thanks. will be applied to next series.
On Mon, Jun 13, 2016 at 3:54 PM, Arnd Bergmann wrote:
> The new rxe infinband driver passes around pointers that have been
> converted to 64-bit integers. This is valid, but causes compile-time
> warnings on all 32-bit architectures:
>
>
thanks. will be applied to next series
On Mon, Jun 13, 2016 at 4:18 PM, Leon Romanovsky wrote:
> On Mon, Jun 13, 2016 at 02:54:53PM +0200, Arnd Bergmann wrote:
>> The rxe driver fails to build on 32-bit because of a 64-bit division:
>>
>> In function `rxe_qp_from_attr':
>>
thanks. will be applied to next series
On Mon, Jun 13, 2016 at 3:54 PM, Arnd Bergmann wrote:
> The newly added rxe driver links against the UDP tunneling code,
> which causes build errors when CONFIG_UDP_TUNNEL is disabled:
>
> ERROR: "setup_udp_tunnel_sock"
Hi,
As reported last week by Or Gerlitz and confirmed by me, there is a kernel
crash
when trying to unenslave all ib slaves (which leads to bonding master
destruction).
I also found that it happens with Ethernet slaves if following the steps:
1. unenslaving all slaves via sysfs
2. deleting
mess with rtnl, and the
second correctly acquires and releases rtnl.
Bug reported by Moni Shoua [EMAIL PROTECTED]
I ran some shallow tests and it seems that the patch fixes the problem.
thanks
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message
Or Gerlitz wrote:
Jay, Moni
I did some tests with 2.6.24-rc1 and the first patch to bonding that Jay
sent last night to netdev. Basic operation and fail over work fine.
However, I see some crashes which are somehow related to destroying the
bond when the slaves are ipoib ones, I don't see
Same thing happens with Ethernet slave - Intel 82546GB Gigabit Ethernet
Controller (rev 03)
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Moni Shoua wrote:
Basically, what I see is that after a while commands like ifconfig or ip
stucks.
I only use sysfs to configure bonding (which also stucks after a while).
Maybe some extra information might help
This is what happens in the kernel while the command 'ip' is stucked in user
Jay Vosburgh wrote:
Moni Shoua [EMAIL PROTECTED] wrote:
Basically, what I see is that after a while commands like ifconfig or ip
stucks.
I only use sysfs to configure bonding (which also stucks after a while).
I've fooled with setting various things in bonding in the
current
Hi,
I've been doing some tests for bonding of 2.6.24-rc1 and noticed some problems.
My first goal was to see how bonding works with IPoIB slaves but I also tried it
with Ethernet.
Basically, what I see is that after a while commands like ifconfig or ip stucks.
I only use sysfs to configure
.
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Jay Vosburgh [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c | 11 +--
drivers/net/bonding/bonding.h |4 ++--
2 files changed, 7 insertions(+), 8 deletions(-)
diff --git a/drivers/net/bonding/bond_main.c b
This is the 7th version of this patch series. See link to V6 below.
Changes from the previous version
-
* Some patches required modifications to remove offsets so they can be applied
with git-apply
* Patch #3 was first modified by Jay and later by me to make it
.
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
Acked-by: Roland Dreier [EMAIL PROTECTED]
---
drivers/infiniband/ulp/ipoib/ipoib.h |4 +++-
drivers/infiniband/ulp/ipoib/ipoib_main.c | 24 +++-
drivers
Moni Shoua wrote:
This is the 7th version of this patch series. See link to V6 below.
I forgot to mention that the patches are relative to
jgarzik/netdev-2.6.git#master.
I couldn't compile the 2.6.24 or the upstream branches so I used master branch
to test the fixes.
-
To unsubscribe from
ipoib_neigh.
This patch detects and prevents this from happenning.
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
Acked-by: Roland Dreier [EMAIL PROTECTED]
---
drivers/infiniband/ulp/ipoib/ipoib_main.c |5 +++--
1 files changed, 3 insertions(+), 2
where for multicast joins taking place after the enslavement another
ip_xxx_mc_map()
is called (eg ip_ib_mc_map() when the bond type is ARPHRD_INFINIBAND)
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
Acked-by: Jay Vosburgh [EMAIL PROTECTED
bonding sometimes uses Ethernet constants (such as MTU and address length) which
are not good when it enslaves non Ethernet devices (such as InfiniBand).
Signed-off-by: Moni Shoua monis at voltaire.com
Acked-by: Jay Vosburgh [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c |3
Delay sending a gratuitous_arp when LINK_STATE_LINKWATCH_PENDING bit
in dev-state field is on. This improves the chances for the arp packet to
be transmitted.
Signed-off-by: Moni Shoua monis at voltaire.com
Acked-by: Jay Vosburgh [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c | 24
slave_cnt is zero
ensures that these functions be used anymore.
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/net/bonding/bond_main.c | 37 -
drivers/net/bonding/bond_sysfs.c |9 +
drivers/net/bonding/bonding.h|3 +++
3 files
Roland Dreier wrote:
It happens only when ib interfaces are slaves of a bonding device.
I thought before that the stuck is in napi_disable() but it's almost right.
I put prints before and after call to napi_disable and see that it is
called twice.
I'll try to investigate in this
I will be near my lab only tomorrow...
I will check this and let you know.
On 10/11/07, Roland Dreier [EMAIL PROTECTED] wrote:
It happens only when ib interfaces are slaves of a bonding device.
I thought before that the stuck is in napi_disable() but it's almost right.
I put prints
Roland Dreier wrote:
I also ran a test for the code in the branch of 2.6.24 and found a problem.
I see that ifconfig down doesn't return (for IPoIB interfaces) and it's
stuck in napi_disable() in the kernel (any idea why?)
For what it's worth, I took the upstream 2.6.23 git tree and
Jay Vosburgh wrote:
David Miller [EMAIL PROTECTED] wrote:
From: Jeff Garzik [EMAIL PROTECTED]
Date: Tue, 09 Oct 2007 20:56:35 -0400
Jeff Garzik wrote:
applied patches 1-9
the only thing that was a hiccup during submission is that your email
subject lines did not contain a notion of
Jay Vosburgh wrote:
Jeff Garzik [EMAIL PROTECTED] wrote:
Moni Shoua wrote:
Jay Vosburgh wrote:
ACK patches 3 - 9.
Roland, are you comfortable with the IB changes in patches 1 and 2?
Jeff, when Roland acks patches 1 and 2, please apply all 9.
-J
Hi Jeff,
Roland acked
Jay Vosburgh wrote:
ACK patches 3 - 9.
Roland, are you comfortable with the IB changes in patches 1 and 2?
Jeff, when Roland acks patches 1 and 2, please apply all 9.
-J
Hi Jeff,
Roland acked the IPoIB patches. If you haven't done so already can you please
apply them?
Jay,
I think that all comments to the patches were discussed and handled.
If you agree, can you please push then to the networking tree so they will be
merged into 2.6.24?
This includes the IPoIB patches (agreed with Roland).
Note that there are *no* patches to net/core (like in V5).
thanks
Roland Dreier wrote:
The action in bonding to a detach of slave is to unregister the master
(see patch 10).
This can't be done from the context of unregister_netdevice itself (it is
protected by rtnl_lock).
I'm confused. Your patch has:
+ ipoib_slave_detach(cpriv-dev);
This patch series is the sixth version (see below link to V5) of the
suggested changes to the bonding driver so it would be able to support
non ARPHRD_ETHER netdevices for its High-Availability (active-backup) mode.
Patches 1-8 were originally submitted in V5 and patch 9 is an addition by Jay.
.
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/infiniband/ulp/ipoib/ipoib.h |4 +++-
drivers/infiniband/ulp/ipoib/ipoib_main.c | 24 +++-
drivers/infiniband/ulp/ipoib/ipoib_multicast.c |3
ipoib_neigh.
This patch detects and prevents this from happenning.
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/infiniband/ulp/ipoib/ipoib_main.c |5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
Index: net-2.6/drivers
) of the port this
IPoIB device is bounded to. The QP is a resource created by the IB HW and the
GID is an identifier burned into the HCA (i have omitted here some details which
are not important for the bonding RFC).
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz
(this is already done by the bonding code).
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/net/bonding/bond_main.c | 87 +++-
drivers/net/bonding/bonding.h |1
2 files changed, 60 insertions(+), 28
where for multicast joins taking place after the enslavement another
ip_xxx_mc_map()
is called (eg ip_ib_mc_map() when the bond type is ARPHRD_INFINIBAND)
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/net/bonding/bond_main.c |5
bonding sometimes uses Ethernet constants (such as MTU and address length) which
are not good when it enslaves non Ethernet devices (such as InfiniBand).
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/net/bonding/bond_main.c |3 ++-
drivers/net/bonding/bond_sysfs.c | 10
Delay sending a gratuitous_arp when LINK_STATE_LINKWATCH_PENDING bit
in dev-state field is on. This improves the chances for the arp packet to
be transmitted.
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/net/bonding/bond_main.c | 24 +---
drivers/net/bonding
slave_cnt is zero
ensures that these functions be used anymore.
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/net/bonding/bond_main.c | 37 +
drivers/net/bonding/bond_sysfs.c |9 +
drivers/net/bonding/bonding.h|3 +++
3 files
Update the don't change MAC of slaves functionality added in
previous changes to be a generic option, rather than something tied to IB
devices, as it's occasionally useful for regular ethernet devices as well.
Adds fail_over_mac option (which is automatically enabled for IB
slaves),
Roland Dreier wrote:
+ ipoib_slave_detach(cpriv-dev);
unregister_netdev(cpriv-dev);
Maybe you already answered this before, but I'm still not clear why
this notifier call can't just be added to the start of
unregister_netdevice(), so we can avoid having driver
A slave of a bonding master that wants to send a notification before
going down should call netdev_slave_detach(). The handling of this notification
will be done outside the context of unregister_netdevice() which is sometimes
necessary, as with IPoIB slave for example.
Signed-off-by: Moni Shoua
master, let the master know that the IPoIB device is
about to unregister (but before calling unregister).
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/infiniband/ulp/ipoib/ipoib.h |7 +++
drivers/infiniband/ulp/ipoib/ipoib_main.c |3 +++
drivers/infiniband/ulp
.
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/infiniband/ulp/ipoib/ipoib.h |4 +++-
drivers/infiniband/ulp/ipoib/ipoib_main.c | 24 +++-
drivers/infiniband/ulp/ipoib/ipoib_multicast.c |3
ipoib_neigh.
This patch detects and prevents this from happenning.
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/infiniband/ulp/ipoib/ipoib_main.c |5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
Index: net-2.6/drivers
) of the port this
IPoIB device is bounded to. The QP is a resource created by the IB HW and the
GID is an identifier burned into the HCA (i have omitted here some details which
are not important for the bonding RFC).
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz
(this is already done by the bonding code).
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/net/bonding/bond_main.c | 87 +++-
drivers/net/bonding/bonding.h |1
2 files changed, 60 insertions(+), 28
where for multicast joins taking place after the enslavement another
ip_xxx_mc_map()
is called (eg ip_ib_mc_map() when the bond type is ARPHRD_INFINIBAND)
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz at voltaire.com
---
drivers/net/bonding/bond_main.c |5
This patch series is the fifth version (see below link to V4) of the
suggested changes to the bonding driver so it would be able to support
non ARPHRD_ETHER netdevices for its High-Availability (active-backup) mode.
Patches 1-10 were originally submitted in V4 and patch 11 is an addition by
bonding sometimes uses Ethernet constants (such as MTU and address length) which
are not good when it enslaves non Ethernet devices (such as InfiniBand).
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/net/bonding/bond_main.c |3 ++-
drivers/net/bonding/bond_sysfs.c | 19
Delay sending a gratuitous_arp when LINK_STATE_LINKWATCH_PENDING bit
in dev-state field is on. This improves the chances for the arp packet to
be transmitted.
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/net/bonding/bond_main.c | 24 +---
drivers/net/bonding
slave_cnt is zero
ensures that these functions be used anymore.
Signed-off-by: Moni Shoua monis at voltaire.com
---
drivers/net/bonding/bond_main.c | 45 +++-
drivers/net/bonding/bonding.h |3 ++
2 files changed, 47 insertions(+), 1 deletion(-)
Index: net
) of the port this
IPoIB device is bounded to. The QP is a resource created by the IB HW and the
GID is an identifier burned into the HCA (i have omitted here some details which
are not important for the bonding RFC).
Signed-off-by: Moni Shoua monis at voltaire.com
Signed-off-by: Or Gerlitz ogerlitz
Update the don't change MAC of slaves functionality added in
previous changes to be a generic option, rather than something tied to IB
devices, as it's occasionally useful for regular ethernet devices as well.
Adds fail_over_mac option (which is automatically enabled for IB
slaves),
Roland, Jay,
Thanks a lot for the comments.
I'd like to summarize the points raised so far
1. Reduce the indentation in patch #4 (Roland)
I will resend
2. Remove the if (n-dev-flags IFF_MASTER) from patch #3 (Roland)
I will resend
3. Consider making ipoib_slave_detach()
Hi all,
This patch series is a bit neglected.
Since our goal is to have bonding support for IPoIB in kernel 2.6.24 it is
very important for us to get comments soon.
We would appreciate if you take some time to look at this and help us push this
code upstream.
thanks
MoniS
-
To
Jay Vosburgh wrote:
Moni Shoua [EMAIL PROTECTED] wrote:
Jay Vosburgh wrote:
Moni Shoua [EMAIL PROTECTED] wrote:
When bonding enslaves non Ethernet devices it takes pointers to functions
in the module that owns the slaves. In this case it becomes unsafe
to keep the bonding master
Jay Vosburgh wrote:
Moni Shoua [EMAIL PROTECTED] wrote:
bonding sometimes uses Ethernet constants (such as MTU and address length)
which
are not good when it enslaves non Ethernet devices (such as InfiniBand).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
drivers/net/bonding
Jay Vosburgh wrote:
Moni Shoua [EMAIL PROTECTED] wrote:
When bonding enslaves non Ethernet devices it takes pointers to functions
in the module that owns the slaves. In this case it becomes unsafe
to keep the bonding master registered after last slave was unenslaved
because we don't know
This patch series is the fourth version (see below link to V3) of the
suggested changes to the bonding driver so it would be able to support
non ARPHRD_ETHER netdevices for its High-Availability (active-backup) mode.
The motivation is to enable the bonding driver on its HA mode to work with
that wish to inform the bonding master
about coming detachment.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
include/linux/if.h |1 +
net/core/dev.c |1 +
2 files changed, 2 insertions(+)
Index: net-2.6/net/core/dev.c
master, let the master know that the IPoIB device is
about to unregister (but before calling unregister).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
drivers/infiniband/ulp/ipoib/ipoib_main.c | 15 +++
1 files changed, 15 insertions(+)
Index: net-2.6/drivers/infiniband/ulp
.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/infiniband/ulp/ipoib/ipoib.h |4 +++-
drivers/infiniband/ulp/ipoib/ipoib_main.c | 17 +++--
drivers/infiniband/ulp/ipoib/ipoib_multicast.c |3 ++-
3 files changed
ipoib_neigh.
This patch detects and prevents this from happenning.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/infiniband/ulp/ipoib/ipoib_main.c |5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
Index: net-2.6/drivers/infiniband/ulp
) of the port this
IPoIB device is bounded to. The QP is a resource created by the IB HW and the
GID is an identifier burned into the HCA (i have omitted here some details which
are not important for the bonding RFC).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL
(this is already done by the bonding code).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c | 87 +++-
drivers/net/bonding/bonding.h |1
2 files changed, 60 insertions(+), 28 deletions
where for multicast joins taking place after the enslavement another
ip_xxx_mc_map()
is called (eg ip_ib_mc_map() when the bond type is ARPHRD_INFINIBAND)
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c |5
bonding sometimes uses Ethernet constants (such as MTU and address length) which
are not good when it enslaves non Ethernet devices (such as InfiniBand).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c |3 ++-
drivers/net/bonding/bond_sysfs.c | 19
Delay sending a gratuitous_arp when LINK_STATE_LINKWATCH_PENDING bit
in dev-state field is on. This improves the chances for the arp packet to
be transmitted.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c | 24 +---
drivers/net/bonding
slave_cnt is zero
ensures that these functions be used anymore.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c | 45 +++-
drivers/net/bonding/bonding.h |3 ++
2 files changed, 47 insertions(+), 1 deletion(-)
Index: net-2.6
It's always wrong to copy symbols from another module without
referencing it.
Michael,
It seems like the preferred approach is to prevent ib_ipoib from being
unloaded while bonding is on top it, right?
It seems like it would handle all safety issues (not just neigh cleanup).
-
To
Roland Dreier wrote:
1. When bonding enslaves an IPoIB device the bonding neighbor holds a
reference to a cleanup function in the IPoIB drives. This makes it unsafe
to
unload the IPoIB module if there are bonding neighbors in the air. So, to
avoid this race one must unload bonding
.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/infiniband/ulp/ipoib/ipoib.h |4 +++-
drivers/infiniband/ulp/ipoib/ipoib_main.c | 17 +++--
drivers/infiniband/ulp/ipoib/ipoib_multicast.c |2 +-
3 files changed
ipoib_neigh.
This patch detects and prevents this from happenning.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/infiniband/ulp/ipoib/ipoib_main.c |5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
Index: net-2.6/drivers/infiniband/ulp
) of the port this
IPoIB device is bounded to. The QP is a resource created by the IB HW and the
GID is an identifier burned into the HCA (i have omitted here some details which
are not important for the bonding RFC).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL
(this is already done by the bonding code).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c | 88 +++-
drivers/net/bonding/bonding.h |1
2 files changed, 61 insertions(+), 28 deletions
where for multicast joins taking place after the enslavement another
ip_xxx_mc_map()
is called (eg ip_ib_mc_map() when the bond type is ARPHRD_INFINIBAND)
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
Signed-off-by: Or Gerlitz [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c |4
This patch series is the third version (see below link to V2) of the
suggested changes to the bonding driver so it would be able to support
non ARPHRD_ETHER netdevices for its High-Availability (active-backup) mode.
The motivation is to enable the bonding driver on its HA mode to work with
bonding sometimes uses Ethernet constants (such as MTU and address length) which
are not good when it enslaves non Ethernet devices (such as InfiniBand).
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c |2 +-
drivers/net/bonding/bond_sysfs.c | 10
Delay sending a gratuitous_arp when LINK_STATE_LINKWATCH_PENDING bit
in dev-state field is on. This improves the chances for the arp packet to
be transmitted.
Signed-off-by: Moni Shoua [EMAIL PROTECTED]
---
drivers/net/bonding/bond_main.c | 25 +
drivers/net/bonding
77 matches
Mail list logo