Neil,
> Please find here:
> http://people.redhat.com/nhorman/dpdk-1.7.0-0.1.gitb20539d68.src.rpm
Your spec file is broken:
A simple patch:
--- dpdk.spec 2014-05-13 22:43:15.89200 +
+++ ../dpdk.spec.vj 2014-05-13 22:42:40.22100 +
@@ -75,7 +75,7 @@
%build
make O=%{ta
Hi Konstantin,
I apologize for my late answer.
I had also read the Niantic documentation that stated to RSS should not
be enabled/disabled on the fly.
In fact, RSS configuration patches are intended to enable the dynamic
changement of the RSS configuration of a port (RETA, hash key, hash
flags)
> Currently, if there is more memory in hugepages than the amount requested by
> dpdk application, the memory is allocated by taking as much memory as
> possible from each socket, starting from first one. For example if a
> system is configured with 8 GB in 2 sockets (4 GB per socket), and dpdk is
2014-05-12 17:35, Julien Cretin:
> This patch set contains three unrelated patches found by running
> TrustInSoft Analyzer [1] on some parts of the testpmd application:
>
> - The first patch fixes a minor signed overflow in a constant
> expression of testpmd.
>
> - The second patch is not a fix
There was an issue with rebuilding the code following a change to
one of the config files inside the "config" directory. If one did
a "make install T=" and then made a modification to the
defconfig_ file (or applied a patch which modified that file)
a subsequent re-run of the make install command w
Hi Bruce
2014-05-14 11:22, Bruce Richardson:
> There was an issue with rebuilding the code following a change to
> one of the config files inside the "config" directory. If one did
> a "make install T=" and then made a modification to the
> defconfig_ file (or applied a patch which modified that f
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Wednesday, May 14, 2014 11:34 AM
> To: Richardson, Bruce
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] mk: allow updates to build config on make
> install
>
> Hi Bruce
>
> 2014-05-14 11:22
On Wed, May 14, 2014 at 12:46:38AM +0200, Vincent JARDIN wrote:
> Neil,
>
> > Please find here:
> > http://people.redhat.com/nhorman/dpdk-1.7.0-0.1.gitb20539d68.src.rpm
>
> Your spec file is broken:
>
> A simple patch:
> --- dpdk.spec 2014-05-13 22:43:15.89200 +
> +++ ../dpdk.spec.vj
2014-05-14 10:51, Richardson, Bruce:
> > -Original Message-
> > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > Sent: Wednesday, May 14, 2014 11:34 AM
> > To: Richardson, Bruce
> > Cc: dev at dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH] mk: allow updates to build config on m
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Wednesday, May 14, 2014 12:56 PM
> To: Richardson, Bruce
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] mk: allow updates to build config on make
> install
>
>
> You would like to have a sy
2014-05-14 12:33, Richardson, Bruce:
> > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> >
> > You would like to have a synchronization of your generated .config file
> > with the configuration template. But it's not possible for the simple
> > reason that you may have modified your
Hi Ivan,
Then may be instead of rss_hash_conf_update() introduce just rss_hash_update()
that will update only hash key and flags dynamically?
In a same manner we have dev_rss_update/dev_rss_query today?
BTW, I never tried myself to update RSSK and/or MRQE.RSS_Field_Enable on the
fly.
Though it i
>
> It seems reasonable.
> So you plan to send a v2 with this algorithm?
>
I haven't started investigating such a solution just yet, but I'll see what I
can come up with and I'll let you know.
Hi,
I'm having a problem with vmxnet-pmd. When I bind vmxnet3 vNIC to
vmxnet3-pmd, I cannot receive ANY packets from vmxnet-pmd.
I'm running dpdk-1.6.0r2 inside ESXi 5.5 VM. The VM has kernel
3.11.0-12-generic.
One thing I notice is that ESXi host does not pass "pdpe1gb" CPU flag to
guest VMs
Hi Olivier,
2014-05-09 12:32, Olivier Matz:
> It is now possible to build all projects from the examples/ directory
> using one command from root directory.
>
> Some illustration of what is possible:
>
> - build examples in the DPDK tree for one target
>
> # install the x86_64-default-linuxap
Hi Oliver,
Apart from performance impact, one more concern:
As I know, theoretical limit for PA on Intel is 52 bits.
I understand that these days no-one using more than 48 bits and it probably
would stay like that for next few years.
Though if we'll occupy these (MAXPHYADDR - 48) bits now, it can
Hi,
I am new to dpdk.
I built the dpdk-1.6.0r1 and even was able to compile the apps.
Then I am trying to bind my NIC to the dpdk application:
#modprobe uio
#
#insmod x86_64-default-linuxapp-gcc/kmod/igb_uio.ko
#
# ./tools/pci_unbind.py --status
Network devices using IGB_UIO driver
==
Hi,
2014-05-14 09:57, HS:
> I'm having a problem with vmxnet-pmd. When I bind vmxnet3 vNIC to
> vmxnet3-pmd, I cannot receive ANY packets from vmxnet-pmd.
>
> I'm running dpdk-1.6.0r2 inside ESXi 5.5 VM. The VM has kernel
> 3.11.0-12-generic.
[...]
> ":13:00.0" (Port 1) is bound to vmxnet3
Hi,
2014-05-14 17:27, Helmut Sim:
> unlike previous cases I went over in this forum, this NIC supposed to be
> supported (or maybe i read it incorrectly???):
>
> # cat lib/librte_eal/common/include/rte_pci_dev_ids.h | grep 82541GI
> #define E1000_DEV_ID_82541GI 0x1076
> #define E
Thanks Thomas for your quick answer.
Meantime I tried a different device, which is supported according to the
rule you mentioned.
it is a 82575EB but i got the exact same behavior.
# grep 'RTE_PCI_DEV_ID_DECL.*82575'
lib/librte_eal/common/include/rte_pci_dev_ids.h
RTE_PCI_DEV_ID_DECL_IGB(PCI_VEND
Hello all,
Here is a patch series that try to:
- make it easier to globally enable/disable a feature used by multiple targets
- make things clearer when compiling DPDK and when you want to cross-compile it
--
David Marchand
David Marchand (3):
config: factorize configurations
config: rena
linux and bsd default configurations now have their own default common
configuration files.
Specific options are then set in the specific files.
This makes it easier to globally enable/disable some features in DPDK for
multiple targets.
Signed-off-by: David Marchand
---
config/common_bsdapp
The "default" part in configuration filenames is misleading.
Rename this as "native", as this is the RTE_MACHINE that is set in these files.
This should make it clearer for people who build DPDK on a system then run it on
another one.
Signed-off-by: David Marchand
---
config/defconfig_i686-defau
"default" machine type refers to core2 -march, so rename the machine directory
to avoid misunderstanding it.
Signed-off-by: David Marchand
---
config/common_bsdapp|2 +-
config/common_linuxapp |2 +-
config/defconfig_i686-native-linuxapp-gcc
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of David Marchand
> Sent: Wednesday, May 14, 2014 3:58 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH 1/3] config: factorize configurations
>
> linux and bsd default configurations now have their own default
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of David Marchand
> Sent: Wednesday, May 14, 2014 3:58 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH 2/3] config: rename "default" configuration file as
> "native"
>
> The "default" part in configuration fi
Fix spelling errors found by codespell
Signed-off-by: Stephen Hemminger
---
Resend because of no response
lib/librte_eal/bsdapp/eal/eal.c |2 +-
lib/librte_eal/bsdapp/eal/include/exec-env/rte_kni_common.h |2 +-
lib/librte_eal/common/include/i686/arch/rt
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of David Marchand
> Sent: Wednesday, May 14, 2014 3:58 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH 3/3] config: rename "default" machine type as
> "core2"
>
> "default" machine type refers to core2 -march
It makes no sense to inline string functions, in fact snprintf
can't be inlined because the function supports variable number of
arguments.
Signed-off-by: Stephen Hemminger
---
Resend because of lack of response
lib/librte_eal/bsdapp/eal/Makefile |1
lib/librte_eal/common/eal_
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Stephen Hemminger
> Sent: Wednesday, May 14, 2014 4:20 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] Don't inline string functions
>
> It makes no sense to inline string functions, in fact snprintf can'
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Stephen Hemminger
> Sent: Wednesday, May 14, 2014 4:18 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] Spelling fixes
>
> Fix spelling errors found by codespell
>
> Signed-off-by: Stephen Hemminger
>
Hi Thomas,
I made the required change in order to support the 82541GI chipset and I
was able to bind it successfully to the igb_uio ==> thank for your advise.
however i am stll not able to bind the 82575EB... although it looks like it
is already supported ;(
I see there is a 82575EB_COPPER and a
2014-05-14 08:19, Stephen Hemminger:
> It makes no sense to inline string functions, in fact snprintf
> can't be inlined because the function supports variable number of
> arguments.
>
> Signed-off-by: Stephen Hemminger
>
> ---
> Resend because of lack of response
Please do not resend patches.
When running "make config" and additional config.orig file is also
generated, which is intended to hold the original, clean configuration
from the template.
When running make install, we first check if there is no existing
.config file, and run make config if not. If there is a file, we then
check
These patches address a number of issues found during development
and testing of QoS in Vyatta 5600 router.
The existing rte scheduler can only be safely configured once per port
because a memory zone has a fixed size once it is created and can never
be freed or change in size.
This patch changes the scheduler to use rte_malloc instead. This allows
for a port to be reconfigured by doing rte_sched_port_f
Fix build error if RTE_SCHED_DEBUG is enabled.
Signed-off-by: Stephen Hemminger
---
lib/librte_sched/rte_sched.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/lib/librte_sched/rte_sched.c 2014-05-14 09:13:57.377235388 -0700
+++ b/lib/librte_sched/rte_sched.c 2014-05
The rte_scheduler will get stuck and not deliver any more packets
if there are two active subports and then one of them stops enqueing
more packets. This is because of abug in how the grinder state machines
are managed.
If a non-zero grinder is assigned (but not yet active), then the dequeue
would
This set of patches is to the igb_uio kernel driver.
Most of the issues are cosmetic, the only serious bug was
support for E1000 on VMware and KVM.
Use pr_info() and dev_info() to print out log messages where appropriate.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c 2014-05-14 09:01:21.149196906
-0700
+++ b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c 2014-05-14 09:01:21.149196906
-0700
@@ -22,6 +22,8 @@
Don't put capitialization and space in name since it will show
up in /proc/interrupts. Instead use driver name like other kernel drivers.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c 2014-05-14 09:01:30.621248162
-0700
+++ b/lib/librte_eal/linuxapp/igb_uio/i
The module parameter is read-only since changing mode after loading
isn't going to work.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c 2014-05-14 09:32:17.214827796
-0700
+++ b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c 2014-05-14 09:32:17.214827796
-0700
@@ -
The igb_uio driver does not handle interrupt mode correctly.
It doesn't do the right thing and start with the most desired
mode and fall back to legacy modes as needed.
It also has a custom pci config lock was broken for any recent kernels.
Instead pci_intx functions that provide the needed access
Fix whitespace and other style issues reported by checkpatch.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c 2014-05-14 09:32:25.366870293
-0700
+++ b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c 2014-05-14 09:32:25.366870293
-0700
@@ -1,23 +1,23 @@
/*-
* GPL
On Wed, 14 May 2014 17:47:01 +0200
Thomas Monjalon wrote:
> 2014-05-14 08:19, Stephen Hemminger:
> > It makes no sense to inline string functions, in fact snprintf
> > can't be inlined because the function supports variable number of
> > arguments.
> >
> > Signed-off-by: Stephen Hemminger
> >
Having a function to list forwarding modes helps to show them
in cli help and in parameters usage witout duplicating code.
Signed-off-by: Thomas Monjalon
---
app/test-pmd/cmdline.c| 13 +++--
app/test-pmd/config.c | 20
app/test-pmd/parameters.c | 3 ++-
app
Includes replacing existing VMXNET3 driver. Our driver was independently
developed based on the OpenSolaris driver (rather than FreeBSD).
After evaluation, it wasn't worth fixing the original code; wholesale
replacement was lower risk since the driver had been used through
the whole product and QA
Convience to allow for devices that handle no checksum offload,
but do do VLAN offload.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_ether/rte_ethdev.h 2014-05-14 11:26:51.817382768 -0700
+++ b/lib/librte_ether/rte_ethdev.h 2014-05-14 11:26:51.817382768 -0700
@@ -510,9 +510,12 @@ st
Many drivers copy/paste same code to atomicly access link information.
Add functions to do this instead.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_ether/rte_ethdev.h 2014-05-14 11:27:07.409461720 -0700
+++ b/lib/librte_ether/rte_ethdev.h 2014-05-14 11:27:07.409461720 -0700
@@ -17
New common code can be used, eliminating driver code.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_pmd_ixgbe/ixgbe_ethdev.c 2014-05-14 11:26:30.381274255
-0700
+++ b/lib/librte_pmd_ixgbe/ixgbe_ethdev.c 2014-05-14 11:37:51.319893900
-0700
@@ -56,7 +56,6 @@
#include
#include
New common code can be used, eliminating driver code.
Signed-off-by: Stephen Hemminger
--- a/lib/librte_pmd_e1000/em_ethdev.c 2014-05-14 11:26:30.381274255 -0700
+++ b/lib/librte_pmd_e1000/em_ethdev.c 2014-05-14 11:42:25.976657022 -0700
@@ -157,58 +157,6 @@ static struct eth_dev_ops eth_em_ops
Allow setting the default Ethernet address used on device.
The underlying drivers allow it but the DPDK was blocking any
attempts to change Ethernet address on device.
For most devices, this is just the same as filling in address
in mac address table entry 0, but for some devices this requires
spe
Replace the Intel VMXNET3 driver with the one developed by Brocade.
The Intel code was a missing features, and had style and other issues.
Major changes:
- supports multi-segment
- supports vlan offload
- per queue stats
Fixes:
- link update reports link change correctly
*** BLURB HERE ***
Vadim Suraev (2):
Pending timers counting fixed
Bug fixed: when timer's callback (for periodic timer) calls
timer_reset for another timer, a per-core updated flag is
raised which prevents reloading of that periodic timer. The
flag is moved to the timer's
---
lib/librte_timer/rte_timer.c |4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/librte_timer/rte_timer.c b/lib/librte_timer/rte_timer.c
index a3d5cca..f98e904 100755
--- a/lib/librte_timer/rte_timer.c
+++ b/lib/librte_timer/rte_timer.c
@@ -173,7 +173,7 @@ timer_set_
---
lib/librte_timer/rte_timer.c | 14 --
lib/librte_timer/rte_timer.h |3 +++
2 files changed, 7 insertions(+), 10 deletions(-)
diff --git a/lib/librte_timer/rte_timer.c b/lib/librte_timer/rte_timer.c
index f98e904..ea2f22a 100755
--- a/lib/librte_timer/rte_timer.c
+++ b/lib/li
Hello,
Thanks for contributing.
Please reformat your patch as described here:
http://dpdk.org/dev#send
--
Thomas
57 matches
Mail list logo
