h can results in
no connection. Example a peer initiated connection with say
tos 8 while usual connection racing can get downgraded to tos 0
which is not desirable.
Patch fixes above issue introduced by commit
commit d021fabf525f ("rds: rdma: add consumer reject")
Reported-by: Yanjun
On 2019/7/10 13:32, Santosh Shilimkar wrote:
From: Gerd Rausch
The proper "tos" value needs to be returned
to user-space (sockopt RDS_INFO_CONNECTIONS).
Fixes: 3eb450367d08 ("rds: add type of service(tos) infrastructure")
Signed-off-by: Gerd Rausch
Reviewed-by: Zhu Yanjun
Thanks. I am OK
On 2019/7/10 13:32, Santosh Shilimkar wrote:
From: Gerd Rausch
Prior to
commit d021fabf525ff ("rds: rdma: add consumer reject")
function "rds_rdma_cm_event_handler_cmn" would always honor a rejected
connection attempt by issuing a "rds_conn_drop".
The commit mentioned above added a "break",
On 2019/7/10 13:32, Santosh Shilimkar wrote:
RDS composite message(rdma + control) user notification needs to be
triggered once the full message is delivered and such a fix was
added as part of commit 941f8d55f6d61 ("RDS: RDMA: Fix the composite
message user notification"). But rds_send_remove_
On 2019/7/9 6:23, David Miller wrote:
From: Zhu Yanjun
Date: Fri, 5 Jul 2019 02:19:26 -0400
This recv cache is to make NIC work steadily when the system memory is
not enough.
The system is supposed to hold onto enough atomic memory to absorb all
reasonable situations like this.
If anythin
On 2019/7/9 4:52, Jakub Kicinski wrote:
On Fri, 5 Jul 2019 02:19:27 -0400, Zhu Yanjun wrote:
A recv cache is added. The size of recv cache is 1000Mb / skb_length.
When the system memory is not enough, this recv cache can make nic work
steadily.
When nic is up, this recv cache and work queue a
87.926043] Unregistered RDS/infiniband transport
"
So IMO, this commit fixes this problem. The root cause is in the commit log.
Zhu Yanjun
On 2019/6/3 20:43, Yanjun Zhu wrote:
Sorry. Add Håkon Bugge
He told me to notice the memory leak when caches are freed.
Zhu Yanjun
On 2019/6/3 20
Sorry. Add Håkon Bugge
He told me to notice the memory leak when caches are freed.
Zhu Yanjun
On 2019/6/3 20:48, Zhu Yanjun wrote:
When KASAN is enabled, after several rds connections are
created, then "rmmod rds_rdma" is run. The following will
appear.
"
BUG rds_ib_incoming (Not tainted): O
On 2019/6/3 11:03, santosh.shilim...@oracle.com wrote:
On 6/1/19 12:54 AM, Zhu Yanjun wrote:
The variable cache_allocs is to indicate how many frags (KiB) are in one
rds connection frag cache.
The command "rds-info -Iv" will output the rds connection cache
statistics as below:
"
RDS IB Connect
On 2019/4/26 8:44, Santosh Shilimkar wrote:
From: Hans Westgaard Ry
RDS doesn't support RDMA on memory apertures that require On Demand
Paging (ODP), such as FS DAX memory. User applications can try to use
RDS to perform RDMA over such memories and since it doesn't report any
failure, it can
On 2019/4/26 8:44, Santosh Shilimkar wrote:
RDS doesn't support RDMA on memory apertures that require On Demand
Paging (ODP), such as FS DAX memory. A sysctl is added to indicate
whether RDMA requiring ODP is supported.
Reviewed-by: Håkon Bugge
Reviewed-tested-by: Zhu Yanjun
Thanks, Santos
On 2019/3/8 6:01, Gerd Rausch wrote:
The proper "tos" value needs to be returned
to user-space (sockopt RDS_INFO_CONNECTIONS).
Fixes: 3eb450367d08 ("rds: add type of service(tos) infrastructure")
Signed-off-by: Gerd Rausch
In RDS/IB, tos is set in this function. Do you still use RoCE device
On 2019/3/7 10:09, Yanjun Zhu wrote:
On 2019/3/7 9:55, Santosh Shilimkar wrote:
On 3/6/2019 5:49 PM, Gerd Rausch wrote:
Prior to
commit d021fabf525ff ("rds: rdma: add consumer reject")
function "rds_rdma_cm_event_handler_cmn" would always honor a rejected
connection
On 2019/3/7 9:55, Santosh Shilimkar wrote:
On 3/6/2019 5:49 PM, Gerd Rausch wrote:
Prior to
commit d021fabf525ff ("rds: rdma: add consumer reject")
function "rds_rdma_cm_event_handler_cmn" would always honor a rejected
connection attempt by issuing a "rds_conn_drop".
The commit mentioned a
On 2019/3/6 15:04, Gerd Rausch wrote:
Prior to
commit d021fabf525ff ("rds: rdma: add consumer reject")
function "rds_rdma_cm_event_handler_cmn" would always honor a rejected
connection attempt by issuing a "rds_conn_drop".
The commit mentioned above added a "break", eliminating
the "fallthrou
On 2019/3/6 0:48, Gerd Rausch wrote:
Hi Santosh,
On 05/03/2019 08.41, Santosh Shilimkar wrote:
On 3/5/2019 8:33 AM, Gerd Rausch wrote:
If there's a mechanism that ensures compatibility with older (pre-4.1) versions
of RDS I am not seeing it.
Thats handled as part of the connection reject ha
On 2019/2/6 0:19, Yang Wei wrote:
From: Yang Wei
dev_consume_skb_irq() should be called when skb xmit done. It makes
drop profiles(dropwatch, perf) more friendly.
Thanks a lot. I am OK.
Zhu Yanjun
Signed-off-by: Yang Wei
---
drivers/net/ethernet/sun/cassini.c | 2 +-
drivers/net/eth
On 2019/1/19 0:30, Shannon Nelson wrote:
On Fri, Jan 18, 2019 at 2:51 AM Thomas Gleixner wrote:
The recent addition of SPDX license identifiers to the files in
drivers/net/ethernet/sun created a licensing conflict.
The cassini driver files contain a proper license notice:
* This program
On 2019/1/9 11:20, Jia-Ju Bai wrote:
On 2019/1/9 10:35, Yanjun Zhu wrote:
On 2019/1/9 10:03, Jia-Ju Bai wrote:
On 2019/1/9 9:24, Yanjun Zhu wrote:
On 2019/1/8 20:57, Jia-Ju Bai wrote:
On 2019/1/8 20:54, Zhu Yanjun wrote:
在 2019/1/8 20:45, Jia-Ju Bai 写道:
In drivers/net/ethernet
On 2019/1/9 10:03, Jia-Ju Bai wrote:
On 2019/1/9 9:24, Yanjun Zhu wrote:
On 2019/1/8 20:57, Jia-Ju Bai wrote:
On 2019/1/8 20:54, Zhu Yanjun wrote:
在 2019/1/8 20:45, Jia-Ju Bai 写道:
In drivers/net/ethernet/nvidia/forcedeth.c, the functions
nv_start_xmit() and nv_start_xmit_optimized
On 2019/1/8 20:57, Jia-Ju Bai wrote:
On 2019/1/8 20:54, Zhu Yanjun wrote:
在 2019/1/8 20:45, Jia-Ju Bai 写道:
In drivers/net/ethernet/nvidia/forcedeth.c, the functions
nv_start_xmit() and nv_start_xmit_optimized() can be concurrently
executed with nv_poll_controller().
nv_start_xmit
line
On 2018/5/12 0:58, Santosh Shilimkar wrote:
On 5/11/2018 12:48 AM, Yanjun Zhu wrote:
On 2018/5/11 13:20, DaeRyong Jeong wrote:
We report the crash: KASAN: null-ptr-deref Read in rds_ib_get_mr
Note that this bug is previously reported by syzkaller.
https://syzkaller.appspot.com/bug?id
On 2018/5/11 18:46, Sowmini Varadhan wrote:
On (05/11/18 15:48), Yanjun Zhu wrote:
diff --git a/net/rds/ib_rdma.c b/net/rds/ib_rdma.c
index e678699..2228b50 100644
--- a/net/rds/ib_rdma.c
+++ b/net/rds/ib_rdma.c
@@ -539,11 +539,17 @@ void rds_ib_flush_mrs(void)
void *rds_ib_get_mr(struct
On 2018/5/11 13:20, DaeRyong Jeong wrote:
We report the crash: KASAN: null-ptr-deref Read in rds_ib_get_mr
Note that this bug is previously reported by syzkaller.
https://syzkaller.appspot.com/bug?id=0bb56a5a48b000b52aa2b0d8dd20b1f545214d91
Nonetheless, this bug has not fixed yet, and we hope
On 2018/5/11 7:31, Qing Huang wrote:
When a system is under memory presure (high usage with fragments),
the original 256KB ICM chunk allocations will likely trigger kernel
memory management to enter slow path doing memory compact/migration
ops in order to complete high order memory allocations.
Add netdev@vger.kernel.org
On 2018/4/26 12:41, Zhu Yanjun wrote:
When skb is sent, it will pass the following functions in soft roce.
rxe_send [rdma_rxe]
ip_local_out
__ip_local_out
ip_output
ip_finish_output
ip_finish_output2
kfree_skb in soft roce module again.
If I am wrong, please correct me.
Zhu Yanjun
On 2018/4/24 16:34, Yanjun Zhu wrote:
Hi, all
rxe_send
ip_local_out
__ip_local_out
nf_hook_slow
In the above call process, nf_hook_slow drops and frees skb, then
-EPERM is returned when
oce, kfree_skb should not be called
in this module.
I will make further investigations about other error handler after
ip_local_out.
If I am wrong, please correct me.
Any reply is appreciated.
Zhu Yanjun
On 2018/4/20 13:46, Yanjun Zhu wrote:
On 2018/4/20 10:19, Doug Ledford wrote:
On Thu, 2018-04
On 2018/4/20 10:19, Doug Ledford wrote:
On Thu, 2018-04-19 at 10:01 -0400, Zhu Yanjun wrote:
When skb is dropped by iptables rules, the skb is freed at the same time
-EPERM is returned. So in softroce, it is not necessary to free skb again.
Or else, crash will occur.
The steps to reproduce:
On 2018/4/17 23:37, Tariq Toukan wrote:
On 16/04/2018 4:02 AM, Zhu Yanjun wrote:
While a faulty cable is used or HCA firmware error, HCA device will
be offline. When the driver is accessing this offline device, the
following call trace will pop out.
"
...
[] dump_stack+0x63/0x81
[] pa
On 2018/1/9 6:47, Shannon Nelson wrote:
Add unlikely() to a few error checking expressions in the Tx
offload handling.
Suggested-by: Yanjun Zhu
Hi,
I am fine with this patch. I have a question. The ipsec feature is
supported in ixgbevf?
Thanks a lot.
Zhu Yanjun
Signed-off-by: Shannon
On 2017/12/20 8:00, Shannon Nelson wrote:
If the skb has a security association referenced in the skb, then
set up the Tx descriptor with the ipsec offload bits. While we're
here, we fix an oddly named field in the context descriptor struct.
v3: added ifdef CONFIG_XFRM_OFFLOAD check around ca
On 2017/12/21 14:39, Yanjun Zhu wrote:
On 2017/12/20 7:59, Shannon Nelson wrote:
This is an implementation of the ipsec hardware offload feature for
the ixgbe driver and Intel's 10Gbe series NICs: x540, x550, 82599.
Hi, Nelson
I notice that the ipsec feature is based on x540, x550,
On 2017/12/20 7:59, Shannon Nelson wrote:
This is an implementation of the ipsec hardware offload feature for
the ixgbe driver and Intel's 10Gbe series NICs: x540, x550, 82599.
Hi, Nelson
I notice that the ipsec feature is based on x540, x550, 82599. But this
ixgbe driver
will also work wi
On 2017/12/8 3:07, David Miller wrote:
From: Zhu Yanjun
Date: Wed, 6 Dec 2017 23:15:15 -0500
Since both tx_ring and first_tx are the head of tx ring, it not
necessary to use two variables. So first_tx is removed.
These are not variables, they are structure members.
Sure. These 2 structure
On 2017/11/8 11:27, Elior, Ariel wrote:
When "NETDEV WATCHDOG: em4 (bnx2x): transmit queue 2 timed out" occurs,
BNX2X_SP_RTNL_TX_TIMEOUT is set. In the function bnx2x_sp_rtnl_task,
bnx2x_nic_unload and bnx2x_nic_load are executed to shutdown and open
NIC. In the function bnx2x_nic_load, bnx2x_a
Please ignore this mail.
Zhu Yanjun
On 2017/11/8 9:58, root wrote:
From: Zhu Yanjun
When "NETDEV WATCHDOG: em4 (bnx2x): transmit queue 2 timed out" occurs,
BNX2X_SP_RTNL_TX_TIMEOUT is set. In the function bnx2x_sp_rtnl_task,
bnx2x_nic_unload and bnx2x_nic_load are executed to shutdown and op
Hi, all
After this patch is applied, the TCP && UDP tests are made.
The TCP bandwidth is 939 Mbits/sec. The UDP bandwidth is 806 Mbits/sec.
So I think this patch can work well.
host1 <-> host2
host1: forcedeth NIC
IP: 1.1.1.107
iperf -s
host2: forcedeth NIC
IP:1.1.1.105
iperf -c 1.1.1.10
On 2017/3/27 15:37, Sowmini Varadhan wrote:
On (03/27/17 03:06), Zhu Yanjun wrote:
Date: Mon, 27 Mar 2017 03:06:26 -0400
From: Zhu Yanjun
To: yanjun@oracle.com, santosh.shilim...@oracle.com,
netdev@vger.kernel.org, linux-r...@vger.kernel.org,
rds-de...@oss.oracle.com, junxiao...@oracl
On 2017/3/13 14:32, Leon Romanovsky wrote:
On Mon, Mar 13, 2017 at 01:43:45AM -0400, Zhu Yanjun wrote:
When rdma_accept fails, rdma_reject is called in it. As such, it is
not necessary to execute rdma_reject again.
It is not always correct, according to the code, rdma_accept can fail
and will
On 2017/3/13 3:43, santosh.shilim...@oracle.com wrote:
On 3/12/17 12:33 PM, Leon Romanovsky wrote:
On Sun, Mar 12, 2017 at 04:07:55AM -0400, Zhu Yanjun wrote:
When rdma_accept fails, rdma_reject is called in it. As such, it is
not necessary to execute rdma_reject again.
Cc: Joe Jin
Cc: Junx
Sorry. I have no test case to show some issue.
But from Linux Kernel Development Second Edition by Robert Love.
Use spin_lock_irq is dangerous since spin_unlock_irq unconditionally
enables interrupts.
We can assume the following scenario:
--->the interrupt is disabled.
spin_lock_irq(
On 2017/2/10 3:08, Tantilov, Emil S wrote:
-Original Message-
From: netdev-ow...@vger.kernel.org [mailto:netdev-ow...@vger.kernel.org] On
Behalf Of Zhu Yanjun
Sent: Wednesday, February 08, 2017 7:03 PM
To: Kirsher, Jeffrey T ; broo...@kernel.org;
da...@davemloft.net; intel-wired-...@lis
From: Zhu Yanjun
>From the realtek data sheet, the PID0 should be bit 0.
Signed-off-by: Zhu Yanjun
---
Change from v1 to v2:
change the commit header.
drivers/net/ethernet/realtek/r8169.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/realtek/r8169.
Hi,
Please comment on this patch.
Zhu Yanjun
On 2016/12/29 11:11, Zhu Yanjun wrote:
>From the realtek data sheet, the PID0 should be bit 0.
Signed-off-by: Zhu Yanjun
---
drivers/net/ethernet/realtek/r8169.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/et
45 matches
Mail list logo
