definitions needed by
the application via linux/rds.h.
Patch 2 provides the setsockopt support, and Patch 3 provides the
getsockopt support.
Sowmini Varadhan (3):
Declare SO_RDS_TRANSPORT and RDS_TRANS_* constants in
uapi/linux/rds.h
Add setsockopt support for SO_RDS_TRANSPORT
Add setsockopt
The currently attached transport for a PF_RDS socket may be obtained
from user space by invoking getsockopt(2) using the SO_RDS_TRANSPORT
option at the SOL_RDS level. The integer optval returned will be one
of the RDS_TRANS_* constants defined in linux/rds.h.
Signed-off-by: Sowmini Varadhan
., RDS_TRANS_TCP. This commit exports
the constant values need by such applications via linux/rds.h
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
include/uapi/linux/rds.h | 10 ++
net/rds/rds.h|5 -
2 files changed, 10 insertions(+), 5 deletions(-)
diff --git
be specified before invoking bind(2) on the socket, and may only
be used once on the socket. An attempt to set the option on a bound
socket, or to invoke the option after a successful SO_RDS_TRANSPORT
attachment, will return EOPNOTSUPP.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
net
Dumazet eduma...@google.com
Cc: Eric W. Biederman ebied...@xmission.com
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
v2: pulled patch #3 out of the RFC patch-set for RDS-TCP netns fixes;
Added Fixes, Acked-by, Cc fields based on mailing list feedback
from Eric Dumazet.
net
,
patch 3 broken off and submitted separately.
Sowmini Varadhan (2):
Make RDS-TCP work correctly when it is set up in a netns other than
init_net
Support multiple RDS-TCP listen endpoints, one per netns.
net/rds/bind.c|3 +-
net/rds/connection.c | 16 +++--
net/rds
indicating
that the netns is getting deleted.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
v2: net_device notifier for synchronous cleanup of sockets.
net/rds/tcp.c | 163 -
net/rds/tcp.h |7 ++-
net/rds
Open the sockets calling sock_create_kern() with the correct struct net
pointer, and use that struct net pointer when verifying the
address passed to rds_bind().
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
v2: David Ahern comments.
net/rds/bind.c|3 ++-
net/rds
indicating
that the netns is getting deleted.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
v2: net_device notifier for synchronous cleanup of sockets.
v3: Cong Wang comments
net/rds/tcp.c | 161 -
net/rds/tcp.h
,
patch 3 broken off and submitted separately.
v3: Cong Wang review comments.
Sowmini Varadhan (2):
Make RDS-TCP work correctly when it is set up in a netns other than
init_net
Support multiple RDS-TCP listen endpoints, one per netns.
net/rds/bind.c|3 +-
net/rds
Open the sockets calling sock_create_kern() with the correct struct net
pointer, and use that struct net pointer when verifying the
address passed to rds_bind().
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
v2: David Ahern comments.
net/rds/bind.c|3 ++-
net/rds
On (07/30/15 11:03), David Ahern wrote:
+write_pnet(conn-c_net, net);
these are typically in wrappers like sock_net and sock_net_set
:
+conn = rds_conn_create(init_net, dp-dp_daddr, dp-dp_saddr,
+ rds_ib_transport, GFP_KERNEL);
I forget what connection
that cleanup_net()
will eventually complete, allowing the pernet -exit to be invoked.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
net/rds/tcp.c | 112 ++--
net/rds/tcp.h |7 ++-
net/rds/tcp_connect.c |6
Open the sockets calling sock_create_kern() with the correct struct net
pointer, and use the correct struct net pointer when verifying the
address passed to rds_bind().
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
net/rds/bind.c|3 ++-
net/rds/connection.c | 16
for lo to become free. Usage count .. being seen in the interval between
ethernet device migration to init_net and the keepalive timeout
Patch 3 in this set is independant of the above two changes, and is
a bugfix/follow up to eeb1bd5c encountered while testing the above.
Sowmini Varadhan
(listen)
socket is a kernel socket (defined in sk_alloc() as having
sk_net_refcnt == 0), then the newsk should also have a 0 sk_net_refcnt
and should not hold a get_net() reference.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
net/core/sock.c |3 ++-
1 files changed, 2
I'm running into a netns refcnt issue, and I suspect that
eeb1bd5c has something to do with it (perhaps we need an
additional change in sk_clone_lock() after eeb1bd5c).
Here's the problem:
When we create an syn_recv sock based on a kernel listen sock, we
take a get_net() ref with a stack
On (07/27/15 12:40), ebied...@xmission.com wrote:
sock_create_kern and friends are specialied interfaces for special
purposes. At a quick read through I don't think we have a single in
tree user doing with them what you are trying to do.
That doesnt change the fact that the architecture is
On (07/27/15 11:13), Cong Wang wrote:
That refcnt should be released in sock destructor too, when the tcp
connection is terminated.
yes, but in my case, the listen socket is opened as part of
the -init indirection in pernet_operations (thus it is a kernel socket)
and the expectation is that
On (07/27/15 11:37), Cong Wang wrote:
dlm uses a kernel TCP socket too, but it allocates a new socket and calls
-accept() by itself. ;)
sure, and rds does this in rds_tcp_accept_one() too.
But the newsk being created in sk_clone_lock is the one on an
incoming syn, i.e., the one that is
On Tue, Jul 21, 2015 at 4:38 PM, Sebastian Fett db_ext...@gmx.de wrote:
Hello!
According to RFC3927 every ARP packet (reply and request) should be sent as
link layer broadcast as long as the sender IP is a link local address. (see
chapter 2.5).
Because broadcast replies are noisy and should
On Wed, Jul 22, 2015 at 9:49 AM, Sebastian Fett db_ext...@gmx.de wrote:
what is your use case?
My problem ist a local network of audio devices. It is a valid possibility
that two halfs of the setup are set up individually (Stage left and stage
right). Both local networks will auto configure
I am working on making rds-tcp to be netns-aware, and in addition
to a few bug fixes that I'm lining up, there's a basic issue with
the way rds-tcp sets up the listen socket that is causing problems
The RDS tcp listen endpoint is created as part of module init.
(rds_tcp_init -
__vxlan_find_mac invokes ether_addr_equal on the eth_addr field,
which triggers unaligned access messages, so rearrange vxlan_fdb
to avoid this as non-intrusively as possible.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
drivers/net/vxlan.c |2 +-
1 files changed, 1
On (07/17/15 16:07), Joe Perches wrote:
On Fri, 2015-07-17 at 22:00 +0200, Sowmini Varadhan wrote:
__vxlan_find_mac invokes ether_addr_equal on the eth_addr field,
which triggers unaligned access messages, so rearrange vxlan_fdb
to avoid this in the most non-intrusive way.
What arch does
__vxlan_find_mac invokes ether_addr_equal on the eth_addr field,
which triggers unaligned access messages, so rearrange vxlan_fdb
to avoid this in the most non-intrusive way.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
v2: Alexander Duyck comments: make eth_addr[] 64b aligned
On 07/18/2015 08:06 PM, Joe Perches wrote:
It seems that this code has had unaligned accesses
on this field even before compare_ether_addr was
converted to ether_addr_equal.
Is sparc64 the only one that emits / ratelimits that
unaligned access message? I looked a little, but I
didn't find a
__vxlan_find_mac invokes ether_addr_equal on the eth_addr field,
which triggers unaligned access messages, so rearrange vxlan_fdb
to avoid this in the most non-intrusive way.
Signed-off-by: Sowmini Varadhan sowmini.varad...@oracle.com
---
v2: Alexander Duyck comments: place eth_addr[] to be 64b
ligned, and this
cannot be changed for nla_data() that is passed up to userspace.
As a result, the put_unaligned() macro needs to be used to
set up potentially unaligned fields such as the xfrm_stats in
copy_to_user_state()
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2: re
. Petersen <martin.peter...@oracle.com>
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
drivers/net/ethernet/intel/i40e/i40e_common.c | 36 +
1 files changed, 36 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_c
in the same
manner as the skb_clone failure: bail out of rds_tcp_data_recv(), and
retry via the deferred call to rds_send_worker() that gets set up on
ENOMEM from rds_tcp_read_sock()
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
net/rds/tcp_recv.c | 11 +--
1 files chan
On (10/21/15 06:22), David Miller wrote:
> memcpy() _never_ works for avoiding unaligned accessed.
>
> I repeat, no matter what you do, no matter what kinds of casts or
> fancy typing you use, memcpy() _never_ works for this purpose.
:
> There is one and only one portable way to access
On (11/16/15 21:14), Jason A. Donenfeld wrote:
>
> But what about in devices for which self-routing might actually be
> useful? For example, let's say that if an incoming skb is headed for
> dst X, it gets encapsulated and sent to dst A, and for dst Y it gets
> encapsulated and sent to dst B, and
> Neat. Though, in my case, I'm not actually just prepending a header.
> I'm doing some more substantial transformations of a packet. And this
> needs to work with v4 too. So I'm not sure implementing a v6 spec will
Understood, that spec was just referenced to indicate that there
are more issues
On (11/02/15 17:26), Nelson, Shannon wrote:
> > I assume you mean .1q
>
> Yes, this is what I had in mind.
I dont think we're quite there yet, even without vlans.
If I turn on/off tcpdump, there's something about the way that the link
is bounced that leaves the device down while tcpdump is
On (10/30/15 22:03), Nelson, Shannon wrote:
> The more common idiom in our driver would be
>
> err = i40e_get_platform_mac_addr(..);
> if (err) {
Ok.
> Have you tested this beyond a compile?
> Do you have a DT model to try this against?
yes.
> In looking at a couple other drivers,
On (10/30/15 02:14), Andy Shevchenko wrote:
>
> Does the following has no stubs?
>
> > + struct i40e_pf *pf = hw->back;
> > + struct device_node *dp = pci_device_to_OF_node(pf->pdev);
> > + const unsigned char *addr;
> > +
> > + addr = of_get_mac_address(dp);
>
> ^^^
I
. Petersen <martin.peter...@oracle.com>
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2: review comments from Andy Shevchenko
drivers/net/ethernet/intel/i40e/i40e_common.c | 32 +
1 files changed, 32 insertions(+), 0 deletions(-)
diff --git
f invoking i40e_macaddr_init() on
platforms that use the default mac address (i.e., when it is not from
OF or idprom) will cause harm, and if it is necessary/possible to
move this invocation to an earlier point in i40e_probe().
Reviewed-by: Martin K. Petersen <martin.peter...@oracle.com>
Signed
On (10/30/15 19:13), Sowmini Varadhan wrote:
> > In looking at a couple other drivers, I see the difference being that
> > they typically are writing the primary mac filter on probe (and any
> > other reset), whereas the i40e "knows" that the default mac address is
>
On (11/01/15 21:03), Nelson, Shannon wrote:
> .. In the meantime, be sure to test what happens over a reset, such as what
> happens when the MTU is changed. This will make sure that the replay
> of mac and vlan filters happens correctly. You'll want to test this
> with and without vlans.
I
. Petersen <martin.peter...@oracle.com>
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2: andy shevchenko comments
v3: more andy shevchenko comments
drivers/net/ethernet/intel/i40e/i40e_common.c | 30 +
1 files changed, 30 insertions(+), 0 de
. Petersen <martin.peter...@oracle.com>
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2, v3: Andy Shevchenko comments
v4: Shannon Nelson review: explicitly set up mac filters before register_netdev
v5: Shannon Nelson code style comments
v6: Shannon Nelson code style
On (11/04/15 21:59), Andy Shevchenko wrote:
>
> Usually the structure of kernel doc is something like following
>
> /**
> * func - summary
> * @paramx: desc
> *
> * Description:
> * Long description in many lines and / or paragraphs
> *
> * Returns:
> * 0 on success or errno otherwise.
>
. Petersen <martin.peter...@oracle.com>
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2, v3: Andy Shevchenko comments
v4: Shannon Nelson review: explicitly set up mac filters before register_netdev
v5: Shannon Nelson code style comments
drivers/net/ethernet/intel/i
On (11/02/15 14:57), Sowmini Varadhan wrote:
> On (11/02/15 17:26), Nelson, Shannon wrote:
> > > I assume you mean .1q
> >
> > Yes, this is what I had in mind.
>
> I dont think we're quite there yet, even without vlans.
>
Ok finally got all the .1q stuff
On (11/05/15 11:29), David Miller wrote:
>
> The intention is to let your patch go in as-is, then try and update
> ixgbe/i40e later in net-next or similar.
Sounds good, I can take care of ixgbe/i40e after that happens.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the
On (11/05/15 11:05), David Miller wrote:
> From: David Miller
> Date: Thu, 05 Nov 2015 10:31:26 -0500 (EST)
>
> > I'll see if I can cook something up.
>
> How does this look?
Looks good to me,
Do you want me to respin patch v7 with this? Or update ixgbe/i40e to use
this
On (10/30/15 18:57), Nelson, Shannon wrote:
> > >
> > > Going along with this being the equivalent of the ixgbe patch, I'd
> > > prefer the new code to be in i40e_main.c, rather than in i40e_common.c.
> > > In the design of our drivers, the common file is essentially a device
> > > specific layer,
On (10/30/15 20:06), Andy Shevchenko wrote:
>
> > +#include "i40e.h"
>
> Why do you need this one exactly?
I needed it to find pf->pdev below.
> > + struct device_node *dp = pci_device_to_OF_node(pf->pdev);
Without it, you will get:
:
CC [M]
On (10/30/15 18:28), Nelson, Shannon wrote:
>
> Going along with this being the equivalent of the ixgbe patch, I'd
> prefer the new code to be in i40e_main.c, rather than in i40e_common.c.
> In the design of our drivers, the common file is essentially a device
> specific layer, and the OS and
A two-part patchset that fixes some "unaligned access" warnings
that showed up my sparc test machines with ipsec set up.
Sowmini Varadhan (2):
crypto/x509: Fix unaligned access in x509_get_sig_params()
Fix unaligned access in xfrm_notify_sa() for DELSA
crypto/asymm
alignment values into consideration when doing kzalloc()
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
crypto/asymmetric_keys/x509_public_key.c |5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/crypto/asymmetric_keys/x509_public_key.c
b/crypto
On sparc, deleting established SAs (e.g., by restarting ipsec
at the peer) results in unaligned access messages via
xfrm_del_sa -> km_state_notify -> xfrm_send_state_notify().
Use an aligned pointer to xfrm_usersa_info for this case.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@o
On (10/21/15 06:54), Sowmini Varadhan wrote:
> But __alignof__(*p) is 8 on sparc, and without the patch I get
> all types of unaligned access. So what do you suggest as the fix?
Even though the alignment is, in fact, 8 (and that comes from
struct xfrm_lifetime_cfg), if uspace is firmly at
On (10/21/15 08:57), Steffen Klassert wrote:
> > --- a/net/xfrm/xfrm_user.c
> > +++ b/net/xfrm/xfrm_user.c
> > @@ -2659,7 +2659,7 @@ static int xfrm_notify_sa(struct xfrm_state *x, const
> > struct km_event *c)
> > if (attr == NULL)
> > goto out_free_skb;
> >
> >
e pong has already been added to the sendq earlier..
in the case of IB, if RDS_LL_SEND_FULL has been set, it takes some
head-scratching to figure out how the pong gets sent, and a few
comments could help clarify that.
but other than that, the contents look good to me, thus
Acked-by: Sowmini Varad
for transport like IB.
Acked-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
ng
is set to 0 so that we do not trigger any reconnect attempts on the
passive side of the tcp socket in the future, in conformance with
commit c82ac7e69efe ("net/rds: RDS-TCP: only initiate reconnect attempt
on outgoing TCP socket.")
Signed-off-by: Sowmini Varadhan <sowmini.varad...@ora
ces.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
net/rds/bind.c |9 -
1 files changed, 8 insertions(+), 1 deletions(-)
diff --git a/net/rds/bind.c b/net/rds/bind.c
index bc6b93e..6192566 100644
--- a/net/rds/bind.c
+++ b/net/rds/bind.c
@@ -196,7 +196,
Thinking back a bit about the kcm proposal:
https://www.mail-archive.com/netdev@vger.kernel.org/msg78696.html
I had a question:
If the user-space has decided to encrypt the http/2 header using tls,
the len (and other http/2 fields) is no longer in the clear for the kernel.
My understanding is
On Wed, Oct 7, 2015 at 11:17 AM, David Ahern wrote:
> Currently, all ipv6 addresses are flushed when the interface is configured
> down, including global, static addresses:
:
>
> Add a new sysctl to make this behavior optional. The new setting defaults to
> flush all
On (10/12/15 15:05), Tom Herbert wrote:
> > There is a different but related problem in this space- existing TLS/DTLS
> > libraries (openssl, gnutls etc) only know how to work with tcp
> > or udp sockets - they do not know anything about PF_RDS or the
> > newly proposed kcm socket type.
> >
>
On Mon, Jul 6, 2015 at 5:03 PM, David Ahern d...@cumulusnetworks.com wrote:
This driver borrows heavily from IPvlan and teaming drivers.
Routing domains (VRF-lite) are created by instantiating a device
and enslaving all routed interfaces that participate in the domain.
As part of the
On Thu, Jul 9, 2015 at 7:19 PM, David Ahern d...@cumulusnetworks.com wrote:
On the to-do list to use cmsg to specify a VRF for outbound packets using
non-connected sockets. I do not believe it is going to work, but need to
look into it.
What about setting ipsec policy for interfaces in the
On Fri, Jul 10, 2015 at 4:39 AM, David Ahern d...@cumulusnetworks.com wrote:
If I set the VRF context (ie., set the SO_BINDTODEVICE for all sockets) of
any RDS, NFS or any other socket app it runs in that VRF context and works
just fine
What if the application wants to do SO_BINDTODEVICE?
--
Only return a conn if the rds_conn_net(conn) matches the struct
net passed to rds_conn_lookup().
Fixes: 467fa15356ac ("RDS-TCP: Support multiple RDS-TCP listen endpoints,
one per netns.")
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
net/rds/con
On Tue, Aug 25, 2015 at 4:52 PM, David Ahern d...@cumulusnetworks.com wrote:
The VRF driver can check the device when the enslave request happens.
Will this work correctly if I set up a bonding interface or SVI,
and want to put the bond-master or SVI in the vrf (but subsequently
want to get,
On Tue, Aug 25, 2015 at 3:50 PM, Florian Fainelli f.faine...@gmail.com wrote:
Hi all,
This patch series implements a L2 only interface concept which basically
denies
any kind of IP address configuration on these interfaces, but still allows
them
to be used as configuration end-points to
On (09/10/15 08:43), roopa wrote:
> If you decide to use a flag, there is IFLA_EXT_MASK which is used to
> specify such filters from userspace today.
>
> /* New extended info filters for IFLA_EXT_MASK */
> #define RTEXT_FILTER_VF (1 << 0)
> #define RTEXT_FILTER_BRVLAN (1 << 1)
>
On (09/10/15 10:13), David Miller wrote:
> I don't think using such a generic netlink flag works best, the
> IFLA_EXT_MASK is definitely more suitable.
Ok, though this more of a IFLA_TRUNCATE_MASK than a IFLA_EXT_MASK.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the
it is desirable to avoid this if the information is
not needed.
This patch adds a the RTEXT_FILTER_SKIP_STATS extended info flag that
can be passed with netlink_request() to avoid statistics computation
for the ifinfo path.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2:
it is desirable to avoid this if the information is
not needed.
This patch adds a the RTEXT_FILTER_SKIP_STATS extended info flag that
can be passed with netlink_request() to avoid statistics computation
for the ifinfo path.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2:
On (09/12/15 00:22), Raghavendra K T wrote:
>
> Sowmini, Thanks for the patch which is more cleaner way without
> breaking current behaviour.
>
> [ Though RTEXT_FILTER_NEED_STATS flag with reverse effect would have
> helped immediately :)]
Agree, but existing legacy usage will not set this
it is desirable to avoid this if the information is
not needed.
This patch adds a the RTEXT_FILTER_SKIP_STATS extended info flag that
can be passed with netlink_request() to avoid statistics comuputation
for the ifinfo path.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
inclu
-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
net/rds/tcp.c | 16
1 files changed, 4 insertions(+), 12 deletions(-)
diff --git a/net/rds/tcp.c b/net/rds/tcp.c
index c42b60b..9d6ddba 100644
--- a/net/rds/tcp.c
+++ b/net/rds/tcp.c
@@ -67,21 +67,13 @@ void rds_tcp_n
socket with other parts of the kernel stack.
v2: fix kbuild-test-robot warnings, comments from Sergei Shtylov
and Santosh Shilimkar.
Sowmini Varadhan (3):
Use a single TCP socket for both send and receive.
Do not bloat sndbuf/rcvbuf in rds_tcp_tune
Set up MSG_MORE
ge in
rds_tcp_accept_one() has been modified to emulate the BGP state
machine: the smaller IP address should back off from the connection attempt.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2: kbuild-test-robot warning around __be32, modify subject line per
Sant
p_sendpage()
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2: Sergei Shtylov, Santosh Shilimkar comments (some parens retained for
readability)
net/rds/tcp_send.c |8 +++-
1 files changed, 7 insertions(+), 1 deletions(-)
diff --git a/net/rds/tcp_send.c
-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
net/rds/tcp.c | 16
1 files changed, 4 insertions(+), 12 deletions(-)
diff --git a/net/rds/tcp.c b/net/rds/tcp.c
index c42b60b..9d6ddba 100644
--- a/net/rds/tcp.c
+++ b/net/rds/tcp.c
@@ -67,21 +67,13 @@ void rds_tcp_n
socket with other parts of the kernel stack.
Sowmini Varadhan (3):
Use a single TCP socket for both send and receive.
Do not bloat sndbuf/rcvbuf in rds_tcp_tune
Set up MSG_MORE and MSG_SENDPAGE_NOTLAST as appropriate in
rds_tcp_xmit
net/rds/connection.c | 22
For the same reasons as 2f53384424 and 35f9c09fe9, rds_tcp_xmit
may have multiple pages to send, so use the MSG_MORE and
MSG_SENDPAGE_NOTLAST as hints to tcp_sendpage()
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
net/rds/tcp_send.c |8 +++-
1 files chan
ge in
rds_tcp_accept_one() has been modified to emulate the BGP state
machine: the smaller IP address should back off from the connection attempt.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
net/rds/connection.c | 22 ++
net/rds/rds.h|
On (09/30/15 08:56), santosh shilimkar wrote:
> Your checkpatch.pl should have complained about commit
> reference in the change-log. You might want to fix that
> for consistency.
It didnt. But ok, I'll fix this nit as well.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
On (09/30/15 08:50), santosh shilimkar wrote:
> > rs_tcp = (struct rds_tcp_connection *)conn->c_transport_data;
> >-WARN_ON(!rs_tcp || rs_tcp->t_sock);
> >+if (rs_tcp->t_sock && inet->inet_saddr < inet->inet_daddr) {
> >+struct sock *nsk = new_sock->sk;
> >
> Any reason you
" or
> - RDS TCP can use "RDS: TCP" or "RDS/TCP:"
Ok, but in this case patch 1/3 the changes affect both core and rds-tcp
modules.
Working on patchv2 that will address Sergei's comments and the
kbuild-test-robot warning as well
>
> $subject
> s/net/rds:/RDS:
&g
I'm doing some experiments that are trying to simultaneously
scaling the number of CPUs, and the number of processes
and encountering getifaddrs() weaknesses.
Others have run into similar things in the past, e.g.,
http://lists.openwall.net/netdev/2014/01/23/119
and more relevant to my
On (09/09/15 14:43), David Miller wrote:
>
> But what we could do is add a flag in the netlink request which
> elides the stats. GLIBC et al. could then start setting the flag.
>
Yes, interestingly that's what I was experimenting with myself
(though I was using a setsockopt in my version).
On (09/21/15 15:36), Tom Herbert wrote:
> segments. What we need to do, which you're probably doing for RDS, is
> do message delineation on the stream as a sequence of:
>
> 1) Read protocol header to determine message length (BPF used here)
right, that's what rds does- first reads the
`ls /sys/devices/channel-devices/vnet-port-0-0/net' is missing without
this change, and applications like NetworkManager are looking in
sysfs for the information.
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
drivers/net/ethernet/sun/sunvnet.c | 17 +++---
On (09/20/15 15:29), Tom Herbert wrote:
>
> Kernel Connection Multiplexor (KCM) is a facility that provides a
> message based interface over TCP for generic application protocols.
> The motivation for this is based on the observation that although
> TCP is byte stream transport protocol with no
On (09/21/15 10:33), Tom Herbert wrote:
> >
> > Some things that were not clear to me from the patch-set:
> >
> > The doc statses that we re-assemble packets the "stated length" -
> > but how will the receiver know the "stated length"?
>
> BPF program returns the length of the next message. In my
On (12/08/15 12:32), Steffen Klassert wrote:
>
> Would be nice if you could share the results. Comments are
Sure, not a problem. Give me some time though, I'm also looking
into the skb_cow_data and other memory-management issues that
were flagged on this thread.
I'll have all this info by
On (12/02/15 12:41), David Laight wrote:
> You are getting 0.7 Gbps with ass-ccm-a-128, scale the esp-null back to
> that and it would use 7/18*71 = 27% of the cpu.
> So 69% of the cpu in the a-128 case is probably caused by the
> encryption itself.
> Even if the rest of the code cost nothing
On (12/01/15 15:57), David Miller wrote:
> >> > Also current versions of SNMP provide more useful information about
> >> > network interface slot information in ifDescription
> >>
> >> Well if they do provide strings, then that is probably a better way
> >> forward than messing with the kernel.
On (12/02/15 13:07), Tom Herbert wrote:
> That's easy enough to add to flow dissector, but is SPI really
> intended to be used an L4 entropy value? We would need to consider the
yes. To quote https://en.wikipedia.org/wiki/Security_Parameter_Index
"This works like port numbers in TCP and UDP
On (12/02/15 14:01), Tom Herbert wrote:
> No, please don't persist is this myopic "we'll get to IPv6 later"
> model! IPv6 is a real protocol, it has significant deployment of the
> Internet, and there are now whole data centers that are IPv6 only
> (e.g. FB), and there are plenty of use cases of
On (12/02/15 13:44), Tom Herbert wrote:
> > IPv6 would be an interesting academic exercise, but it's going
> > to be a while before we get RDS-TCP to go over IPv6.
> >
> Huh? Who said anything about RDS-TCP? I thought you were trying to
> improve IPsec performance...
yes, and it would be nice to
On (12/03/15 09:45), Steffen Klassert wrote:
> pcrypt(echainiv(authenc(hmac(sha1-ssse3),cbc-aes-aesni)))
>
> Result:
>
> iperf -c 10.0.0.12 -t 60
>
> Client connecting to 10.0.0.12, TCP port 5001
> TCP window size: 45.0 KByte (default)
;martin.peter...@oracle.com>
Signed-off-by: Sowmini Varadhan <sowmini.varad...@oracle.com>
---
v2, v3: Andy Shevchenko comments
v4: Shannon Nelson review: explicitly set up mac filters before register_netdev
v5: Shannon Nelson code style comments
v6: Shannon Nelson code style comments
v7: Ensure
1 - 100 of 628 matches
Mail list logo