test on s390x, which depends on the read value matching what was
written. I've confirmed that the test now passes on big and little endian
systems.
Signed-off-by: Jason Baron
Fixes: 438ac88009bc ("net: fastopen: robustness and endianness fixes for
SipHash")
Cc: Ard Biesheuvel
Cc:
On 6/11/20 5:19 PM, jim.cro...@gmail.com wrote:
> trimmed..
>
Currently I think there not enough "levels" to map something like
drm.debug to the new dyn dbg feature. I don't think it is intrinsic
but I couldn't find the bit of the code where the 5-bit level in struct
_ddebug
On 5/29/20 7:52 PM, David Miller wrote:
> From: Jason Baron
> Date: Wed, 27 May 2020 16:25:28 -0400
>
>> Port based allow rules must currently allow all fragments since the
>> port number is not included in the 1rst fragment. We want to restrict
>> allowing all
, demonstrating the new behavior.
Jason Baron (2):
net: sched: cls-flower: include ports in 1rst fragment
selftests: tc_flower: add destination port tests
net/core/flow_dissector.c | 4 +-
net/sched/cls_flower.c | 3 +-
.../testing/selftests/net
ing a rule which
allows/blocks specific ports and allows all non-first ip fragments
(via setting ip_flags to frag/nofirstfrag).
Cc: Jamal Hadi Salim
Cc: Cong Wang
Cc: Jiri Pirko
Signed-off-by: Jason Baron
---
net/core/flow_dissector.c | 4 +++-
net/sched/cls_flower.c| 3 ++-
2 files chang
Verify that tc flower can match on destination port for
udp/tcp for both non-fragment and first fragment cases.
Cc: Jamal Hadi Salim
Cc: Cong Wang
Cc: Jiri Pirko
Signed-off-by: Jason Baron
---
.../testing/selftests/net/forwarding/tc_flower.sh | 73 +-
1 file changed, 72
lds do not cover all the cases where TFO may fail, but other
failures, such as SYN/ACK + data being dropped, will result in the
connection not becoming established. And a connection blackhole after
session establishment shows up as a stalled connection.
Signed-off-by: Jason Baron
Cc: Eric Dumazet
On 10/22/19 2:17 PM, Yuchung Cheng wrote:
> On Mon, Oct 21, 2019 at 7:14 PM Neal Cardwell wrote:
>>
>> On Mon, Oct 21, 2019 at 5:11 PM Jason Baron wrote:
>>>
>>>
>>>
>>> On 10/21/19 4:36 PM, Eric Dumazet wrote:
>>>>
On 10/21/19 4:36 PM, Eric Dumazet wrote:
> On Mon, Oct 21, 2019 at 12:53 PM Christoph Paasch wrote:
>>
>
>> Actually, longterm I hope we would be able to get rid of the
>> blackhole-detection and fallback heuristics. In a far distant future where
>> these middleboxes have been weeded out ;-)
>
On 10/21/19 2:02 PM, Yuchung Cheng wrote:
> Thanks for the patch. Detailed comments below
>
> On Fri, Oct 18, 2019 at 4:58 PM Neal Cardwell wrote:
>>
>> On Fri, Oct 18, 2019 at 3:03 PM Jason Baron wrote:
>>>
>>> The TCPI_OPT_SYN_DATA bit as part of
ainly not cover all the cases where TFO may fail, but
other failures, such as SYN/ACK + data being dropped, will result in the
connection not becoming established. And a connection blackhole after
session establishment shows up as a stalled connection.
Signed-off-by: Jason Baron
Cc: Eric Dumazet
Cc
On 9/10/19 4:38 PM, Eric Dumazet wrote:
> On Tue, Sep 10, 2019 at 10:11 PM Thomas Higdon wrote:
>>
>>
> ...
>> Because an additional 32-bit member in struct tcp_info would cause
>> a hole on 64-bit systems, we reserve a struct member '_reserved'.
> ...
>> diff --git a/include/uapi/linux/tcp.h b
write queue
> is empty")
> Signed-off-by: Eric Dumazet
> Cc: Jason Baron
> Reported-by: Vladimir Rutsky
> Cc: Soheil Hassas Yeganeh
> Cc: Neal Cardwell
> ---
> net/ipv4/tcp.c | 30 --
> 1 file changed, 20 insertion
On 8/17/19 12:26 PM, Eric Dumazet wrote:
>
>
> On 8/17/19 4:19 PM, Jason Baron wrote:
>>
>>
>> On 8/17/19 12:26 AM, Eric Dumazet wrote:
>>> As Jason Baron explained in commit 790ba4566c1a ("tcp: set SOCK_NOSPACE
>>> under memory pressure"
SO_SNDTIMEO to adjust their write timeout. This mirrors the behavior
that Eric Dumazet introduced for tcp sockets.
Signed-off-by: Jason Baron
Cc: Eric Dumazet
Cc: Ursula Braun
Cc: Karsten Graul
---
net/smc/smc_tx.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/net
On 8/17/19 12:26 AM, Eric Dumazet wrote:
> As Jason Baron explained in commit 790ba4566c1a ("tcp: set SOCK_NOSPACE
> under memory pressure"), it is crucial we properly set SOCK_NOSPACE
> when needed.
>
> However, Jason patch had a bug, because the 'n
On 6/14/19 4:53 PM, Jason Baron wrote:
>
>
> On 6/13/19 5:20 PM, Willem de Bruijn wrote:
>>>>> @@ -237,6 +237,7 @@ static inline int find_next_netdev_feature(u64
>>>>> feature, unsigned long start)
>>>>>
On 6/13/19 5:20 PM, Willem de Bruijn wrote:
@@ -237,6 +237,7 @@ static inline int find_next_netdev_feature(u64
feature, unsigned long start)
NETIF_F_GSO_GRE_CSUM | \
NETIF_F_GSO_IPXIP4 |
On 6/13/19 1:15 PM, Alexander Duyck wrote:
> On Wed, Jun 12, 2019 at 4:14 PM Jason Baron wrote:
>>
>> Now that the stack supports UDP GRO, we can enable udp gso for virtual
>> devices. If packets are looped back locally, and UDP GRO is not enabled
>> then they will b
.
Tested by connecting two namespaces via macvlan and then ran
udpgso_bench_tx:
before:
udp tx: 2068 MB/s35085 calls/s 35085 msg/s
after (no UDP_GRO):
udp tx: 3438 MB/s58319 calls/s 58319 msg/s
after (UDP_GRO):
udp tx: 8037 MB/s 136314 calls/s 136314 msg/s
Signed-off-by: Jason
Add docs for /proc/sys/net/ipv4/tcp_fastopen_key
Signed-off-by: Jason Baron
Signed-off-by: Christoph Paasch
Cc: Jeremy Sowden
Acked-by: Yuchung Cheng
---
Documentation/networking/ip-sysctl.txt | 20
1 file changed, 20 insertions(+)
diff --git a/Documentation/networking
ceive a 32-byte value as output
if requested. If a 16-byte value is used to set the primary key via
TCP_FASTOPEN_KEY, then any previously set backup key will be removed.
Signed-off-by: Jason Baron
Signed-off-by: Christoph Paasch
Acked-by: Yuchung Cheng
---
net/ipv4/
se of this infrastructure in subsequent patches.
Suggested-by: Igor Lubashev
Signed-off-by: Jason Baron
Signed-off-by: Christoph Paasch
Acked-by: Yuchung Cheng
---
include/net/tcp.h | 41 ++-
include/uapi/linux/snmp.h | 1 +
net/ipv4/proc.c| 1 +
net
Demonstrate how the primary and backup TFO keys can be rotated while
minimizing the number of client cookies that are rejected.
Signed-off-by: Jason Baron
Signed-off-by: Christoph Paasch
Acked-by: Yuchung Cheng
---
tools/testing/selftests/net/.gitignore | 1 +
tools/testing
Christoph Paasch
Signed-off-by: Jason Baron
Acked-by: Yuchung Cheng
---
net/ipv4/tcp_fastopen.c | 73 +
1 file changed, 37 insertions(+), 36 deletions(-)
diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c
index 018a484..3889ad2 100644
ne key is set, userspace will simply read back
that single key as follows:
# echo "x-x-x-x" > /proc/sys/net/ipv4/tcp_fastopen_key
# cat /proc/sys/net/ipv4/tcp_fastopen_key
x-x-x-x
Signed-off-by: Jason Baron
Signed-off-by: Christoph Paasch
Acked-by: Yuchung Cheng
---
net/i
Jason
Changes in v2:
-spelling fixes in ip-sysctl.txt (Jeremy Sowden)
-re-base to latest net-next
Christoph Paasch (1):
tcp: introduce __tcp_fastopen_cookie_gen_cipher()
Jason Baron (5):
tcp: add backup TFO key infrastructure
tcp: add support to TCP_FASTOPEN_KEY for optional backup key
On 5/24/19 7:17 PM, Yuchung Cheng wrote:
> On Thu, May 23, 2019 at 4:31 PM Yuchung Cheng wrote:
>>
>> On Thu, May 23, 2019 at 12:14 PM David Miller wrote:
>>>
>>> From: Jason Baron
>>> Date: Wed, 22 May 2019 16:39:32 -0400
>>>
>>>>
Add docs for /proc/sys/net/ipv4/tcp_fastopen_key
Signed-off-by: Christoph Paasch
Signed-off-by: Jason Baron
---
Documentation/networking/ip-sysctl.txt | 20
1 file changed, 20 insertions(+)
diff --git a/Documentation/networking/ip-sysctl.txt
b/Documentation/networking/ip
is used to set the primary
key via TCP_FASTOPEN_KEY, then any previously set backup key will be
removed.
Signed-off-by: Christoph Paasch
Signed-off-by: Jason Baron
---
net/ipv4/tcp.c | 30 --
1 file changed, 20 insertions(+), 10 deletions(-)
diff --git a/net/ipv4/t
Jason
Christoph Paasch (1):
tcp: introduce __tcp_fastopen_cookie_gen_cipher()
Jason Baron (5):
tcp: add backup TFO key infrastructure
tcp: add support to TCP_FASTOPEN_KEY for optional backup key
tcp: add support for optional TFO backup key to
/proc/sys/net/ipv4/tcp_fastopen_key
Document
Demonstrate how the primary and backup TFO keys can be rotated while
minimizing the number of client cookies that are rejected.
Signed-off-by: Christoph Paasch
Signed-off-by: Jason Baron
---
tools/testing/selftests/net/.gitignore | 1 +
tools/testing/selftests/net/Makefile
se of this infrastructure in subsequent patches.
Suggested-by: Igor Lubashev
Signed-off-by: Christoph Paasch
Signed-off-by: Jason Baron
---
include/net/tcp.h | 41 ++-
include/uapi/linux/snmp.h | 1 +
net/ipv4/proc.c| 1 +
net/ipv4/sysctl_net_ipv4.c | 2 +-
ne key is set, userspace will simply read back
that single key as follows:
# echo "x-x-x-x" > /proc/sys/net/ipv4/tcp_fastopen_key
# cat /proc/sys/net/ipv4/tcp_fastopen_key
x-x-x-x
Signed-off-by: Christoph Paasch
Signed-off-by: Jason Baron
---
net/i
From: Christoph Paasch
Restructure __tcp_fastopen_cookie_gen() to take a 'struct crypto_cipher'
argument and rename it as __tcp_fastopen_cookie_gen_cipher(). Subsequent
patches will provide different ciphers based on which key is being used for
the cookie generation.
Signed-off-by: J
On 2/26/19 6:59 PM, Al Viro wrote:
> On Tue, Feb 26, 2019 at 03:35:39PM -0500, Jason Baron wrote:
>
>>> I understand what the unix_dgram_peer_wake_me() is doing; I understand
>>> what unix_dgram_poll() is using it for. What I do not understand is
>>>
On 2/26/19 2:03 PM, Al Viro wrote:
> On Tue, Feb 26, 2019 at 03:31:32PM +, Rainer Weikusat wrote:
>> Al Viro writes:
>>> On Tue, Feb 26, 2019 at 06:28:17AM +, Al Viro wrote:
>>
>> [...]
>>
>>
* if after relocking we see that unix_peer(sk) now
is equal to other, we arrange f
as reported as a hang when /dev/log is closed.
The fix is to signal POLLOUT if the socket is marked as SOCK_DEAD, which
means a subsequent write() will get -ECONNREFUSED.
Reported-by: Ian Lance Taylor
Cc: David Rientjes
Cc: Rainer Weikusat
Cc: Eric Dumazet
Signed-off-by: Jason Baron
---
v2: u
' step. Nevertheless, this has been
observed when the syslog daemon closes /dev/log. Tested against
a reproducer that re-creates the syslog hang.
The proposed fix is to move the wake_up_interruptible_all() call
after the 'free all skbs' step.
Reported-by: Ian Lance Taylor
Cc: Rainer W
kbs' step would in fact cause a wakeup and
a POLLOUT return. So the race here is probably fairly rare because
it means there are no skbs that thread 1 queued and that thread 1
schedules before the 'free all skbs' step. Nevertheless, this has
been observed in the wild via syslog.
The
In prepartion for using some of the high order feature bits, make sure that
virtio-net uses 64-bit values everywhere.
Signed-off-by: Jason Baron
Cc: "Michael S. Tsirkin"
Cc: Jason Wang
Cc: virtio-...@lists.oasis-open.org
---
hw/net/virtio-net.c
tention
is that device feature bits are to grow down from bit 63, since the
transports are starting from bit 24 and growing up.
Signed-off-by: Jason Baron
Cc: "Michael S. Tsirkin"
Cc: Jason Wang
Cc: virtio-...@lists.oasis-open.org
---
drivers/net/virtio_net.c| 23 ++
bsequently overwrite it later if desired via: 'ethtool -s'.
Linkspeed and duplex settings can be set as:
'-device virtio-net,speed=1,duplex=full'
where speed is [-1...INT_MAX], and duplex is ["half"|"full"].
Signed-off-by: Jason Baron
Cc: "Michael
* only do speed/duplex read in virtnet_config_changed_work() on LINK_UP
changes from v2:
* move speed/duplex read into virtnet_config_changed_work() so link up changes
are detected
Jason Baron (1):
virtio_net: propagate linkspeed/duplex settings from the hypervisor
drivers/net/virtio_net.c
On 01/04/2018 01:22 PM, Michael S. Tsirkin wrote:
> On Thu, Jan 04, 2018 at 01:12:30PM -0500, Jason Baron wrote:
>>
>>
>> On 01/04/2018 12:05 PM, Michael S. Tsirkin wrote:
>>> On Thu, Jan 04, 2018 at 12:16:44AM -0500, Jason Baron wrote:
>>>> The ability
On 01/04/2018 12:05 PM, Michael S. Tsirkin wrote:
> On Thu, Jan 04, 2018 at 12:16:44AM -0500, Jason Baron wrote:
>> The ability to set speed and duplex for virtio_net is useful in various
>> scenarios as described here:
>>
>> 16032be virtio_net: add ethtool support
On 01/04/2018 11:27 AM, Michael S. Tsirkin wrote:
> On Thu, Jan 04, 2018 at 12:16:44AM -0500, Jason Baron wrote:
>> The ability to set speed and duplex for virtio_net is useful in various
>> scenarios as described here:
>>
>> 16032be virtio_net: add ethtool support
bsequently overwrite it later if desired via: 'ethtool -s'.
Linkspeed and duplex settings can be set as:
'-device virtio-net,speed=1,duplex=full'
where speed is [-1...INT_MAX], and duplex is ["half"|"full"].
Signed-off-by: Jason Baron
Cc: "Michael
s,
-Jason
linux changes:
changes from v2:
* move speed/duplex read into virtnet_config_changed_work() so link up changes
are detected
Jason Baron (1):
virtio_net: propagate linkspeed/duplex settings from the hypervisor
drivers/net/virtio_net.c| 19 ++-
include/
tention
is that device feature bits are to grow down from bit 63, since the
transports are starting from bit 24 and growing up.
Signed-off-by: Jason Baron
Cc: "Michael S. Tsirkin"
Cc: Jason Wang
Cc: virtio-...@lists.oasis-open.org
---
drivers/net/virtio_net.c| 19 ++
In prepartion for using some of the high order feature bits, make sure that
virtio-net uses 64-bit values everywhere.
Signed-off-by: Jason Baron
Cc: "Michael S. Tsirkin"
Cc: Jason Wang
Cc: virtio-...@lists.oasis-open.org
---
hw/net/virtio-net.c
On 12/27/2017 04:43 PM, David Miller wrote:
> From: Jason Baron
> Date: Fri, 22 Dec 2017 16:54:01 -0500
>
>> The ability to set speed and duplex for virtio_net in useful in various
>> scenarios as described here:
>>
>> 16032be virtio_net: add ethtool su
bsequently overwrite it later if desired via: 'ethtool -s'.
Linkspeed and duplex settings can be set as:
'-device virtio-net,speed=1,duplex=full'
where speed is [-1...INT_MAX], and duplex is ["half"|"full"].
Signed-off-by: Jason Baron
Cc: "Michae
In prepartion for using some of the high order feature bits, make sure that
virtio-net uses 64-bit values everywhere.
Signed-off-by: Jason Baron
Cc: "Michael S. Tsirkin"
Cc: Jason Wang
---
hw/net/virtio-net.c| 54 +-
include/hw/vir
htool commands.
Introduce a new feature flag, VIRTIO_NET_F_SPEED_DUPLEX, which allows
the hypervisor to export a linkspeed and duplex setting. The user can
subsequently overwrite it later if desired via: 'ethtool -s'.
Signed-off-by: Jason Baron
Cc: "Michael S. Tsirkin"
Cc:
-Jason
linux changes:
Jason Baron (1):
virtio_net: propagate linkspeed/duplex settings from the hypervisor
drivers/net/virtio_net.c
From: Jason Baron
sock_reset_flag() maps to __clear_bit() not the atomic version clear_bit().
Thus, we need smp_mb(), smp_mb__after_atomic() is not sufficient.
Fixes: 3c7151275c0c ("tcp: add memory barriers to write space paths")
Cc: Eric Dumazet
Cc: Oleg Nesterov
Signed-off-by: J
On 01/23/2017 01:04 PM, Eric Dumazet wrote:
On Mon, 2017-01-23 at 11:56 -0500, Jason Baron wrote:
On 01/23/2017 09:30 AM, Oleg Nesterov wrote:
Hello,
smp_mb__after_atomic() looks wrong and misleading, sock_reset_flag() does the
non-atomic __clear_bit() and thus it can not guarantee test_bit
On 01/23/2017 09:30 AM, Oleg Nesterov wrote:
Hello,
smp_mb__after_atomic() looks wrong and misleading, sock_reset_flag() does the
non-atomic __clear_bit() and thus it can not guarantee test_bit(SOCK_NOSPACE)
(non-atomic too) won't be reordered.
Indeed. Here's a bit of discussion on it:
http:/
From: Jason Baron
Using a Mac OSX box as a client connecting to a Linux server, we have found
that when certain applications (such as 'ab'), are abruptly terminated
(via ^C), a FIN is sent followed by a RST packet on tcp connections. The
FIN is accepted by the Linux stack but the R
On 01/11/2017 10:48 AM, Eric Dumazet wrote:
> On Thu, 2017-01-05 at 16:33 -0500, Jason Baron wrote:
>
>>
>> +/* Accept RST for rcv_nxt - 1 after a FIN.
>> + * When tcp connections are abruptly terminated from Mac OSX (via ^C), a
>> + * FIN is sent followed by a RS
On 01/11/2017 12:17 AM, Christoph Paasch wrote:
Hello Jason,
(resending as Gmail sent out with HTML)
On 05/01/17 - 16:33:28, Jason Baron wrote:
Using a Mac OSX box as a client connecting to a Linux server, we have found
that when certain applications (such as 'ab'), are abruptly
. 1:1(0) ack 1 win 32768
0.200 accept(3, ..., ...) = 4
// Client closes the connection
0.300 < F. 1:1(0) ack 1 win 32768
// now send rst with same sequence
0.300 < R. 1:1(0) ack 1 win 32768
// make sure we are in TCP_CLOSE
0.400 %{
assert tcpi_state == 7
}%
Signed-off-by: Jason Baron
---
/proc/net/route")
Cc: Andy Whitcroft
Reported-by: Jason Baron
Signed-off-by: Alexander Duyck
---
net/ipv4/fib_trie.c | 21 +
1 file changed, 9 insertions(+), 12 deletions(-)
Ok. Works for me.
Feel free to add:
Reviewed-and-Tested-by: Jason Baron
Thanks,
-Jason
On 11/04/2016 02:43 PM, Alexander Duyck wrote:
On Fri, Nov 4, 2016 at 7:45 AM, Jason Baron wrote:
From: Jason Baron
When read() is called on /proc/net/route requesting a size that is one
entry size (128 bytes) less than m->size or greater, the resulting output
has missing and/or duplic
From: Jason Baron
When read() is called on /proc/net/route requesting a size that is one
entry size (128 bytes) less than m->size or greater, the resulting output
has missing and/or duplicate entries. Since m->size is typically PAGE_SIZE,
for a PAGE_SIZE of 4,096 this means that reads requ
On 10/03/2016 08:19 PM, David Miller wrote:
> From: "Baron, Jason"
> Date: Mon, 3 Oct 2016 20:24:32 +
>
>> Or should I just send the incremental at this point?
> Incremental is required at this point.
Hi David,
Ok. The above question was sent out erroneously. I have already posted
the incr
The group list must be freed prior to freeing the command otherwise
we have a use-after-free.
Signed-off-by: Jason Baron
Cc: Yuval Mintz
Cc: Ariel Elior
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet
Hi,
On 09/23/2016 03:24 AM, Nicholas Piggin wrote:
On Fri, 23 Sep 2016 14:42:53 +0800
"Hillf Danton" wrote:
The select(2) syscall performs a kmalloc(size, GFP_KERNEL) where size grows
with the number of fds passed. We had a customer report page allocation
failures of order-4 for this allocat
From: Jason Baron
Currently, we can have high order page allocations that specify
GFP_ATOMIC when configuring multicast MAC address filters.
For example, we have seen order 2 page allocation failures with
~500 multicast addresses configured.
Convert the allocation for 'mcast_list'
From: Jason Baron
Currently, we can have high order page allocations that specify
GFP_ATOMIC when configuring multicast MAC address filters.
For example, we have seen order 2 page allocation failures with
~500 multicast addresses configured.
Convert the allocation for the pending list to be
+0x10/0x40
[1207325.864263] [] ? kthread_create_on_node+0x180/0x180
[1207325.871288] [] ret_from_fork+0x42/0x70
[1207325.877183] [] ? kthread_create_on_node+0x180/0x180
v2:
-make use of list_next_entry()
-only use PAGE_SIZE allocations
Jason Baron (2):
bnx2x: allocate mac filtering '
On 09/20/2016 07:30 AM, David Laight wrote:
From: Jason Baron
Sent: 19 September 2016 19:34
...
sizeof(struct bnx2x_mcast_list_elem) = 24. So there are 170 per
page on x86. So if we want to fit 2,048 elements, we need 12 pages.
If you only need to save the mcast addresses you could use a
On 09/20/2016 11:00 AM, Mintz, Yuval wrote:
The question I rose was whether it actually makes a difference under
such circumstances whether the device would actually filter those
multicast addresses or be completely multicast promiscuous.
e.g., whether it's significant to be filtering out multi
for the pending list to be done in PAGE_SIZE
increments.
Signed-off-by: Jason Baron
While I appreciate the effort, I wonder whether it's worth it:
- The hardware [even in its newer generation] provides an approximate
based classification [I.e., hashed] with 256 bins.
When configurin
the pending list to be done in PAGE_SIZE
increments.
Signed-off-by: Jason Baron
While I appreciate the effort, I wonder whether it's worth it:
- The hardware [even in its newer generation] provides an approximate
based classification [I.e., hashed] with 256 bins.
When configuring 500 mult
increments.
Signed-off-by: Jason Baron
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c | 131 ++---
1 file changed, 94 insertions(+), 37 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
index
PAGE_SIZE
increments.
Signed-off-by: Jason Baron
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 85
1 file changed, 57 insertions(+), 28 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ma
+0x10/0x40
[1207325.864263] [] ? kthread_create_on_node+0x180/0x180
[1207325.871288] [] ret_from_fork+0x42/0x70
[1207325.877183] [] ? kthread_create_on_node+0x180/0x180
Jason Baron (2):
bnx2x: allocate mac filtering 'mcast_list' in PAGE_SIZE increments
bnx2x: allocate mac filte
Hi,
After looking at this further we found that there is actually
a rate limit on 'rst' packets sent by OSX on a closed socket.
Its set to 250 per second and controlled via:
net.inet.icmp.icmplim. Increasing that limit resolves the
issue, but the default is apparently 250.
Thanks,
-Jason
On 07
From: Jason Baron
The per-socket rate limit for 'challenge acks' was introduced in the
context of limiting ack loops:
commit f2b2c582e824 ("tcp: mitigate ACK loops for connections as tcp_sock")
And I think it can be extended to rate limit all 'challenge acks' on
On 07/01/2016 02:16 PM, One Thousand Gnomes wrote:
>> yes, we do in fact see a POLLRDHUP from the FIN in this case and
>> read of zero, but we still have more data to write to the socket, and
>> b/c the RST is dropped here, the socket stays in TIME_WAIT until
>> things eventually time out...
>
>
On 07/01/2016 01:08 PM, Rick Jones wrote:
> On 07/01/2016 08:10 AM, Jason Baron wrote:
>> I'm wondering if anybody else has run into this...
>>
>> On Mac OSX 10.11.5 (latest version), we have found that when tcp
>> connections are abruptly terminated (via ^C), a F
I'm wondering if anybody else has run into this...
On Mac OSX 10.11.5 (latest version), we have found that when tcp
connections are abruptly terminated (via ^C), a FIN is sent followed
by an RST packet. The RST is sent with the same sequence number as the
FIN, and thus dropped since the stack only
On 06/22/2016 02:51 PM, Eric Dumazet wrote:
On Wed, 2016-06-22 at 11:43 -0700, Eric Dumazet wrote:
On Wed, 2016-06-22 at 14:18 -0400, Jason Baron wrote:
For 1/2, the getting the correct memory barrier, should I re-submit
that as a separate patch?
Are you sure a full memory barrier (smp_mb
On 06/22/2016 01:34 PM, Eric Dumazet wrote:
On Wed, 2016-06-22 at 11:32 -0400, Jason Baron wrote:
From: Jason Baron
When SO_SNDBUF is set and we are under tcp memory pressure, the effective
write buffer space can be much lower than what was set using SO_SNDBUF. For
example, we may have set
From: Jason Baron
When SO_SNDBUF is set and we are under tcp memory pressure, the effective
write buffer space can be much lower than what was set using SO_SNDBUF. For
example, we may have set the buffer to 100kb, but we may only be able to
write 10kb. In this scenario poll()/select()/epoll
From: Jason Baron
sock_reset_flag() maps to __clear_bit() not the atomic version clear_bit(),
hence we need an smp_mb() there, smp_mb__after_atomic() is not sufficient.
Fixes: 3c7151275c0c ("tcp: add memory barriers to write space paths")
Cc: Eric Dumazet
Signed-off-by: Jason Baron
with SOCK_QUEUE_SHRUNK
Jason Baron (2):
tcp: replace smp_mb__after_atomic() with smp_mb() in tcp_poll()
tcp: reduce cpu usage when SO_SNDBUF is set
include/net/sock.h | 6 ++
net/ipv4/tcp.c | 26 +++---
net/ipv4/tcp_input.c | 5 +++--
3 files changed, 28 insertions(
On 06/20/2016 06:29 PM, Eric Dumazet wrote:
> On Mon, 2016-06-20 at 17:23 -0400, Jason Baron wrote:
>> From: Jason Baron
>>
>> When SO_SNDBUF is set and we are under tcp memory pressure, the effective
>> write buffer space can be much lower than what was set using SO
From: Jason Baron
When SO_SNDBUF is set and we are under tcp memory pressure, the effective
write buffer space can be much lower than what was set using SO_SNDBUF. For
example, we may have set the buffer to 100kb, but we may only be able to
write 10kb. In this scenario poll()/select()/epoll
On 12/04/2015 12:03 PM, Joe Perches wrote:
> On Fri, 2015-12-04 at 11:47 -0500, Jason Baron wrote:
>> When DYNAMIC_DEBUG is enabled we have this wrapper from
>> include/linux/dynamic_debug.h:
>>
>> if (unlikely(descriptor.flags & _DPRINTK_FLAGS_PRINT))
>&g
On 12/04/2015 11:12 AM, Dmitry Vyukov wrote:
> On Thu, Dec 3, 2015 at 9:51 PM, Joe Perches wrote:
>> (adding lkml as this is likely better discussed there)
>>
>> On Thu, 2015-12-03 at 15:42 -0500, Jason Baron wrote:
>>> On 12/03/2015 03:24 PM, Joe Perches wrote:
&g
On 12/03/2015 03:24 PM, Joe Perches wrote:
> On Thu, 2015-12-03 at 15:10 -0500, Jason Baron wrote:
>> On 12/03/2015 03:03 PM, Joe Perches wrote:
>>> On Thu, 2015-12-03 at 14:32 -0500, Jason Baron wrote:
>>>> On 12/03/2015 01:52 PM, Aaron Conole wrote:
>>>>
On 12/03/2015 03:03 PM, Joe Perches wrote:
> On Thu, 2015-12-03 at 14:32 -0500, Jason Baron wrote:
>> On 12/03/2015 01:52 PM, Aaron Conole wrote:
>>> I think that as a minimum, the following patch should be evaluted,
>>> but am unsure to whom I should submit it (af
On 12/03/2015 01:52 PM, Aaron Conole wrote:
> Dmitry Vyukov writes:
>> On Thu, Dec 3, 2015 at 6:02 PM, Eric Dumazet wrote:
>>> On Thu, Dec 3, 2015 at 7:55 AM, Dmitry Vyukov wrote:
On Thu, Dec 3, 2015 at 3:48 PM, Eric Dumazet wrote:
>>
>> No, I don't. But pr_debug always computes
On 11/30/2015 04:50 PM, Johannes Weiner wrote:
> On Mon, Nov 30, 2015 at 04:08:18PM -0500, Jason Baron wrote:
>> We're trying to move to the updated API, so this should be:
>> static_branch_unlikely(&memcg_sockets_enabled_key)
>>
>> see: include/linux/jump_
Hi,
On 11/24/2015 04:52 PM, Johannes Weiner wrote:
> The unified hierarchy memory controller is going to use this jump
> label as well to control the networking callbacks. Move it to the
> memory controller code and give it a more generic name.
>
> Signed-off-by: Johannes Weiner
> Acked-by: Mich
On 11/24/2015 10:21 AM, Eric Dumazet wrote:
> On Tue, Nov 24, 2015 at 6:18 AM, Dmitry Vyukov wrote:
>> Hello,
>>
>> The following program triggers use-after-free in sock_wake_async:
>>
>> // autogenerated by syzkaller (http://github.com/google/syzkaller)
>> #include
>> #include
>> #include
>>
still has a pointer to the
> corresponding peer_wait queue. There's no way to forcibly deregister a
> wait queue with epoll.
>
> Based on an idea by Jason Baron, the patch below changes the code such
> that a wait_queue_t belonging to the client socket is enqueued on the
>
1 - 100 of 144 matches
Mail list logo
