a gpio if it exists
> */
> +#ifdef CONFIG_GPIOLIB
> static int try_toggle_control_gpio(struct device *dev,
> struct gpio_desc **desc,
> const char *name, int index,
> @@ -2224,7 +2225,16 @@ static int try_toggle_control_gpio(struct device *dev,
>
> return 0;
> }
> -#endif
> +#else
> +static int try_toggle_control_gpio(struct device *dev,
> + struct gpio_desc **desc,
> + const char *name, int index,
> + int value, unsigned int nsdelay)
> +{
> + return 0;
> +}
> +#endif /* CONFIG_GPIOLIB */
> +#endif /* CONFIG_OF */
>
> /*
> * smc_init(void)
> --
> 2.2.0
>
>
^ permalink raw reply
* Re: [PATCH net v2] net: smc91x: Fix build without gpiolib
From: David Miller @ 2014-12-12 16:21 UTC (permalink / raw)
To: tony; +Cc: tklauser, nico, netdev
In-Reply-To: <20141212161100.GH2950@atomide.com>
From: Tony Lindgren <tony@atomide.com>
Date: Fri, 12 Dec 2014 08:11:00 -0800
> * Tobias Klauser <tklauser@distanz.ch> [141212 08:09]:
>> If GPIOLIB=n the following build errors occur:
>>
>> drivers/net/ethernet/smsc/smc91x.c: In function 'try_toggle_control_gpio':
>> drivers/net/ethernet/smsc/smc91x.c:2204:2: error: implicit declaration of function 'devm_gpiod_get_index' [-Werror=implicit-function-declaration]
>> drivers/net/ethernet/smsc/smc91x.c:2204:7: warning: assignment makes pointer from integer without a cast [enabled by default]
>> drivers/net/ethernet/smsc/smc91x.c:2213:2: error: implicit declaration of function 'gpiod_direction_output' [-Werror=implicit-function-declaration]
>> drivers/net/ethernet/smsc/smc91x.c:2216:3: error: implicit declaration of function 'devm_gpiod_put' [-Werror=implicit-function-declaration]
>> drivers/net/ethernet/smsc/smc91x.c:2222:2: error: implicit declaration of function 'gpiod_set_value_cansleep' [-Werror=implicit-function-declaration]
>>
>> Since the toggling of the GPIOs is an optional feature, define
>> try_toggle_control_gpio only if GPIOLIB is enabled.
>
> Oops sorry about that, I guess my systems have it all selected so
> randconfig builds did not expose it.
>
> Would it make sense to add the missing gpio stub into
> include/linux/gpio/consumer.h instead?
I really don't like these kinds of things at all.
If GPIO is required to fully program the hardware properly,
the smc91x driver should have it as a dependency.
^ permalink raw reply
* Re: [PATCH net-next v10 0/7] cxgb4/cxgbi: misc. fixes for cxgb4i
From: David Miller @ 2014-12-12 16:22 UTC (permalink / raw)
To: kxie; +Cc: linux-scsi, netdev, hariprasad, anish, hch, James.Bottomley,
michaelc
In-Reply-To: <201412120313.sBC3DQhF028321@localhost6.localdomain6>
From: Karen Xie <kxie@chelsio.com>
Date: Thu, 11 Dec 2014 19:13:26 -0800
> This patch set fixes cxgb4i's tx credit calculation and adds handling of
> additional rx message and negative advice types. It also removes the duplicate
> code in cxgb4i to set the outgoing queues of a packet.
Series applied, thanks.
^ permalink raw reply
* Re: [PATCH net] cxgb4: Add support for QSA modules
From: David Miller @ 2014-12-12 16:24 UTC (permalink / raw)
To: hariprasad; +Cc: netdev, leedom, anish, nirranjan
In-Reply-To: <1418366277-13317-1-git-send-email-hariprasad@chelsio.com>
From: Hariprasad Shenai <hariprasad@chelsio.com>
Date: Fri, 12 Dec 2014 12:07:57 +0530
> Firmware 1.12.25.0 added support for QSA module, adding the driver code for it.
> Also fixes some ethtool get settings for other module types.
>
> Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Applied.
^ permalink raw reply
* Re: [PATCH net] net: ethernet: smsc: Allow to select SMC91X for nios2
From: David Miller @ 2014-12-12 16:26 UTC (permalink / raw)
To: tklauser; +Cc: netdev
In-Reply-To: <1418376068-28649-1-git-send-email-tklauser@distanz.ch>
From: Tobias Klauser <tklauser@distanz.ch>
Date: Fri, 12 Dec 2014 10:21:08 +0100
> This chip is present on the Nios2 Development Kit 2C35.
>
> Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Applied.
^ permalink raw reply
* Re: [PATCH net] net: ethernet: davicom: Allow to select DM9000 for nios2
From: David Miller @ 2014-12-12 16:26 UTC (permalink / raw)
To: tklauser; +Cc: netdev
In-Reply-To: <1418376071-28709-1-git-send-email-tklauser@distanz.ch>
From: Tobias Klauser <tklauser@distanz.ch>
Date: Fri, 12 Dec 2014 10:21:11 +0100
> This chip is present on older revisions of the DE2 development kit.
>
> Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Applied.
^ permalink raw reply
* Re: [PATCH net v2] net: smc91x: Fix build without gpiolib
From: Tobias Klauser @ 2014-12-12 16:27 UTC (permalink / raw)
To: David Miller; +Cc: tony, nico, netdev
In-Reply-To: <20141212.112119.1395396406830853805.davem@davemloft.net>
On 2014-12-12 at 17:21:19 +0100, David Miller <davem@davemloft.net> wrote:
> From: Tony Lindgren <tony@atomide.com>
> Date: Fri, 12 Dec 2014 08:11:00 -0800
>
> > * Tobias Klauser <tklauser@distanz.ch> [141212 08:09]:
> >> If GPIOLIB=n the following build errors occur:
> >>
> >> drivers/net/ethernet/smsc/smc91x.c: In function 'try_toggle_control_gpio':
> >> drivers/net/ethernet/smsc/smc91x.c:2204:2: error: implicit declaration of function 'devm_gpiod_get_index' [-Werror=implicit-function-declaration]
> >> drivers/net/ethernet/smsc/smc91x.c:2204:7: warning: assignment makes pointer from integer without a cast [enabled by default]
> >> drivers/net/ethernet/smsc/smc91x.c:2213:2: error: implicit declaration of function 'gpiod_direction_output' [-Werror=implicit-function-declaration]
> >> drivers/net/ethernet/smsc/smc91x.c:2216:3: error: implicit declaration of function 'devm_gpiod_put' [-Werror=implicit-function-declaration]
> >> drivers/net/ethernet/smsc/smc91x.c:2222:2: error: implicit declaration of function 'gpiod_set_value_cansleep' [-Werror=implicit-function-declaration]
> >>
> >> Since the toggling of the GPIOs is an optional feature, define
> >> try_toggle_control_gpio only if GPIOLIB is enabled.
> >
> > Oops sorry about that, I guess my systems have it all selected so
> > randconfig builds did not expose it.
> >
> > Would it make sense to add the missing gpio stub into
> > include/linux/gpio/consumer.h instead?
>
> I really don't like these kinds of things at all.
>
> If GPIO is required to fully program the hardware properly,
> the smc91x driver should have it as a dependency.
GPIO is not needed in all configurations of the chip, they are optional.
That's why I was going for the ugly #ifdef solution.
It seems that there are stubs for the functions in question in
linux/gpio/consumer.h already as Tony suggested. Thus, adding an
#include <linux/gpio/consumer.h> to the driver should be enough. I'll
send an updated patch.
^ permalink raw reply
* Re: [PATCH net v2] net: smc91x: Fix build without gpiolib
From: David Miller @ 2014-12-12 16:30 UTC (permalink / raw)
To: tklauser; +Cc: tony, nico, netdev
In-Reply-To: <20141212162746.GJ16916@distanz.ch>
From: Tobias Klauser <tklauser@distanz.ch>
Date: Fri, 12 Dec 2014 17:27:46 +0100
> GPIO is not needed in all configurations of the chip, they are optional.
> That's why I was going for the ugly #ifdef solution.
>
> It seems that there are stubs for the functions in question in
> linux/gpio/consumer.h already as Tony suggested. Thus, adding an
> #include <linux/gpio/consumer.h> to the driver should be enough. I'll
> send an updated patch.
So for the "configurations" that need it, how does the user figure out
that GPIO is needed?
In my opinion, if the code blocks enabling the configurations that
need this are enabled, so should GPIO be depended upon.
I think, at a minimum, when CONFIG_OF is enabled smsc91x should
require GPIO.
^ permalink raw reply
* Re: [PATCH net v2] net: smc91x: Fix build without gpiolib
From: Tony Lindgren @ 2014-12-12 16:34 UTC (permalink / raw)
To: David Miller; +Cc: tklauser, nico, netdev
In-Reply-To: <20141212.113020.884577050283133272.davem@davemloft.net>
* David Miller <davem@davemloft.net> [141212 08:32]:
> From: Tobias Klauser <tklauser@distanz.ch>
> Date: Fri, 12 Dec 2014 17:27:46 +0100
>
> > GPIO is not needed in all configurations of the chip, they are optional.
> > That's why I was going for the ugly #ifdef solution.
> >
> > It seems that there are stubs for the functions in question in
> > linux/gpio/consumer.h already as Tony suggested. Thus, adding an
> > #include <linux/gpio/consumer.h> to the driver should be enough. I'll
> > send an updated patch.
>
> So for the "configurations" that need it, how does the user figure out
> that GPIO is needed?
>
> In my opinion, if the code blocks enabling the configurations that
> need this are enabled, so should GPIO be depended upon.
>
> I think, at a minimum, when CONFIG_OF is enabled smsc91x should
> require GPIO.
Well in many cases these control GPIOs, or some part of them, are
hardwired with pulls. That's why they are optional.
Regards,
Tony
^ permalink raw reply
* Re: [PATCH net v2] net: smc91x: Fix build without gpiolib
From: Tobias Klauser @ 2014-12-12 16:45 UTC (permalink / raw)
To: David Miller; +Cc: tony, nico, netdev
In-Reply-To: <20141212.113020.884577050283133272.davem@davemloft.net>
On 2014-12-12 at 17:30:20 +0100, David Miller <davem@davemloft.net> wrote:
> From: Tobias Klauser <tklauser@distanz.ch>
> Date: Fri, 12 Dec 2014 17:27:46 +0100
>
> > GPIO is not needed in all configurations of the chip, they are optional.
> > That's why I was going for the ugly #ifdef solution.
> >
> > It seems that there are stubs for the functions in question in
> > linux/gpio/consumer.h already as Tony suggested. Thus, adding an
> > #include <linux/gpio/consumer.h> to the driver should be enough. I'll
> > send an updated patch.
>
> So for the "configurations" that need it, how does the user figure out
> that GPIO is needed?
Probably ionly by consulting the manual of the board usingthe chip.
Which presumably very few users will do.
> In my opinion, if the code blocks enabling the configurations that
> need this are enabled, so should GPIO be depended upon.
>
> I think, at a minimum, when CONFIG_OF is enabled smsc91x should
> require GPIO.
Agreed. This is the better solution, causing less surprises for the
user. Should this be a "select GPIOLIB if OF" then?
^ permalink raw reply
* Re: [PATCH v2 0/6] net-PPP: Deletion of a few unnecessary checks
From: SF Markus Elfring @ 2014-12-12 16:56 UTC (permalink / raw)
To: David Miller
Cc: Sergei Shtylyov, Paul Mackerras, linux-ppp, netdev, Eric Dumazet,
linux-kernel, kernel-janitors, Julia Lawall
In-Reply-To: <20141212.105106.1162681032492664308.davem@davemloft.net>
> You are asking me to invest a lot of time for a very trivial
> set of changes.
I find the proposed six update steps also trivial so far.
> Why don't you just integrate the feedback you were given and
> resubmit your patch series, just like any other developer would?
I find the requested redesign and reorganisation of involved data structures
not so easy and therefore more challenging at the moment.
Where should "the error pointers" be stored instead?
Is such a software improvement even a kind of add-on to my patch series?
Regards,
Markus
^ permalink raw reply
* Re: [bisected] xfrm: TCP connection initiating PMTU discovery stalls on v3.
From: Wolfgang Walter @ 2014-12-12 16:58 UTC (permalink / raw)
To: Eric Dumazet
Cc: Thomas Jarosch, netdev, Eric Dumazet, Herbert Xu,
Steffen Klassert
In-Reply-To: <1418238603.27198.37.camel@edumazet-glaptop2.roam.corp.google.com>
[-- Attachment #1: Type: text/plain, Size: 3614 bytes --]
Hello Eric,
my plan to use e114a710aa5058c0ba4aa1dfb105132aefeb5e04 to make it easier to apply tcp-refine-TSO-autosizing was a bad idea as that needs further patches :-). Backporting
tcp-refine-TSO-autosizing
https://patchwork.ozlabs.org/patch/418506/
seems to be the easier solution.
I tested if 3.18 and 3.17.6 are affected and they are. I applied tcp-refine-TSO-autosizing and the one-liner below and then both work fine.
Attached is
1) my backport of tcp-refine-TSO-autosizing for 3.14.26
2) my backport of tcp-refine-TSO-autosizing for 3.17.6
For 3.18 tcp-refine-TSO-autosizing just applies fine.
I tested 3.16.26 + patches, 3.17.6 + patches and 3.18.2 + patches.
And here again the one-liner:
====================
Subject: tcp: ensure that sk_gso_max_segs = 1 if GSO is disabled
sk_setup_caps() may disable GSO even if the network device
supports it. This may happens after the creation time of the
socket (e.g. triggered by xfrm, PMTU or routing changes).
This fixes hangs of local tcp connections over ipsec tunnels where
pmtu is lower than the mtu of the interface.
---
net/core/sock.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/net/core/sock.c b/net/core/sock.c
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1559,6 +1559,7 @@ void sk_setup_caps(struct sock *sk, struct dst_entry *dst)
if (sk->sk_route_caps & NETIF_F_GSO)
sk->sk_route_caps |= NETIF_F_GSO_SOFTWARE;
sk->sk_route_caps &= ~sk->sk_route_nocaps;
+ sk->sk_gso_max_segs = 1;
if (sk_can_gso(sk)) {
if (dst->header_len) {
sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
--
1.7.10.4
=======================
Do you think the backports are ok? Then Thomas my test them, too.
What should be done now?
First the one line should go into 3.19 as soon as tcp-refine-TSO-autosizing
arrived there?
If both are in Torvald's who should I ask for inclusion into stable?
Am Mittwoch, 10. Dezember 2014, 11:10:03 schrieb Eric Dumazet:
> On Wed, 2014-12-10 at 19:34 +0100, Wolfgang Walter wrote:
> > So gso is on. When the hang happens sk_setup_caps is called from
> > inet_sk_rebuild_header(). Now the path
> >
> > sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
> >
> > is taken as dst->header_len is now non zero.
> >
> > This is the reason why later calls of sk_can_gso() return false.
> >
> > I'll try to change the above patch to
> >
> > @@ -1585,6 +1585,8 @@ void sk_setup_caps(struct sock *sk, struct dst_entry
> > *dst) sk->sk_gso_max_size = dst->dev->gso_max_size;
> >
> > sk->sk_gso_max_segs = dst->dev->gso_max_segs;
> >
> > }
> >
> > }
> >
> > + if (sk_can_gso(sk)) {
> > + sk->sk_gso_max_segs = 1;
> >
> > }
> >
> > }
> > EXPORT_SYMBOL_GPL(sk_setup_caps);
> >
> > so that the case that GSO is disabled because of dst->header_len != 0 sets
> > sk_gso_max_segs, too.
>
> Sounds good, or maybe simply :
>
> diff --git a/net/core/sock.c b/net/core/sock.c
> index
> 9a56b2000c3f374fb95aedada3327447816a9512..edca31319dfee57cabe5b376505324bea
> 07f767a 100644 --- a/net/core/sock.c
> +++ b/net/core/sock.c
> @@ -1577,6 +1577,7 @@ void sk_setup_caps(struct sock *sk, struct dst_entry
> *dst) if (sk->sk_route_caps & NETIF_F_GSO)
> sk->sk_route_caps |= NETIF_F_GSO_SOFTWARE;
> sk->sk_route_caps &= ~sk->sk_route_nocaps;
> + sk->sk_gso_max_segs = 1;
> if (sk_can_gso(sk)) {
> if (dst->header_len) {
> sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
Regards and thank you very much for your help,
--
Wolfgang Walter
Studentenwerk München
Anstalt des öffentlichen Rechts
[-- Attachment #2: 0002-tcp-refine-TSO-autosizing.patch --]
[-- Type: text/x-patch, Size: 7394 bytes --]
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -134,7 +134,7 @@ struct tcp_sock {
/* inet_connection_sock has to be the first member of tcp_sock */
struct inet_connection_sock inet_conn;
u16 tcp_header_len; /* Bytes of tcp header to send */
- u16 xmit_size_goal_segs; /* Goal for segmenting output packets */
+ u16 gso_segs; /* Max number of segs per GSO packet */
/*
* Header prediction flags
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -825,47 +825,29 @@ static unsigned int tcp_xmit_size_goal(struct sock *sk, u32 mss_now,
int large_allowed)
{
struct tcp_sock *tp = tcp_sk(sk);
- u32 xmit_size_goal, old_size_goal;
-
- xmit_size_goal = mss_now;
-
- if (large_allowed && sk_can_gso(sk)) {
- u32 gso_size, hlen;
-
- /* Maybe we should/could use sk->sk_prot->max_header here ? */
- hlen = inet_csk(sk)->icsk_af_ops->net_header_len +
- inet_csk(sk)->icsk_ext_hdr_len +
- tp->tcp_header_len;
-
- /* Goal is to send at least one packet per ms,
- * not one big TSO packet every 100 ms.
- * This preserves ACK clocking and is consistent
- * with tcp_tso_should_defer() heuristic.
- */
- gso_size = sk->sk_pacing_rate / (2 * MSEC_PER_SEC);
- gso_size = max_t(u32, gso_size,
- sysctl_tcp_min_tso_segs * mss_now);
-
- xmit_size_goal = min_t(u32, gso_size,
- sk->sk_gso_max_size - 1 - hlen);
-
- xmit_size_goal = tcp_bound_to_half_wnd(tp, xmit_size_goal);
-
- /* We try hard to avoid divides here */
- old_size_goal = tp->xmit_size_goal_segs * mss_now;
-
- if (likely(old_size_goal <= xmit_size_goal &&
- old_size_goal + mss_now > xmit_size_goal)) {
- xmit_size_goal = old_size_goal;
- } else {
- tp->xmit_size_goal_segs =
- min_t(u16, xmit_size_goal / mss_now,
- sk->sk_gso_max_segs);
- xmit_size_goal = tp->xmit_size_goal_segs * mss_now;
- }
+ u32 new_size_goal, size_goal, hlen;
+
+ if (!large_allowed || !sk_can_gso(sk))
+ return mss_now;
+
+ /* Maybe we should/could use sk->sk_prot->max_header here ? */
+ hlen = inet_csk(sk)->icsk_af_ops->net_header_len +
+ inet_csk(sk)->icsk_ext_hdr_len +
+ tp->tcp_header_len;
+
+ new_size_goal = sk->sk_gso_max_size - 1 - hlen;
+ new_size_goal = tcp_bound_to_half_wnd(tp, new_size_goal);
+
+ /* We try hard to avoid divides here */
+ size_goal = tp->gso_segs * mss_now;
+ if (unlikely(new_size_goal < size_goal ||
+ new_size_goal >= size_goal + mss_now)) {
+ tp->gso_segs = min_t(u16, new_size_goal / mss_now,
+ sk->sk_gso_max_segs);
+ size_goal = tp->gso_segs * mss_now;
}
- return max(xmit_size_goal, mss_now);
+ return max(size_goal, mss_now);
}
static int tcp_send_mss(struct sock *sk, int *size_goal, int flags)
diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c
--- a/net/ipv4/tcp_cong.c
+++ b/net/ipv4/tcp_cong.c
@@ -290,7 +290,7 @@ bool tcp_is_cwnd_limited(const struct sock *sk, u32 in_flight)
left = tp->snd_cwnd - in_flight;
if (sk_can_gso(sk) &&
left * sysctl_tcp_tso_win_divisor < tp->snd_cwnd &&
- left < tp->xmit_size_goal_segs)
+ left < min_t(u32, tp->gso_segs, sk->sk_gso_max_segs))
return true;
return left <= tcp_max_tso_deferred_mss(tp);
}
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1432,6 +1432,27 @@ static bool tcp_nagle_check(bool partial, const struct tcp_sock *tp,
((nonagle & TCP_NAGLE_CORK) ||
(!nonagle && tp->packets_out && tcp_minshall_check(tp)));
}
+
+/* Return how many segs we'd like on a TSO packet,
+ * to send one TSO packet per ms
+ */
+static u32 tcp_tso_autosize(const struct sock *sk, unsigned int mss_now)
+{
+ u32 bytes, segs;
+
+ bytes = min(sk->sk_pacing_rate >> 10,
+ sk->sk_gso_max_size - 1 - MAX_TCP_HEADER);
+
+ /* Goal is to send at least one packet per ms,
+ * not one big TSO packet every 100 ms.
+ * This preserves ACK clocking and is consistent
+ * with tcp_tso_should_defer() heuristic.
+ */
+ segs = max_t(u32, bytes / mss_now, sysctl_tcp_min_tso_segs);
+
+ return min_t(u32, segs, sk->sk_gso_max_segs);
+}
+
/* Returns the portion of skb which can be sent right away */
static unsigned int tcp_mss_split_point(const struct sock *sk,
const struct sk_buff *skb,
@@ -1633,7 +1654,7 @@ static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len,
*
* This algorithm is from John Heffner.
*/
-static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb)
+static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb, u32 max_segs)
{
struct tcp_sock *tp = tcp_sk(sk);
const struct inet_connection_sock *icsk = inet_csk(sk);
@@ -1663,8 +1684,7 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb)
limit = min(send_win, cong_win);
/* If a full-sized TSO skb can be sent, do it. */
- if (limit >= min_t(unsigned int, sk->sk_gso_max_size,
- tp->xmit_size_goal_segs * tp->mss_cache))
+ if (limit >= max_segs * tp->mss_cache)
goto send_now;
/* Middle in queue won't get any more data, full sendable already? */
@@ -1857,6 +1877,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
unsigned int tso_segs, sent_pkts;
int cwnd_quota;
int result;
+ u32 max_segs;
sent_pkts = 0;
@@ -1870,6 +1891,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
}
}
+ max_segs = tcp_tso_autosize(sk, mss_now);
while ((skb = tcp_send_head(sk))) {
unsigned int limit;
@@ -1900,10 +1922,22 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
nonagle : TCP_NAGLE_PUSH))))
break;
} else {
- if (!push_one && tcp_tso_should_defer(sk, skb))
+ if (!push_one && tcp_tso_should_defer(sk, skb, max_segs))
break;
}
+ limit = mss_now;
+ if (tso_segs > 1 && !tcp_urg_mode(tp))
+ limit = tcp_mss_split_point(sk, skb, mss_now,
+ min_t(unsigned int,
+ cwnd_quota,
+ max_segs),
+ nonagle);
+
+ if (skb->len > limit &&
+ unlikely(tso_fragment(sk, skb, limit, mss_now, gfp)))
+ break;
+
/* TCP Small Queues :
* Control number of packets in qdisc/devices to two packets / or ~1 ms.
* This allows for :
@@ -1914,8 +1948,8 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
* of queued bytes to ensure line rate.
* One example is wifi aggregation (802.11 AMPDU)
*/
- limit = max_t(unsigned int, sysctl_tcp_limit_output_bytes,
- sk->sk_pacing_rate >> 10);
+ limit = max(2 * skb->truesize, sk->sk_pacing_rate >> 10);
+ limit = min_t(u32, limit, sysctl_tcp_limit_output_bytes);
if (atomic_read(&sk->sk_wmem_alloc) > limit) {
set_bit(TSQ_THROTTLED, &tp->tsq_flags);
@@ -1930,18 +1964,6 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
break;
}
- limit = mss_now;
- if (tso_segs > 1 && !tcp_urg_mode(tp))
- limit = tcp_mss_split_point(sk, skb, mss_now,
- min_t(unsigned int,
- cwnd_quota,
- sk->sk_gso_max_segs),
- nonagle);
-
- if (skb->len > limit &&
- unlikely(tso_fragment(sk, skb, limit, mss_now, gfp)))
- break;
-
TCP_SKB_CB(skb)->when = tcp_time_stamp;
if (unlikely(tcp_transmit_skb(sk, skb, 1, gfp)))
--
1.7.10.4
[-- Attachment #3: 0001-tcp-refine-TSO-autosizing.patch --]
[-- Type: text/x-patch, Size: 6898 bytes --]
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -131,7 +131,7 @@ struct tcp_sock {
/* inet_connection_sock has to be the first member of tcp_sock */
struct inet_connection_sock inet_conn;
u16 tcp_header_len; /* Bytes of tcp header to send */
- u16 xmit_size_goal_segs; /* Goal for segmenting output packets */
+ u16 gso_segs; /* Max number of segs per GSO packet */
/*
* Header prediction flags
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -836,47 +836,29 @@ static unsigned int tcp_xmit_size_goal(struct sock *sk, u32 mss_now,
int large_allowed)
{
struct tcp_sock *tp = tcp_sk(sk);
- u32 xmit_size_goal, old_size_goal;
-
- xmit_size_goal = mss_now;
-
- if (large_allowed && sk_can_gso(sk)) {
- u32 gso_size, hlen;
-
- /* Maybe we should/could use sk->sk_prot->max_header here ? */
- hlen = inet_csk(sk)->icsk_af_ops->net_header_len +
- inet_csk(sk)->icsk_ext_hdr_len +
- tp->tcp_header_len;
-
- /* Goal is to send at least one packet per ms,
- * not one big TSO packet every 100 ms.
- * This preserves ACK clocking and is consistent
- * with tcp_tso_should_defer() heuristic.
- */
- gso_size = sk->sk_pacing_rate / (2 * MSEC_PER_SEC);
- gso_size = max_t(u32, gso_size,
- sysctl_tcp_min_tso_segs * mss_now);
-
- xmit_size_goal = min_t(u32, gso_size,
- sk->sk_gso_max_size - 1 - hlen);
-
- xmit_size_goal = tcp_bound_to_half_wnd(tp, xmit_size_goal);
-
- /* We try hard to avoid divides here */
- old_size_goal = tp->xmit_size_goal_segs * mss_now;
-
- if (likely(old_size_goal <= xmit_size_goal &&
- old_size_goal + mss_now > xmit_size_goal)) {
- xmit_size_goal = old_size_goal;
- } else {
- tp->xmit_size_goal_segs =
- min_t(u16, xmit_size_goal / mss_now,
- sk->sk_gso_max_segs);
- xmit_size_goal = tp->xmit_size_goal_segs * mss_now;
- }
+ u32 new_size_goal, size_goal, hlen;
+
+ if (!large_allowed || !sk_can_gso(sk))
+ return mss_now;
+
+ /* Maybe we should/could use sk->sk_prot->max_header here ? */
+ hlen = inet_csk(sk)->icsk_af_ops->net_header_len +
+ inet_csk(sk)->icsk_ext_hdr_len +
+ tp->tcp_header_len;
+
+ new_size_goal = sk->sk_gso_max_size - 1 - hlen;
+ new_size_goal = tcp_bound_to_half_wnd(tp, new_size_goal);
+
+ /* We try hard to avoid divides here */
+ size_goal = tp->gso_segs * mss_now;
+ if (unlikely(new_size_goal < size_goal ||
+ new_size_goal >= size_goal + mss_now)) {
+ tp->gso_segs = min_t(u16, new_size_goal / mss_now,
+ sk->sk_gso_max_segs);
+ size_goal = tp->gso_segs * mss_now;
}
- return max(xmit_size_goal, mss_now);
+ return max(size_goal, mss_now);
}
static int tcp_send_mss(struct sock *sk, int *size_goal, int flags)
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1484,6 +1484,27 @@ static bool tcp_nagle_check(bool partial, const struct tcp_sock *tp,
((nonagle & TCP_NAGLE_CORK) ||
(!nonagle && tp->packets_out && tcp_minshall_check(tp)));
}
+
+/* Return how many segs we'd like on a TSO packet,
+ * to send one TSO packet per ms
+ */
+static u32 tcp_tso_autosize(const struct sock *sk, unsigned int mss_now)
+{
+ u32 bytes, segs;
+
+ bytes = min(sk->sk_pacing_rate >> 10,
+ sk->sk_gso_max_size - 1 - MAX_TCP_HEADER);
+
+ /* Goal is to send at least one packet per ms,
+ * not one big TSO packet every 100 ms.
+ * This preserves ACK clocking and is consistent
+ * with tcp_tso_should_defer() heuristic.
+ */
+ segs = max_t(u32, bytes / mss_now, sysctl_tcp_min_tso_segs);
+
+ return min_t(u32, segs, sk->sk_gso_max_segs);
+}
+
/* Returns the portion of skb which can be sent right away */
static unsigned int tcp_mss_split_point(const struct sock *sk,
const struct sk_buff *skb,
@@ -1687,7 +1708,7 @@ static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len,
* This algorithm is from John Heffner.
*/
static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb,
- bool *is_cwnd_limited)
+ bool *is_cwnd_limited, u32 max_segs)
{
struct tcp_sock *tp = tcp_sk(sk);
const struct inet_connection_sock *icsk = inet_csk(sk);
@@ -1717,8 +1738,7 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb,
limit = min(send_win, cong_win);
/* If a full-sized TSO skb can be sent, do it. */
- if (limit >= min_t(unsigned int, sk->sk_gso_max_size,
- tp->xmit_size_goal_segs * tp->mss_cache))
+ if (limit >= max_segs * tp->mss_cache)
goto send_now;
/* Middle in queue won't get any more data, full sendable already? */
@@ -1915,6 +1935,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
int cwnd_quota;
int result;
bool is_cwnd_limited = false;
+ u32 max_segs;
sent_pkts = 0;
@@ -1928,6 +1949,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
}
}
+ max_segs = tcp_tso_autosize(sk, mss_now);
while ((skb = tcp_send_head(sk))) {
unsigned int limit;
@@ -1960,10 +1982,23 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
break;
} else {
if (!push_one &&
- tcp_tso_should_defer(sk, skb, &is_cwnd_limited))
+ tcp_tso_should_defer(sk, skb, &is_cwnd_limited,
+ max_segs))
break;
}
+ limit = mss_now;
+ if (tso_segs > 1 && !tcp_urg_mode(tp))
+ limit = tcp_mss_split_point(sk, skb, mss_now,
+ min_t(unsigned int,
+ cwnd_quota,
+ max_segs),
+ nonagle);
+
+ if (skb->len > limit &&
+ unlikely(tso_fragment(sk, skb, limit, mss_now, gfp)))
+ break;
+
/* TCP Small Queues :
* Control number of packets in qdisc/devices to two packets / or ~1 ms.
* This allows for :
@@ -1974,8 +2009,8 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
* of queued bytes to ensure line rate.
* One example is wifi aggregation (802.11 AMPDU)
*/
- limit = max_t(unsigned int, sysctl_tcp_limit_output_bytes,
- sk->sk_pacing_rate >> 10);
+ limit = max(2 * skb->truesize, sk->sk_pacing_rate >> 10);
+ limit = min_t(u32, limit, sysctl_tcp_limit_output_bytes);
if (atomic_read(&sk->sk_wmem_alloc) > limit) {
set_bit(TSQ_THROTTLED, &tp->tsq_flags);
@@ -1988,18 +2023,6 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
break;
}
- limit = mss_now;
- if (tso_segs > 1 && !tcp_urg_mode(tp))
- limit = tcp_mss_split_point(sk, skb, mss_now,
- min_t(unsigned int,
- cwnd_quota,
- sk->sk_gso_max_segs),
- nonagle);
-
- if (skb->len > limit &&
- unlikely(tso_fragment(sk, skb, limit, mss_now, gfp)))
- break;
-
TCP_SKB_CB(skb)->when = tcp_time_stamp;
if (unlikely(tcp_transmit_skb(sk, skb, 1, gfp)))
--
1.7.10.4
^ permalink raw reply
* Re: [PATCH net v2] net: smc91x: Fix build without gpiolib
From: David Miller @ 2014-12-12 16:58 UTC (permalink / raw)
To: tklauser; +Cc: tony, nico, netdev
In-Reply-To: <20141212164528.GK16916@distanz.ch>
From: Tobias Klauser <tklauser@distanz.ch>
Date: Fri, 12 Dec 2014 17:45:29 +0100
> On 2014-12-12 at 17:30:20 +0100, David Miller <davem@davemloft.net> wrote:
>> In my opinion, if the code blocks enabling the configurations that
>> need this are enabled, so should GPIO be depended upon.
>>
>> I think, at a minimum, when CONFIG_OF is enabled smsc91x should
>> require GPIO.
>
> Agreed. This is the better solution, causing less surprises for the
> user. Should this be a "select GPIOLIB if OF" then?
If GPIO is a child node in the Kconfig hierarchy (ie. has no
dependencies of it's own), then yes. Otherwise, a depends
will need to be used, because select does not recursively
trigger a select'd nodes dependencies.
^ permalink raw reply
* Re: [PATCH v2 0/6] net-PPP: Deletion of a few unnecessary checks
From: David Miller @ 2014-12-12 16:59 UTC (permalink / raw)
To: elfring
Cc: sergei.shtylyov, paulus, linux-ppp, netdev, eric.dumazet,
linux-kernel, kernel-janitors, julia.lawall
In-Reply-To: <548B1E44.6050005@users.sourceforge.net>
From: SF Markus Elfring <elfring@users.sourceforge.net>
Date: Fri, 12 Dec 2014 17:56:36 +0100
> Where should "the error pointers" be stored instead?
A local variable, before you assign it into the datastructure.
^ permalink raw reply
* RE: [net PATCH] fib_trie: Fix trie balancing issue if new node pushes down existing node
From: David Laight @ 2014-12-12 17:05 UTC (permalink / raw)
To: 'Alexander Duyck', David Miller,
alexander.h.duyck@redhat.com
Cc: netdev@vger.kernel.org
In-Reply-To: <548B132B.5000606@gmail.com>
From: Alexander Duyck
...
> >> One has to be mindful with this code that what it's doing now might
> >> be intentional. For example, it might be doing things this way
> >> on purpose in order to minimize rebalancing during route flaps.
> >>
> >> Barring anything like that, I think your change is fine.
> > Alex, in case it isn't clear, I'm hoping that you might have some
> > thoughts on this aspect of your changes. :)
>
> Route flaps should at most trigger an inflate without a deflate due to
> the way the resize logic works. If this occurs often it would be useful
> since then there would already be space in the tree for the next time
> around.
One way to deal with 'flapping' is to leave deleted items in the tree
(marked so they won't be used).
Deleted entries do need to be garbage collected at some point.
Doing so during insert is more than adequate, and possibly only for
nodes that are traversed while processing the insert itself
(and maybe only is the ratio of valid to invalid entries is low).
This might also mean that you don't need to actively run any timeouts.
Timed out entries just stay put - marked invalid by their timestamp.
David
^ permalink raw reply
* Re: [PATCH v2 0/6] net-PPP: Deletion of a few unnecessary checks
From: SF Markus Elfring @ 2014-12-12 17:22 UTC (permalink / raw)
To: David Miller
Cc: Sergei Shtylyov, Paul Mackerras, linux-ppp, netdev, Eric Dumazet,
linux-kernel, kernel-janitors, Julia Lawall
In-Reply-To: <20141212.115922.687789059853236747.davem@davemloft.net>
>> Where should "the error pointers" be stored instead?
> A local variable, before you assign it into the datastructure.
Will it be acceptable for you that anyone (or even me) will introduce
such a change with a seventh (and eventually eighth) update step here?
Do you want any other sequence for source code preparation of
the requested software improvement?
Regards,
Markus
^ permalink raw reply
* Re: [bisected] xfrm: TCP connection initiating PMTU discovery stalls on v3.
From: Eric Dumazet @ 2014-12-12 17:27 UTC (permalink / raw)
To: Wolfgang Walter
Cc: Thomas Jarosch, netdev, Eric Dumazet, Herbert Xu,
Steffen Klassert
In-Reply-To: <2630078.LdJXSsPB40@h2o.as.studentenwerk.mhn.de>
On Fri, 2014-12-12 at 17:58 +0100, Wolfgang Walter wrote:
> This fixes hangs of local tcp connections over ipsec tunnels where
> pmtu is lower than the mtu of the interface.
Again this is a work around, the one liner patch is not a clean patch.
Something is wrong, we should fix GSO path instead.
Why ? Because we can queue a GSO packet in a qdisc, then later call
sk_setup_caps() (too late)
^ permalink raw reply
* Re: am335x: cpsw: interrupt failure
From: Felipe Balbi @ 2014-12-12 17:32 UTC (permalink / raw)
To: Yegor Yefremov; +Cc: Felipe Balbi, netdev, N, Mugunthan V
In-Reply-To: <CAGm1_kuaqXrU4zhNUZeY85+OeXHApSUDjJdsXcyBdvHTNhLCmQ@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 556 bytes --]
Hi,
On Fri, Dec 12, 2014 at 01:00:51PM +0100, Yegor Yefremov wrote:
> U-Boot version: 2014.07
> Kernel config is omap2plus with enabled USB
>
> # cat /proc/version
> Linux version 3.18.0 (user@user-VirtualBox) (gcc version 4.8.3
> 20140320 (prerelease) (Sourcery CodeBench Lite 2014.05-29) ) #6 SMP
> Mon Dec 8 22:47:43 CET 2014
Wasn't GCC 4.8.x total crap for building ARM kernels ? IIRC it was even
blacklisted. Can you try with 4.9.x just to make sure ?
I'll try to run your test case here once i'm back from vacations.
--
balbi
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]
^ permalink raw reply
* About ARP state change from Delay to Stale
From: Sriram Raghunathan @ 2014-12-12 17:37 UTC (permalink / raw)
To: netdev, linux-kernel
Hi,
I'd a question about ARP entry state change from DELAY to STALE. I'm
referring to 3.10 Master kernel, where this particular state change
is not notified on neigh_update. In case, on creating a patch for the
same, would it affect the kernel behavior ?
Could you please share as why this state change is not notified
right now?
Sriram
^ permalink raw reply
* Re: [PATCH v2 0/6] net-PPP: Deletion of a few unnecessary checks
From: Julia Lawall @ 2014-12-12 18:46 UTC (permalink / raw)
To: SF Markus Elfring
Cc: David Miller, Sergei Shtylyov, Paul Mackerras, linux-ppp, netdev,
Eric Dumazet, linux-kernel, kernel-janitors, Julia Lawall
In-Reply-To: <548B0A29.8050503@users.sourceforge.net>
> I hope that a bit more constructive suggestions will be contributed by
> involved
> software developers around the affected source code. Now it seems
> that a small code clean-up becomes a more challenging development task.
This is often the case. Doing something half way is not useful.
julia
^ permalink raw reply
* Re: [PATCH v2 0/6] net-PPP: Deletion of a few unnecessary checks
From: Eric Dumazet @ 2014-12-12 19:08 UTC (permalink / raw)
To: SF Markus Elfring
Cc: David Miller, Sergei Shtylyov, Paul Mackerras, linux-ppp, netdev,
linux-kernel, kernel-janitors, Julia Lawall
In-Reply-To: <548B2468.5050402@users.sourceforge.net>
On Fri, 2014-12-12 at 18:22 +0100, SF Markus Elfring wrote:
> Will it be acceptable for you that anyone (or even me) will introduce
> such a change with a seventh (and eventually eighth) update step here?
> Do you want any other sequence for source code preparation of
> the requested software improvement?
The thing is : We are in the merge window, tracking bugs added in latest
dev cycle.
Having to deal with patches like yours is adding pressure
on the maintainer (and other developers) at the wrong time.
^ permalink raw reply
* Re: am335x: cpsw: interrupt failure
From: Yegor Yefremov @ 2014-12-12 19:19 UTC (permalink / raw)
To: Felipe Balbi; +Cc: netdev, N, Mugunthan V
In-Reply-To: <20141212173210.GI7549@saruman>
On Fri, Dec 12, 2014 at 6:32 PM, Felipe Balbi <balbi@ti.com> wrote:
> Hi,
>
> On Fri, Dec 12, 2014 at 01:00:51PM +0100, Yegor Yefremov wrote:
>> U-Boot version: 2014.07
>> Kernel config is omap2plus with enabled USB
>>
>> # cat /proc/version
>> Linux version 3.18.0 (user@user-VirtualBox) (gcc version 4.8.3
>> 20140320 (prerelease) (Sourcery CodeBench Lite 2014.05-29) ) #6 SMP
>> Mon Dec 8 22:47:43 CET 2014
>
> Wasn't GCC 4.8.x total crap for building ARM kernels ? IIRC it was even
> blacklisted. Can you try with 4.9.x just to make sure ?
Will do.
> I'll try to run your test case here once i'm back from vacations.
Wish you good vacation!
Yegor
^ permalink raw reply
* Re: [PATCH iproute2 v3] ip: Simplify executing ip cmd within network ns
From: vadim4j @ 2014-12-12 19:26 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: Vadim Kochan, netdev
In-Reply-To: <20141212060436.GA876@angus-think.lan>
On Fri, Dec 12, 2014 at 08:04:36AM +0200, vadim4j@gmail.com wrote:
> On Thu, Dec 11, 2014 at 05:26:40PM -0800, Stephen Hemminger wrote:
> > On Fri, 12 Dec 2014 00:32:51 +0200
> > Vadim Kochan <vadim4j@gmail.com> wrote:
> >
> > > +
> > > +#define NETNS_RUN_DIR "/var/run/netns"
> > > +#define NETNS_ETC_DIR "/etc/netns"
> > > +
> > > +#ifndef CLONE_NEWNET
> > > +#define CLONE_NEWNET 0x40000000 /* New network namespace (lo, device, names sockets, etc) */
> > > +#endif
> > > +
> > > +#ifndef MNT_DETACH
> > > +#define MNT_DETACH 0x00000002 /* Just detach from the tree */
> > > +#endif /* MNT_DETACH */
> > > +
> > > +/* sys/mount.h may be out too old to have these */
> > > +#ifndef MS_REC
> > > +#define MS_REC 16384
> > > +#endif
> > > +
> > > +#ifndef MS_SLAVE
> > > +#define MS_SLAVE (1 << 19)
> > > +#endif
> > > +
> > > +#ifndef MS_SHARED
> > > +#define MS_SHARED (1 << 20)
> > > +#endif
> > > +
> > > +extern int netns_switch(char *netns);
> > > +
> >
> > Maybe it would be cleaner to introduce a new file netns.h
> > with this and the setnetns syscall wrapper
> >
> And MS_*/MNT_* defines to new mount.h ?
>
> Thanks,
>
I think it can go all together (netns + mount things) to the new
include/namespace.h header, what do you think ?
And what about netns_switch func - let it be in utils.c/utils.h or
add new lib/namespace.c which will be linked to the libutil ?
Thanks,
^ 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