Hello,
On Fri, 2 Oct 2015, Simon Horman wrote:
> Since 6f2bcea9917d ("ipvs: Pass ipvs into ip_vs_in_icmp and ip_vs_in_icmp_v6")
> the net net variable in ip_vs_out() appears to be unused so remove it.
>
> Signed-off-by: Simon Horman
> ---
> net/netfilter/ipvs/ip_vs_core.c | 1 -
> 1 f
From: Peter Nørlund
Date: Tue, 6 Oct 2015 07:24:47 +0200
> This fixes
>
> net/built-in.o: In function `fib_rebalance':
> fib_semantics.c:(.text+0x9df14): undefined reference to `__divdi3'
>
> and
>
> net/built-in.o: In function `fib_rebalance':
> net/ipv4/fib_semantics.c:572: undefined refere
From: Peter Nørlund
Date: Mon, 5 Oct 2015 23:31:58 +0200
> What is the proper way to supply fixes for this kind of errors? Should
> I submit the original patches again with the bug fixed, or should I
> submit a patch fixing this particular problem? I imagine bisecting
> becomes annoying when a pa
From: Cong Wang
Date: Mon, 5 Oct 2015 10:53:57 -0700
> On Thu, Oct 1, 2015 at 8:49 AM, David Ahern wrote:
>> Eric's net namespace changes in 1b75097dd7a26 leaves net unreferenced if
>> CONFIG_IP_VS_IPV6 is not enabled:
>>
>> ../net/netfilter/ipvs/ip_vs_core.c: In function ‘ip_vs_out’:
>> ../net/
From: Tom Lendacky
Date: Mon, 5 Oct 2015 10:51:03 -0500
> The kasprintf function can return NULL if the allocation fails. Check for
> successful allocation before attempting to use the returned buffer.
>
> Signed-off-by: Tom Lendacky
Applied, thanks.
--
To unsubscribe from this list: send the
From: Mark Brown
Date: Mon, 5 Oct 2015 15:57:22 +0100
> On Mon, Oct 05, 2015 at 09:29:31AM -0400, j...@ringle.org wrote:
>> From: Jon Ringle
>>
>> The only time that it makes sense to call a custom provided reg_update_bits
>> function, is the register being updated is one that has volatile bits
From: Carol Soto
Date: Mon, 5 Oct 2015 10:24:19 -0500
>
>
> On 10/4/2015 3:03 AM, Or Gerlitz wrote:
>> On 9/29/2015 9:38 PM, cls...@linux.vnet.ibm.com wrote:
>>> From: Carol L Soto
>>>
>>> Test interrupts fails if not all completion vectors called
>>> request_irq. This case can happen if only
From: Mark Brown
Date: Mon, 5 Oct 2015 15:25:31 +0100
> On Mon, Oct 05, 2015 at 06:16:09AM -0700, David Miller wrote:
>
>> >> Applied.
>
>> > Thanks David. However, I've sent a v3 patch, and also expecting feedback
>> > from Mark Brown on the regmap portion of it.
>
>> Please send me relative
Hi all,
After merging the net-next tree, today's linux-next build (i386 defconfig)
failed like this:
net/built-in.o: In function `fib_rebalance':
fib_semantics.c:(.text+0x9d6d4): undefined reference to `__divdi3'
Caused by commit
0e884c78ee19 ("ipv4: L3 hash-based multipath")
I have reverted
This fixes
net/built-in.o: In function `fib_rebalance':
fib_semantics.c:(.text+0x9df14): undefined reference to `__divdi3'
and
net/built-in.o: In function `fib_rebalance':
net/ipv4/fib_semantics.c:572: undefined reference to `__aeabi_ldivmod'
Fixes: 0e884c78ee19 ("ipv4: L3 hash-based multipath"
On Mon, 5 Oct 2015 15:23:10 -0700
Alexander Duyck wrote:
> Normally the correct comment would include something like the
> following just before the signed-off-by line.
> Fixes: 0e884c78ee19 ("ipv4: L3 hash-based multipath")
>
Got it - I misunderstood the format. I'll submit a revised patc
On Mon, Oct 5, 2015 at 10:43 AM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> This is another step on the way to per-world clean cut. Introduce world
> ops hooks which each world can implement in world-specific way.
> Also introduce world infrastructure including function for port worlds
> change.
>
On Sun, Oct 4, 2015 at 2:25 PM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> Introduce a helper to ask HW for change of the port mode (world).
>
> Signed-off-by: Jiri Pirko
Acked-by: Scott Feldman
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majo
On Mon, Oct 5, 2015 at 10:43 AM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> Carve out TLV processing helpers into separate files.
>
> Signed-off-by: Jiri Pirko
Acked-by: Scott Feldman
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.
> > The second one seems to be trickier. It looks like a race wrt. PADT
> > message reception. Reproducing the bug will probably require to
> > generate some PADT flooding to a host that creates and releases PPPoE
> > connections.
Ok I think I can see the potential race here, specifically the PADT
On Mon, Oct 5, 2015 at 10:43 AM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> Since "rocker.c" is going to be split into multiple files, start with
> renaming original "rocker.c" file to "rocker_main.c". Multiple code
> parts are going to be cut from "rocker_main.c" later on.
>
> Signed-off-by: Jiri
On Mon, Oct 5, 2015 at 10:43 AM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> Since "rocker.h" file is going to be used for different purpose,
> rename the hardware-specific header to "rocker_hw.h".
>
> Signed-off-by: Jiri Pirko
Acked-by: Scott Feldman
--
To unsubscribe from this list: send the li
On Mon, Oct 5, 2015 at 10:43 AM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> No need to pass rocker_port around to alloc/free rocker functions,
> since they now use switchdev_trans for memory management storage.
> With the param removal, shorten the name of the functions since they now
> has nothing
On Sat, 2015-10-03 at 17:19 +0200, Christoph Hellwig wrote:
> This ensures the dma mask that is supported by the driver is recorded
> in the device structure.
>
> Signed-off-by: Christoph Hellwig
> ---
> drivers/net/ethernet/amd/pcnet32.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
On 10/05/2015 04:39 PM, Tom Herbert wrote:
This provides an example of a driver calling the skb_csum_offload_check.
Signed-off-by: Tom Herbert
---
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 6 +++---
drivers/net/ethernet/mellanox/mlx4/en_tx.c | 20 +++-
2 files ch
There are multiple packages and channels connecting to NCSI enabled
interface. The available packages and channels are probed with NCSI
packets when NCSI device is started. Among the available packages
and channels, only one channel should be selected to be active to
provide service. When the activ
According to NCSI spec (DSP0222), the NCSI enabled interface can
connected to multiple packages simultaneously, up to 8 packages.
Each package includes multiple channels, up to 32 channels. At
one moment, one channel is enabled to provide service to the NCSI
enabled interface. Besides, each channel
This patchset is prototype requesting for comments.
This series of patches intends to support NCSI stack, which is specified by
DSP0222. One NCSI enabled interface potentially connects to multiple packages
and channels, but there is one active channel at once when hardware arbitration
is disabled.
On Mon, Oct 5, 2015 at 10:43 AM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> This patchset allows new rocker worlds to be easily added in future (like eBPF
> based one I have been working on). The main part of the patchset is the OF-DPA
> carve-out. It resuts in OF-DPA specific file. Clean cut.
>
>
On 10/05/2015 04:39 PM, Tom Herbert wrote:
Add skb_csum_offload_chk driver helper function to determine if a
device with limited checksum offload capabilities is able to offload the
checksum for a given packet.
This patch includes:
- The skb_csum_offload_chk function. Returns true if checksum
NCSI packet encapsulated in ethernet frame is recognized by the
ethernet protocol field (0x88F8). According to NCSI spec, the
destination MAC field should be 0xFF's, but the source MAC field
could be arbitrary one since channel field in NCSI header indicates
the source of the packet.
There are 3 t
This enables IP/UDP/TCP offload checksum according to information
passed on from bootloader through device-tree. The offload doesn't
working properly when the interface works in NCSI mode.
Signed-off-by: Gavin Shan
---
drivers/net/ethernet/faraday/ftgmac100.c | 6 +-
1 file changed, 5 insert
This adds netlink support so that userland can retrieve the NCSI
topology on the specified interface. Also, more netlink commands
are introduced to retrieve information from the active channel
or configure the active channel with provided options.
Signed-off-by: Gavin Shan
---
include/uapi/linux
The patch replaces variable name "use_nc_si" with "use_ncsi". No
function changed.
Signed-off-by: Gavin Shan
---
drivers/net/ethernet/faraday/ftgmac100.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/faraday/ftgmac100.c
b/drivers/net/eth
The NIC has the possibility to connect to NCSI package and channel.
This supports NCSI enabled interface. When the network device is
registered, the accompanying NCSI device is registered. When the
interface is to be brought up, the NCSI device is started to probe
NCSI topology and choose one activ
On 10/5/15 3:14 PM, Daniel Borkmann wrote:
One scenario that comes to mind ... what happens when there are kernel
pointers stored in skb->cb[] (either from the current layer or an old
one from a different layer that the skb went through previously, but
which did not get overwritten)?
Socket filt
On 10/5/15 3:02 PM, Kees Cook wrote:
the array maps that hold FDs (BPF_MAP_TYPE_PROG_ARRAY and
>BPF_MAP_TYPE_PERF_EVENT_ARRAY) don't have lookup/update accessors
>from the program side, so programs cannot see or manipulate
>those pointers.
>For the former only bpf_tail_call() is allowed that take
On Wed, Sep 30, 2015 at 8:33 PM, Thomas F Herbert
wrote:
> Add support for 802.1ad including the ability to push and pop double
> tagged vlans. Add support for 802.1ad to netlink parsing and flow
> conversion. Uses double nested encap attributes to represent double
> tagged vlan. Inner TPID encode
On Mon, 2015-10-05 at 14:24 +0200, Guillaume Nault wrote:
> On Mon, Oct 05, 2015 at 04:08:51AM +, Matt Bennett wrote:
> > Hi, I am seeing this panic occur occasionally however I am unsure how to
> > go about reproducing it. Is it enough to simply keep creating and
> > tearing down the PPP inter
Hi all,
Today's linux-next merge of the net-next tree got a conflict in:
net/ipv6/xfrm6_output.c
between commit:
93efac3f2e03 ("ipv6: Fix IPsec pre-encap fragmentation check")
from the ipsec tree and commit:
7d8c6e391575 ("ipv6: Pass struct net through ip6_fragment")
from the net-next
Hi all,
Today's linux-next merge of the net-next tree got a conflict in:
net/ipv4/inet_connection_sock.c
between commit:
2306c704ce28 ("inet: fix race in reqsk_queue_unlink()")
from the net tree and commit:
079096f103fa ("tcp/dccp: install syn_recv requests into ehash table")
from the
Hi all,
Today's linux-next merge of the net-next tree got a conflict in:
drivers/net/usb/asix_common.c
between commit:
f6194bcf03e4 ("net: usb: asix: Fix crash on skb alloc failure")
from the net tree and commit:
6a570814cd43 ("asix: Continue processing URB if no RX netdev buffer")
(amo
On Mon, 5 Oct 2015 21:26:39 +0200
Jesper Dangaard Brouer wrote:
> My only problem left, is I want a perf measurement that pinpoint these
> kind of spots. The difference in L1-icache-load-misses were significant
> (1,278,276 vs 2,719,158). I tried to somehow perf record this with
> different per
This provides an example of a driver calling the skb_csum_offload_check.
Signed-off-by: Tom Herbert
---
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 6 +++---
drivers/net/ethernet/mellanox/mlx4/en_tx.c | 20 +++-
2 files changed, 18 insertions(+), 8 deletions(-)
diff --
Add skb_csum_offload_chk driver helper function to determine if a
device with limited checksum offload capabilities is able to offload the
checksum for a given packet.
This patch includes:
- The skb_csum_offload_chk function. Returns true if checksum is
offloadable, else false. Optionally, i
Same thing as skb_transport_offset but returns the offset of the inner
transport header (when skb->encpasulation is set).
Signed-off-by: Tom Herbert
---
include/linux/skbuff.h | 5 +
1 file changed, 5 insertions(+)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 4398411..
When drivers have support for offloading transport IP checksums, they
will indicate this in the features for the device. As described in
skbuff.h, a driver will usually advertise NETIF_F_HW_CSUM,
NETIF_F_IP_CSUM, and/or NETIF_F_IPV6_CSUM. The first of these
(NETIF_F_HW_CSUM) is the preferred method
On Mon, Oct 5, 2015 at 3:23 PM, Alexander Duyck
wrote:
> On 10/05/2015 03:14 PM, Peter Nørlund wrote:
>>
>> commit 0e884c78ee19e902f300ed147083c28a0c6302f0 ("ipv4: L3 hash-based
>> multipath") broke compilation for i386 and arm due to dependency on
>> 64-bit division.
>>
>> The problem is simply s
On 10/05/2015 04:26 PM, Steven Schlansker wrote:
On Sep 25, 2015, at 7:58 PM, Guenter Roeck wrote:
On 09/25/2015 02:37 PM, Steven Schlansker wrote:
Thank you for the patches to try, I'll build a kernel with them early next week
and report back. It sounds like it may not match my problem e
On Sep 25, 2015, at 7:58 PM, Guenter Roeck wrote:
> On 09/25/2015 02:37 PM, Steven Schlansker wrote:
>>
>>
>> Thank you for the patches to try, I'll build a kernel with them early next
>> week
>> and report back. It sounds like it may not match my problem exactly so we'll
>> see.
>>
>
> Fo
(trimmed Cc list a little)
On Mon, 5 Oct 2015 14:20:45 -0700 Andi Kleen wrote:
> > My only problem left, is I want a perf measurement that pinpoint these
> > kind of spots. The difference in L1-icache-load-misses were significant
> > (1,278,276 vs 2,719,158). I tried to somehow perf record thi
Previously, the CT_ATTR_FLAGS attribute, when nested under the
OVS_ACTION_ATTR_CT, encoded a 32-bit bitmask of flags that modify the
semantics of the ct action. It's more extensible to just represent each
flag as a nested attribute, and this requires no additional error
checking to reject flags tha
These comments hadn't caught up to their implementations, fix them.
Fixes: 7f8a436 "openvswitch: Add conntrack action"
Signed-off-by: Joe Stringer
Acked-by: Pravin B Shelar
---
v3: No change.
v2: Acked.
---
include/uapi/linux/openvswitch.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(
The ct_state field was initially added as an 8-bit field, however six of
the bits are already being used and use cases are already starting to
appear that may push the limits of this field. This patch extends the
field to 32 bits while retaining the internal representation of 8 bits.
This should co
The ct action uses parts of the flow key, so we need to ensure that it
is valid before executing that action.
Fixes: 7f8a436 "openvswitch: Add conntrack action"
Signed-off-by: Joe Stringer
Acked-by: Pravin B Shelar
---
v3: No change.
v2: Acked.
---
net/openvswitch/actions.c | 6 ++
1 file c
If ovs_fragment() was unable to fragment the skb due to an L2 header
that exceeds the supported length, skbs would be leaked. Fix the bug.
Fixes: 7f8a436 "openvswitch: Add conntrack action"
Signed-off-by: Joe Stringer
Acked-by: Pravin B Shelar
---
v3: Acked.
v2: Drop if condition by returning in
The userspace side of the Open vSwitch conntrack changes is currently
undergoing review, which has highlighted some minor bugs in the existing
conntrack implementation in the kernel, as well as pointing out some
future-proofing that can be done on the interface to reduce the need for
additional com
Previously, if userspace specified ct_state bits in the flow key which
are currently undefined (and therefore unsupported), then they would be
ignored. This could cause unexpected behaviour in future if userspace is
extended to support additional bits but attempts to communicate with the
current ve
On 10/05/2015 03:14 PM, Peter Nørlund wrote:
commit 0e884c78ee19e902f300ed147083c28a0c6302f0 ("ipv4: L3 hash-based
multipath") broke compilation for i386 and arm due to dependency on
64-bit division.
The problem is simply solved by using the proper macro for 64-bit
divison.
Signed-off-by: Peter
commit 0e884c78ee19e902f300ed147083c28a0c6302f0 ("ipv4: L3 hash-based
multipath") broke compilation for i386 and arm due to dependency on
64-bit division.
The problem is simply solved by using the proper macro for 64-bit
divison.
Signed-off-by: Peter Nørlund
---
net/ipv4/fib_semantics.c | 4 ++-
On 10/05/2015 10:48 PM, Alexei Starovoitov wrote:
In order to let unprivileged users load and execute eBPF programs
teach verifier to prevent pointer leaks.
Verifier will prevent
- any arithmetic on pointers
(except R10+Imm which is used to compute stack addresses)
- comparison of pointers
- p
On Mon, Oct 5, 2015 at 2:12 PM, Alexei Starovoitov wrote:
> On 10/5/15 2:00 PM, Kees Cook wrote:
>>
>> On Mon, Oct 5, 2015 at 1:48 PM, Alexei Starovoitov
>> wrote:
>>>
>>> >In order to let unprivileged users load and execute eBPF programs
>>> >teach verifier to prevent pointer leaks.
>>> >Verifier
On 10/05/2015, 07:29 PM, Roland Dreier wrote:
> From: Roland Dreier
>
> Backports of 41fc014332d9 ("fib_rules: fix fib rule dumps across
> multiple skbs") introduced a regression in "ip rule show" - it ends up
> dumping the first rule over and over and never exiting, because 3.19
> and earlier ar
On Sat, 3 Oct 2015 15:32:36 +0200 Christoph Hellwig wrote:
> This series consolidates the code to implement configfs attributes
> by providing the ->show and ->store method in common code and using
> container_of in the methods to access the containing structure.
>
> This reduces source and bin
Peter Nørlund wrote:
> What is the proper way to supply fixes for this kind of errors? Should
> I submit the original patches again with the bug fixed, or should I
> submit a patch fixing this particular problem?
The latter, relative fix, with 'Fixes: ("name of commit") tag'.
--
To unsubscribe f
On 10/5/15 2:16 PM, Andy Lutomirski wrote:
On Mon, Oct 5, 2015 at 2:12 PM, Alexei Starovoitov wrote:
On 10/5/15 2:00 PM, Kees Cook wrote:
On Mon, Oct 5, 2015 at 1:48 PM, Alexei Starovoitov
wrote:
In order to let unprivileged users load and execute eBPF programs
teach verifier to prevent po
On Mon, 05 Oct 2015 04:26:38 -0700 (PDT)
David Miller wrote:
> From: kbuild test robot
> Date: Mon, 5 Oct 2015 18:54:18 +0800
>
> > tree:
> > https://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git
> > master head: 937317c7c1097aa878a5000e3aab616eb5c590c0 commit:
> > 0e884c78ee19e9
> My only problem left, is I want a perf measurement that pinpoint these
> kind of spots. The difference in L1-icache-load-misses were significant
> (1,278,276 vs 2,719,158). I tried to somehow perf record this with
> different perf events without being able to pinpoint the location (even
> thoug
On Mon, Oct 5, 2015 at 2:12 PM, Alexei Starovoitov wrote:
> On 10/5/15 2:00 PM, Kees Cook wrote:
>>
>> On Mon, Oct 5, 2015 at 1:48 PM, Alexei Starovoitov
>> wrote:
>>>
>>> >In order to let unprivileged users load and execute eBPF programs
>>> >teach verifier to prevent pointer leaks.
>>> >Verifier
On 10/5/15 2:00 PM, Kees Cook wrote:
On Mon, Oct 5, 2015 at 1:48 PM, Alexei Starovoitov wrote:
>In order to let unprivileged users load and execute eBPF programs
>teach verifier to prevent pointer leaks.
>Verifier will prevent
>- any arithmetic on pointers
> (except R10+Imm which is used to c
On Mon, Oct 5, 2015 at 1:48 PM, Alexei Starovoitov wrote:
> In order to let unprivileged users load and execute eBPF programs
> teach verifier to prevent pointer leaks.
> Verifier will prevent
> - any arithmetic on pointers
> (except R10+Imm which is used to compute stack addresses)
> - comparis
I think it is time to liberate eBPF from CAP_SYS_ADMIN.
As was discussed when eBPF was first introduced two years ago
the only piece missing in eBPF verifier is 'pointer leak detection'
to make it available to non-root users.
Patch 1 adds this pointer analysis.
The eBPF programs, obviously, need to
In order to let unprivileged users load and execute eBPF programs
teach verifier to prevent pointer leaks.
Verifier will prevent
- any arithmetic on pointers
(except R10+Imm which is used to compute stack addresses)
- comparison of pointers
- passing pointers to helper functions
- indirectly pass
since eBPF programs and maps use kernel memory consider it 'locked' memory
from user accounting point of view and charge it against RLIMIT_MEMLOCK limit.
This limit is typically set to 64Kbytes by distros, so almost all
bpf+tracing programs would need to increase it, since they use maps,
but kernel
Mon, Oct 05, 2015 at 08:32:33PM CEST, marcelo.leit...@gmail.com wrote:
>On Wed, Sep 30, 2015 at 03:51:41PM +0200, Jiri Pirko wrote:
>> Hi Dave/all.
>>
>> There is quite common for drivers to speak with HW using "packet-like"
>> messages. In our case, we put message into skb and push that out as
>>
On 10/05/2015 06:59 AM, Vlastimil Babka wrote:
On 10/02/2015 12:18 PM, Konstantin Khlebnikov wrote:
When openvswitch tries allocate memory from offline numa node 0:
stats = kmem_cache_alloc_node(flow_stats_cache, GFP_KERNEL |
__GFP_ZERO, 0)
It catches VM_BUG_ON(nid < 0 || nid >= MAX_NUMNODES ||
On Mon, 2015-10-05 at 09:23 -0700, Tom Herbert wrote:
>
> 1) Drivers may advertise NETIF_F_HW_CSUM. The stack will indicate
> checksum offload exclusively using the
> CHECKSUM_PARTIAL/csum_start/csum_offset interface. No additional
> interfaces (bits in skbuff should not be needed)
> 2) A driver m
On Mon, Oct 5, 2015 at 6:24 PM, Carol Soto wrote:
>
>
> On 10/4/2015 3:03 AM, Or Gerlitz wrote:
>>
>> On 9/29/2015 9:38 PM, cls...@linux.vnet.ibm.com wrote:
>>>
>>> From: Carol L Soto
>>>
>>> Test interrupts fails if not all completion vectors called
>>> request_irq. This case can happen if only
From: Craig Gallek
These patches introduce a structure for reporting udp statistics
through netlink. Simple datagram and byte counters are included.
Craig Gallek (2):
sock_diag: create udp_info for udp socket stats
sock_diag: initial udp_info metrics
include/linux/inet_diag.h | 1 +
From: Craig Gallek
struct udp_info is intended to be the UDP equivalent of tcp_info for the
INET_DIAG_INFO netlink attribute of TCPDIAG_GETSOCK and
SOCK_DIAG_BY_FAMILY.
Unfortunately, since clients may assume that all instances of the
INET_DIAG_INFO response attribute are serialized versions of
From: Craig Gallek
Define per-UDP socket metrics for counting datagrams in and out and
bytes in and out. These four metrics are also exposed though the
INET_DIAG_UDP_INFO netlink attribute of the SOCK_DIAG_BY_FAMILY interface.
Performance test configuration to maximize cache misses across CPU s
On Mon, Oct 5, 2015 at 11:40 AM, Jiri Benc wrote:
> On Mon, 5 Oct 2015 10:58:17 -0700, Pravin B Shelar wrote:
>> --- a/drivers/net/vxlan.c
>> +++ b/drivers/net/vxlan.c
>> @@ -2337,6 +2337,51 @@ static int vxlan_change_mtu(struct net_device *dev,
>> int new_mtu)
>> return 0;
>> }
>>
>> +st
On Fri, 2 Oct 2015 14:50:44 -0700
Andrew Morton wrote:
> On Fri, 2 Oct 2015 15:40:39 +0200 Jesper Dangaard Brouer
> wrote:
>
> > > Thus, I need introducing new code like this patch and at the same time
> > > have to reduce the number of instruction-cache misses/usage. In this
> > > case we so
Hello Dave,
Could you please backport the following to stable-4.1.y:
commit 74e98eb085889b0d2d4908f59f6e00026063014f -- RDS: verify
the underlying transport exists before creating a connection
This fixes CVE-2015-6937.
commit 48900cb6af4282fa0fb6ff4d72a81aa3dadb5
Hello Arnd,
thanks for picking up this y2038 api issue.
On 09/30/2015 01:26 PM, Arnd Bergmann wrote:
> The can subsystem communicates with user space using a bcm_msg_head
> header, which contains two timestamps. This is problematic for
> multiple reasons:
>
> a) The structure layout is currently
On Mon, 5 Oct 2015 10:58:17 -0700, Pravin B Shelar wrote:
> --- a/drivers/net/vxlan.c
> +++ b/drivers/net/vxlan.c
> @@ -2337,6 +2337,51 @@ static int vxlan_change_mtu(struct net_device *dev,
> int new_mtu)
> return 0;
> }
>
> +static int egress_ipv4_tun_info(struct net_device *dev, struc
On Wed, Sep 30, 2015 at 03:51:41PM +0200, Jiri Pirko wrote:
> Hi Dave/all.
>
> There is quite common for drivers to speak with HW using "packet-like"
> messages. In our case, we put message into skb and push that out as
> an ordinary packet. HW then sends us reply in a packet, similar to other
> r
On 10/5/2015 10:56 AM, Santosh Shilimkar wrote:
Darn. Header hunk remained in the repo. Resending it.
From 4bebdd7a4d2960b2ff6c40b27156d041ea270765 Mon Sep 17 00:00:00 2001
From: Santosh Shilimkar
Date: Thu, 10 Sep 2015 11:57:14 -0700
Subject: [PATCH v3 05/14] RDS: defer the over_batch work t
> On Oct 5, 2015, at 9:23 AM, Tom Herbert wrote:
>
> 4) To help drivers for devices with limited offload capabilities we'll
> define a helper function to check for typical restrictions (.e.g. IPv4
> only, TCP/UDP only. no encapsulation, no IPv6 extension headers,
> etc.). I am working on this hel
On 10/5/15 12:01 PM, kbuild test robot wrote:
Hi David,
[auto build test ERROR on v4.3-rc4 -- if it's inappropriate base, please ignore]
net-next patches can *not* be applied to Linus' tree. If you are going
to run the build bot for patches submitted to netdev, please be sure to
apply those
On Fri, Oct 2, 2015 at 5:57 PM, Jesse Gross wrote:
> On Fri, Oct 2, 2015 at 12:32 PM, Pravin Shelar wrote:
>> On Thu, Oct 1, 2015 at 11:00 PM, Jiri Benc wrote:
>>> On Thu, 1 Oct 2015 17:11:56 -0700, Pravin Shelar wrote:
I dont see point of adding this code when IPv6 sampling not support by
Hi David,
[auto build test ERROR on v4.3-rc4 -- if it's inappropriate base, please ignore]
config: x86_64-randconfig-x002-201540 (attached as .config)
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64
All error/warnings (new ones prefixed by >>):
In
Current process gives up if its send work over the batch limit.
The work queue will get kicked to finish off any other requests.
This fixes remainder condition from commit 443be0e5affe ("RDS: make
sure not to loop forever inside rds_send_xmit").
The restart condition is only for the case where we
One need to take rds socket reference while using it and release it
once done with it. rds_add_bind() code path does not do that so
lets fix it.
Signed-off-by: Santosh Shilimkar
Signed-off-by: Santosh Shilimkar
---
net/rds/bind.c | 16 +++-
1 file changed, 11 insertions(+), 5 deleti
In Transport indepedent rds_sendmsg(), we shouldn't make decisions based
on RDS_LL_SEND_FULL which is used to manage the ring for RDMA based
transports. We can safely issue rds_send_xmit() and the using its
return value take decision on deferred work. This will also fix
the scenario where at times
Just in case we are still handling the QP receive completion while the
rds_ibdev is released, drop the connection instead of crashing the kernel.
Signed-off-by: Santosh Shilimkar
Signed-off-by: Santosh Shilimkar
---
net/rds/ib_cm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff -
synchronize_rcu() slowing down un-necessarily the socket shutdown
path. It is used just kfree() the ip addresses in rds_ib_remove_ipaddr()
which is perfect usecase for kfree_rcu();
So lets use that to gain some speedup.
Signed-off-by: Santosh Shilimkar
Signed-off-by: Santosh Shilimkar
---
net/
Similar to what we did with receive CQ completion handling, we split
the transmit completion handler so that it lets us implement batched
work completion handling.
We re-use the cq_poll routine and makes use of RDS_IB_SEND_OP to
identify the send vs receive completion event handler invocation.
Si
One global lock protecting hash-tables with 1024 buckets isn't
efficient and it shows up in a massive systems with truck
loads of RDS sockets serving multiple databases. The
perf data clearly highlights the contention on the rw
lock in these massive workloads.
When the contention gets worse, the c
[v3]
Updated patch "[PATCH v2 05/14] RDS: defer the over_batch work to
send worker" as per David Miller's comment [4] to avoid the magic
value usage. Patch now makes use of already available but unused
send_batch_count module parameter. Rest of the patches are same as
earlier version v2 [3]
[v2]:
On Mon, Oct 5, 2015 at 10:44 AM, Jiri Pirko wrote:
> From: Jiri Pirko
>
> Carve out OF-DPA would specific code from the common file to the world
> file. This change required struct rocker and struct rocker_port split
> into world specific struct ofdpa and struct ofdpa_port. Along with this
> the
While transitioning to netdev based vport we broke OVS
feature which allows user to retrieve tunnel packet egress
information for lwtunnel devices. Following patch fixes it
by introducing ndo operation to get the tunnel egress info.
Same ndo operation can be used for lwtunnel devices and compat
ov
Fix below warning by marking rds_ib_fmr_wq static
net/rds/ib_rdma.c:87:25: warning: symbol 'rds_ib_fmr_wq' was not declared.
Should it be static?
Signed-off-by: Santosh Shilimkar
Signed-off-by: Santosh Shilimkar
---
net/rds/ib_rdma.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
RDS IB mr pool has its own workqueue 'rds_ib_fmr_wq', so we need
to use queue_delayed_work() to kick the work. This was hurting
the performance since pool maintenance was less often triggered
from other path.
Signed-off-by: Santosh Shilimkar
Signed-off-by: Santosh Shilimkar
---
net/rds/ib_rdma.
RDS bind and release locking scheme is very inefficient. It
uses RCU for maintaining the bind hash-table which is great but
it also needs to hold spinlock for [add/remove]_bound(). So
overall usecase, the hash-table concurrent speedup doesn't pay off.
In fact blocking nature of synchronize_rcu() ma
1 - 100 of 252 matches
Mail list logo
